In the same seven days, PwC enrolled its three hundred sixty-four thousand consultants in a training program built around Claude and stood up an internal finance practice that runs on it. Anthropic and PwC made the joint pitch in a press release on Thursday: insurance underwriting cycles compressed from ten weeks to ten days, an HR transformation prototype delivered in a week. On Friday, OpenAI shipped a personal finance product that connects ChatGPT to twelve thousand banks and brokerages through Plaid. Mint shut down two years ago and the category never quite reformed. The default just got written by the model company.

Three different moves. One pattern. The frontier labs stopped selling tokens this quarter. They are running operations.

Three moves in one week could be coincidence. The capital lined up behind them is what makes it durable. OpenAI’s deployment unit launched with named partners that include McKinsey, Bain, and Capgemini, plus Goldman Sachs and SoftBank on the money side. Anthology Fund, the joint vehicle Menlo Ventures runs with Anthropic, led a thirty million dollar round into a social commerce company called Nectar the same week. The lab is no longer the supplier in the supply chain. The lab is the table where consulting firms, banks, and growth funds now sit.

A founder I was on the phone with on Wednesday put it this way before I could ask. “We used to pitch CIOs that we’d be the integration layer between the model and their workflow. That sentence does not work this week.” It does not. OpenAI just hired the integration layer.

Here is what gets compressed.

For the past eighteen months, a familiar category of early-stage company sold itself with a version of the same pitch: we will take a frontier model and put it inside your enterprise. Some of those companies were good. Some had a real wedge in a vertical. Most were betting on a margin pool between the lab and the customer that the lab had not noticed yet. The lab noticed.

The same is true for personal finance. Copilot, Monarch, Rocket Money, and Cleo have spent two years competing to be the chat surface where a millennial sees her credit card spend. They were also betting on a margin pool. ChatGPT now sits inside that pool with a Plaid connector and a hundred million weekly users. Some of those companies still have a moat in coaching, advice, or community. The default does not.

The SMB market got the same treatment. Anthropic opened a small business product on Wednesday with native connectors to QuickBooks, PayPal, HubSpot, Canva, DocuSign, Google Workspace, and Microsoft 365. Until last week, “AI for SMB bookkeeping” was a sensible pitch deck slide. It now competes against the connector pack that comes free with the model.

The cap chart and the macro print

While the labs were doing all of this, the public data underneath turned against them.

April CPI printed on Tuesday at 3.8% annual, with the monthly read at six tenths of a point. Producer prices the next day were louder: one point four percent month over month, the largest single-month gain since March of 2022. Senate Republicans confirmed Kevin Warsh as the next Federal Reserve chair on Wednesday afternoon, fifty-four to forty-five. He was sworn in on Friday. The “we get rate cuts in the second half” thesis that growth-stage models have quietly been leaning on did not die forever, though it is closed for the conversation the new chair will have at his first FOMC on June 17.

Capital priced none of that this week. Anduril closed a five billion dollar round on Wednesday at a sixty-one billion dollar post-money. Thrive Capital and Andreessen Horowitz led. The valuation roughly doubled in ten months. The same day, the Department of Defense announced an agreement for ten thousand low-cost hypersonic weapons with Anduril, CoAspire, Leidos, and Zone 5 named as contractors. Two days later, Cerebras priced its initial public offering at one hundred eighty-five dollars per share, above its range, and opened the next morning near three hundred fifty. The book was reportedly twenty times oversubscribed. By Friday afternoon, Bloomberg reported that Anthropic had begun talks for another thirty billion dollar round at a nine hundred billion dollar valuation.

Read those three lines without the macro context and the cycle never tightened. Read them with the macro context, and the gap between what the public data said and what private capital priced is the story.

The disconnect is not irrational. Defense budgets are appropriated, not voted on each month, and the autonomy buyer is uncorrelated with a Fed cut. Frontier AI capex runs on multi-year compute contracts. The names that moved this week were the ones whose customers do not check ten-year yields before signing.

For everyone else, the hardest piece of news arrived more quietly.

The real ceiling

Anthropic re-metered third-party agent harnesses on its paid plans on Thursday. Each plan tier now has a separate monthly credit limit for Claude Code and the Agent SDK, starting around twenty dollars at the bottom. Sam Altman responded a few hours later with a two-month Codex promotion for new business customers. None of that sounds dramatic until you read what ServiceNow and Uber told The Information earlier this month, which is that both companies blew through their entire annual AI token budgets before half the year was finished.

The agent economy is hitting a real ceiling. The ceiling is the cost of a session.

An agent that writes a ten-minute report at five cents per thousand tokens is cheap. An agent that runs for nine hours to refactor a codebase, retries when it fails, and burns a hundred dollars in inference per attempt is not cheap. Companies signed contracts last year on the assumption that volume would scale gracefully. Volume is scaling faster than the contracts assumed.

A few categories get more interesting in that light. Orchestration software that routes tasks to cheaper models. Software stacks for non-Nvidia silicon, now that Cerebras has a public mark. Finance and operations software for AI workloads, the budget controls and per-task monitoring that have existed for cloud spend since the early 2010s and barely exist for inference. None of those is a glamorous category. All of them sit on top of a real and growing customer pain.

Microsoft made a related point this week with less fanfare. The company shipped a vulnerability research system called MDASH, an orchestration of more than a hundred specialized agents over an ensemble of frontier and distilled models. MDASH reported sixteen critical Windows vulnerabilities in Tuesday’s Patch Tuesday cycle, four of them remote code execution flaws. On CyberGym, a public security benchmark, Microsoft claims MDASH outscored Anthropic’s latest research-preview model, called Mythos, by about five points. The number is vendor-claimed for now. The structural read is interesting anyway. Defensive AI started outperforming offensive AI, and the gap closed through careful orchestration of cheaper models. No bigger training run involved. Underpriced category.

What changes from here

A few categories move up the list for any investor doing diligence next week. Forward-deployed engineering teams in markets OpenAI’s services arm will not reach first, meaning regulated healthcare, the European Union, and parts of Latin America. Software that controls the inference bill. Defense autonomy at Series A and Series B, where dual-use buyers and appropriated budgets are insulated from the rate path. Power and grid infrastructure for the data centers that are now displacing residential ratepayers in places like Lake Tahoe, where the utility told forty-nine thousand residents last week that it will stop serving them after May of 2027 in favor of hyperscaler load. Defensive cyber, after MDASH.

A few move down. Generic AI services consulting. AI-native finance and CFO startups that compete with what PwC stood up. Personal finance chat assistants that compete with the new ChatGPT default. SMB bookkeeping AI that competes with Anthropic’s connector pack. Mobile-first coding agents after Codex moved into the ChatGPT mobile app. Anything sold to enterprises last year on the assumption that token prices would fall faster than usage rose.

The questions that experienced investors will ask founders this week are sharper than usual. If a Big Four firm stood up a vertical practice around a competitor’s model on Wednesday, what part of the wedge survives? What does monthly agent burn per active customer look like, and at what contract value does the unit economic flip? If the founder is quoting benchmark numbers, which independent lab has reproduced them?

For our late-stage book, the actionable read is narrow. The cleanest secondary mark for Anthropic this week is the Forge institutional price of two hundred sixty-four dollars and fifty-seven cents, set on Friday. Tokenized SPVs trading on retail platforms imply a much higher number; Anthropic publicly disavowed them on Tuesday, which makes them an unenforceable price signal. The Bloomberg report of a thirty billion dollar primary above nine hundred billion would reset the book if it closes. OpenAI’s last verified mark of eight hundred fifty-two billion from April held quiet. Cerebras gave the public market its first non-Nvidia AI silicon comp at roughly fifty-six billion fully diluted. Anduril at sixty-one billion is the new reference for autonomy. SpaceX is the same watch path it has been for two months, with the S-1 still the catalyst that matters.

The week ahead has its own. Google I/O on Tuesday, where the new Gemini is expected to land short of the frontier ceiling. Nvidia and Walmart earnings on Wednesday and Thursday. Independent reproduction of the MDASH and Mythos benchmark numbers when it comes. And the slow approach of the June 17 FOMC, the new chair’s first.

That meeting is the one that will tell us whether the disconnect between the cap chart and the macro print closes from the top, from the bottom, or for a while not at all.

They fail because they build the wrong thing.

That’s the core problem Bill Albert has spent decades trying to solve.

He’s worked across academia and industry, led global customer experience at Mach49, and now runs Greenlight Idea Lab, where he helps companies validate ideas before they burn time and capital. His focus is simple: reduce the risk of building products nobody wants.

This matters more now than ever.

You can ship a product in days. AI tools cut development time to near zero. But that speed creates a new problem. You can go very fast in the wrong direction.

Here’s how Bill thinks about avoiding that.

Subscribe now

Most founders validate the wrong thing

A common pattern shows up in early-stage startups.

Founders say they’ve “talked to customers.” They feel confident. They start building.

Then the product launches—and nothing happens.

The issue is not effort. It’s what they validated.

There are two separate questions:

1. Is this a real problem worth solving?

2. Will people actually use or pay for this solution?

Most teams jump straight to the second question without answering the first.

They assume the problem exists. They focus on features, design, and speed.

Bill sees this constantly. Products that are easy to use. Clean UI. Thoughtful flows.

But nobody cares.

The difference between interest and demand

One of the biggest traps in customer discovery is mistaking positive feedback for real demand.

People will tell you your idea is great. They’ll say they would use it. They might even ask to be notified when it launches.

None of that matters.

What matters is behavior.

Bill looks for increasing levels of commitment:

• “Send me an email when it’s ready”

• “I’ll join a demo”

• “I’ll bring my team”

• “I’ll sign an agreement”

• “Here’s my credit card”

Each step requires more risk from the user. More time. More reputation. More money.

That’s where signal lives.

If people won’t move up that ladder, you don’t have strong demand yet.

Why “talking to users” is often useless

Many founders rely on small sets of conversations. Friends, early contacts, warm intros.

That creates bias in three ways:

• They’re talking to the wrong audience

• They’re asking leading questions

• They’re looking for confirmation, not truth

Even worse, they ignore negative feedback.

Bill describes a common scenario. A founder hears 58 minutes of hesitation and doubt. Then two minutes of mild enthusiasm.

They leave convinced the product is a hit.

This is human nature. But it’s dangerous.

Real validation requires pressure.

You need to create situations where users can easily reject your idea. If it survives that, then you might have something.

The biggest mistake: falling in love with the solution

Founders are wired to build.

That’s the problem.

They start with an idea. They imagine the product. They picture the outcome.

Then they try to prove it’s right.

Bill pushes the opposite approach.

Start with the problem. Stay there longer than you want to.

Test whether the problem is painful, frequent, and worth solving.

He uses exercises like forcing users to “spend” a fixed budget across different problems. The ones that attract the most “spend” are the ones that matter most.

This forces prioritization. It removes vague answers.

If your problem isn’t winning that competition, it’s not strong enough.

Speed is not your advantage anymore

AI has changed how fast teams can build.

What used to take a month now takes a day. Sometimes an hour.

This creates the illusion that you should just launch and learn.

Bill disagrees.

Fast iteration is useful. But skipping validation upfront leads to wasted cycles.

You end up building multiple products, hoping one works. That costs time, money, and credibility.

His focus is different:

Time to first revenue.

Not time to launch.

Not time to prototype.

Revenue is proof that someone values what you built.

Everything else is a guess.

What real product-market fit looks like

Many founders rely on soft signals.

Signups. Engagement. Positive feedback.

Bill looks for something stronger.

One simple benchmark is the Sean Ellis test:

Ask users: How would you feel if this product disappeared?

If more than 40% say “very disappointed,” you’re on the right track.

But even that is not enough.

You still need behavioral proof:

• Are people paying?

• Are they committing time and resources?

• Are they bringing others in?

The closer you get to real-world commitment, the more confident you can be.

Where AI helps—and where it hurts

AI is powerful for product discovery.

It can help you:

• Understand a new domain quickly

• Generate interview questions

• Synthesize qualitative data

• Identify patterns faster

But there’s a line.

Bill avoids synthetic users and AI-generated personas.

Why?

They tend to be overly positive. They don’t push back. They don’t reflect real behavior.

Until AI can replicate true human decision-making under risk, it can’t replace real users.

Use AI to move faster. Not to replace validation.

A better way to evaluate startup ideas

If you’re building or investing, Bill suggests a simple framework:

1. What does the product do?

2. Who is it for?

3. What problem does it solve?

4. What evidence proves this problem matters?

5. What proof shows your solution works?

If any of these answers are vague, that’s a red flag.

Strong teams can explain all five clearly. And back them with data.

The takeaway

You don’t need months of research.

You don’t need large budgets.

But you do need discipline.

Spend a few weeks validating the problem. Test demand with real signals. Look for commitment, not compliments.

It’s cheap insurance.

Because once you start building, every mistake compounds.

And in a world where building is easy, knowing what to build is everything.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:

00:01 Introduction to Bill Albert

00:30 Bill’s Background and Academic Roots

02:50 Transition from Academia to Industry

03:30 The Problem of Building Products Nobody Wants

05:08 Joining Mach49 and Focus on Product Validation

07:05 Early UX Research in Japan

09:26 Measuring UX and Industry Gaps

11:07 UX Research Misconceptions on Sample Size

12:30 Shift from Usability to Design and Brand

12:50 Common Mistakes in Early-Stage Product Development

14:31 Validating Problems vs Solutions

17:11 Why Talking to Customers Isn’t Enough

18:40 Stress Testing Product Ideas

18:48 Framework for Knowing When a Product Is Ready

19:57 Common Product Failure Patterns

21:23 Evaluating Startups as an Investor

23:06 Market Trends and AI Impact

25:50 Favorite Tools and AI in Research

29:35 AI’s Role in Product Discovery

30:47 Merging Roles: UX, PM, Engineering

31:57 AI: Easier or More Dangerous for Discovery

32:09 Speed vs Insight in Product Development

33:13 Faster Iteration Cycles in Startups

34:18 Future of Product Development and AI

Transcript

Brian Bell (00:01:01): Hey everyone, welcome back to the Ignite Podcast. Today we’re thrilled to have Bill Albert on the mic. He is the founder of Greenlight Idea Lab, a product validation and UX research expert who has spent decades helping companies de-risk innovation, previously leading global customer experience at Mach 49 and authoring one of the foundational books on measuring user experience and product validation. Thanks for coming on the pod, Bill.

Bill Albert (00:01:24): Yeah, it’s my pleasure. Thanks for having me.

Brian Bell (00:01:26): Yeah, so I’d love to start with your origin story. What’s your background?

Bill Albert (00:01:29): So my background’s probably a little bit unusual, and when I tell it, just to give it a caveat, it makes perfect sense to me, but may not make sense to other people. So here we go. So in college, I studied geography, and I went all the way through my PhD. And for my research, I was really focused on spatial cognition, how people navigate in real life. in virtual environments. When I was finishing up, I started into a postdoc looking at the design of navigation systems in cars. And at that time, they were only in Japan. They weren’t in the US yet. And it really got me thinking about kind of design and cognition and how people process information. And at that point, I sort of There’s sort of a fork in the road for people at that stage of academia or industry. I went into industry, but I always wanted to have a connection into academia. I started working for actually a UX team in 1999. I learned a ton about design and usability and really kind of core foundational skills. I jumped over to Fidelity Investments for about seven years running a research team within Fidelity working on a lot of enterprise applications and for both kind of B2B, B2C and really learned a ton. At that point I met what became my mentor Tom Tullis and we ended up doing a lot of stuff together writing books and research and all that and it’s very transformative in terms of my life trajectory after fidelity an opportunity opened up to head up the user experience center at Bentley University it’s a business school but they had a UX center there that kind of operates like a consultancy or you could think of as a teaching hospital for people in UX so we had a staff we had graduate students working on real client engagements and that was wonderful I really enjoyed that I did that for a long time you know again kind of having the kind of being in an academic setting but doing really practical grounded research was very important to me and then what happened after that this is probably in 2021 I was seeing so many products that were we could make them easy to use but no one would want to use them And so that really got me thinking about Yeah, if you can make the most beautiful designed product, it’s perfect in every way, but yeah, nobody cares Exactly, and so A big problem, so I started trying to answer this question like, should we build it? Should we build it? Because I saw a lot of wasted money building product that was never going to be successful.

Brian Bell (00:04:17): Most painful thing as a product manager, as a new product manager, where I’m like, yeah, we’re going to build this. I remember I built this beautiful map-based report back when I was in mobile ads over a decade ago. One internal product of the quarter or year, and everybody was like, high-fiving me. It’s such a great product, and nobody used it.

Bill Albert (00:04:37): Yeah, exactly.

Brian Bell (00:04:38): Nobody cared.

Bill Albert (00:04:39): So what happened was as I was starting to like, okay, I’m ready for a new challenge, an opportunity came along to head up the global CX team at Mach 49, does a corporate venture building. And that really gave me kind of an extra energy, a kick in my step in terms of my career focused on early product validation. And my background is in UX was really about measurement. I’m a data person. As you mentioned, the book I wrote on measuring UX really fit nicely into this new focus around product validation and was there for a few years left and started up Greenlight Idea Lab where I’ve been running this agency for about the last year and a half. Wow. Really focused on product validation for venture studios, corporate ventures, product innovation teams.

Brian Bell (00:05:28): Makes sense. Yeah, I mean, they definitely have a need for someone like you. Circle back to Japan really quick. What was it about Japan as you studied UX coming out of Japan that inspired you? Because it strikes me, especially in the late 90s, early 2000s, they were just like ahead on so many dimensions as a producer of UX. And not so much anymore, but definitely back then they were.

Bill Albert (00:05:50): Yeah I mean it was really interesting so so I went to Japan through a fellowship that had been organized between the Japanese government and Toyota the car company to bring new PhDs over to Japan to work on like technology whatever that means and in Boston where I’m based I was doing my postdoc at Nissan R&D so Nissan hosted me but I was being paid through the Japanese government and they sort of didn’t care what I worked on and what the cool thing was is at Nissan they hadn’t really done any research like we’re building these navigation systems but we have no idea like how to design them in terms of the the experience the UI how how do people process that information what’s too much information what’s not enough we take all that for granted now but back then people just didn’t know and at at Nissan at the plant where I was working they had a driving simulator that was up on a hydraulic lift it had six degrees of freedom and it was literally like this million dollar machine that was essentially only used to take executives is like basically the world’s most expensive video game here jump in the driving simulator do it for a couple minutes and wow that’s cool and then you get out and I got there and I said like this is an amazing opportunity for us to really do more research that no one else has and so we started kind of working with that as a way to really understand kind of cognition during driving we used eye tracking all this cool technology so it was there they had the technology to do the research but they really weren’t research minded like like I was so it was a bit of a clash for at least for the first few months and then you know then they sort of got used to me and what I wanted to do so it worked out I lived in Japan 20 years ago actually when I washed out of Wall Street and what struck me about living there and visiting sense is just how intentional everything is designed in their society yeah they really really think everything through right just just from end to end for the food packaging you know is very innovative how they wrap food and how you unpack it just just every little thing there is just really thought through yeah and even the details like the speed at which you would wrap or unwrap something like you said intentional it’s very like there’s a process of doing it and in the U.S. we’re just like you know rip it apart all this so you’re totally right there they think about design very deeply in a very different way I feel like some of it is is more surface level and they’re not going deep into UX actually for my measuring UX book they translate into Japanese I went there on like a little book tour and I was trying to say let’s build more rigor let’s start measuring experience and it was like totally new concept at that time and kind of 2010 that’s interesting

Brian Bell (00:08:57): Yeah. So you spent decades measuring user experience. What’s a widely accepted UX practice that you think is fundamentally flawed?

Bill Albert (00:09:06): Oh boy, there’s a lot to choose from. I’m going to pick one thing that’s a little bit more nuanced. For people that do UX research, oftentimes we’re trying to find pain points or usability issues as somebody’s using a product. And we can do that with a very small sample size. We can do it with 5, 8, 10 people because what we’re looking for is problem detection we’re just trying to identify problems we’re not trying to measure opinion or preferences where we would need hundreds of people to get something reliable and what I see is in people that either do research or are consumers of UX research they sort of conflate the two they say how can you only test with 10 people and be sure that this is a problem well our Goal or outcome is very different. And sometimes the researchers can do the opposite. Well, they’ll ask people, do you like the red button or the blue button with 10 people? And that’s nowhere close to being reliable. It’s a preference and we need a much larger sample size. So it’s kind of a overall kind of a misuse or misunderstanding of sample sizes related to the research question.

Brian Bell (00:10:21): Yeah, I think that’s really insightful. So looking back at your career, what did you believe about users or product design, UX design that you’ve since completely reversed?

Bill Albert (00:10:32): Oh boy. I think the one thing that Early on, and it’s partly a product of when I sort of started in this field in, let’s say, 1999, early 2000s. We were really focused on just usability, like can people complete the tasks? And we were kind of less focused on the design of it. The design didn’t really matter that much, the branding. And as we’ve sort of figured out more about how to design good experiences or at least make them more usable I’ve really come around to see the power that visual design and brand particularly how that affects kind of their overall experience and so that’s something where I’ve really kind of shifted my attitude probably in the last 10 years about how important that is on top of kind of the more you know basic UX stuff that we’re trying to do.

Brian Bell (00:11:33): Yeah. So back to, you know, you’re helping a lot of venture studios do product discovery or product validation. What do you think a lot of early stage companies and venture studios, people building experiences get wrong?

Bill Albert (00:11:48): I think this is in some way, this is kind of an easy one for me, but it’s so fundamental. There’s a well-known quote. I think it’s fall in love with the problem, not the solution. And I see founders senior leadership falling in love with the solution hey I’ve got a great idea I think there’s going to be a real market for this and go ahead and let’s let’s start designing and building it and launch and learn and it’s great to be excited but oftentimes they have done nothing about validating the pain the problem that they’re solving and I see that over and over and over again and then what we end up with is a product in search of a problem it’s it’s really common and probably to me one of the biggest reasons why either startups fail or they run just run out of runway they overspend capital early on they have to go back to the drawing board it has a whole host of problems but it’s always more fun to start building than going to these fundamental questions it’s very enticing to build it especially when you’ve got a great idea but you gotta like pause that and validate the problem yeah

Brian Bell (00:13:04): it’s kind of like the old adage you know measure twice cut once you really really should make sure what you’re building people actually want so how do you how do you know when people actually you know have the problem or really want the solution

Bill Albert (00:13:19): Yeah, so it’s a really interesting point you just made, Brian. Those are kind of two sort of different things. So one is, how do you know that this is a problem worth solving? Because it could be a problem, but it could be what we’d say a paper cut. It’s really not that big a deal. The only way that we can really truly validate that it’s a problem worth solving is by collecting a lot of data from the target users. We do things like a spend exercise. You have a hundred dollars, euros, whatever to solve problems, right? And we have four in our kind of hypothesized panes. You add two or three things. How would you spend that money, right? virtual money to solve that and we can see if there’s one pain that is just like getting the lion’s share of the money and there are other research techniques to really hone in on this is the big pain right once we do that then the question you said like how do we know people will want it then we’re looking at the value proposition and actually the product solution and for that we’re starting to go from what people say to what they do we’re looking at a demonstration behavioral demonstration of the demand for it so for example going from send me an email when the product is available to let me sign up for a demo a demo with my boss to uh let me sign an MOU here’s my credit card like all the different ladders of kind of commitment that we would be looking for as we develop that prototype or that product that there’s an increase in that level of commitment whether it’s the time of the tension the reputation or the money that people are willing to put forth so that’s how we validate as the product is being developed. But it’s kind of two different ways of validation in that order. The problem, the idea, the value prop, and then the product solution.

Brian Bell (00:15:22): So founders often say they’ve talked to customers, right? They’ve validated it. Why is that almost always insufficient?

Bill Albert (00:15:27): Not most of the time. Sometimes it’s friends and family. They’re always going to tell you they love it.

Brian Bell (00:15:31): Mom loves my product.

Bill Albert (00:15:33): Absolutely. She always will, no matter what the product is, guaranteed.

Brian Bell (00:15:37): My mom thinks I’m the best basketball player on the court.

Bill Albert (00:15:41): I hear you. But they’re oftentimes not asking the right questions the right way. They’re not talking to the right customer profile. Sometimes they only talk to buyers and not the users or vice versa. They’re just, it’s research done very quickly and oftentimes really a confirmation of You know, what they already believe. The way I, or at Greenlight, how we approach the validation is we really want to put it through a stress test. We want to pressure test this as best we can. We want to give people as many Different ways of telling us, no, they don’t like it. They’re not going to use it. And even after that, if they’re saying, yes, I have to have it, right? Then we know we’ve got something. So we try to attack that validation in different ways, different ways of stress testing it.

Brian Bell (00:16:34): Yeah. So what is that framework for knowing when a product is actually ready for market versus just still kind of in that illusion territory?

Bill Albert (00:16:42): Yeah, kind of going back to what I said a few minutes ago, the way we know it is when we’re seeing those behavioral signals. One of the kind of more common measurements for product validation is the Sean Ellis test. And in that, for those that aren’t aware, it’s one simple question is, how would you feel if this product weren’t available? If you had it and I took it away from you? And if more than 40% of the people say, I would be very upset because That’s sort of at least a signal that you have something. Now, I think that’s a great way to start, but it’s not sufficient. We need to look at what those behavioral signals are that matter to that product, to your company, and making sure that we’re hitting those targets. That’s really when we know, all right, we know we have something that there’s going to be true customer demand for. And until that, it’s not an illusion, but it’s a lot of assumptions going on.

Brian Bell (00:17:41): So you’ve seen hundreds of products over the years. What’s the most common failure pattern look like? And what does success look like?

Bill Albert (00:17:49): There’s products that have already been developed and they’re not necessarily new. They’re not trying to do anything new, but they’re being improved somehow. For those types of products, the biggest mistake or failures are really not understanding kind of ultimately what customers care about, how it’s going to make their lives better. for new product, especially in an innovation space. it’s really figuring out what is that what is that problem that people care about what is what are those blockers how can I remove the blockers right when you’re able to do that then you’re gonna be successful so there’s just that is really to me that the difference between success and failure ultimately I mean people are get so enamored with with new technology and AI but ultimately it’s about solving a problem and you I know I’m kind of harping on that, but it is so fundamental to product validation.

Brian Bell (00:18:47): So put your investor, your VC hat on for a second and you’re evaluating early stage startups. How would you evaluate that they’ve validated that this is a problem?

Bill Albert (00:18:57): Well, the first thing, I mean, whenever I’m learning about a new product, like, you know, tell me about the product. What does it do? Right. I’m trying to understand the value proposition. Then I want to know who is it for? What is your ideal customer profile? and then I want to know what problem are you solving and then I want to understand something about the context of use and if they’re not able to give a really crisp clear explanation of the problem then I’m starting to wonder a little bit and I want to know what data they have to support what is the evidence to support that that is a problem worth Solving and we get kind of into the weeds around the data and the methodology that’s sort of my thing anyway so I like going down there but that’s really what I’m looking for is I want you to convince me in many different ways that there’s really strong evidence that this is the problem that you’re that you’re going after that really matters to people and then And then the idea. So once you know that problem, then there’s the idea or the value proposition. There’s a lot of different ideas to solve the same problem. And then we want to have evidence that you have the right idea to solve that.

Brian Bell (00:20:11): So what are you seeing in the market? You know, it’s 2026 now. Like, what are you seeing out there that’s shifting?

Bill Albert (00:20:18): Well, I mean, obviously it’s AI. Everything is AI. I mean, try to find a product that doesn’t have AI somehow incorporated into it. Sometimes they used to switch in the old days. We’d say IA because that’s information architecture, but another start. But what we’re seeing now with So many products, you know, through vibe coding and everyone saying, you know, hey, we can build this product really quickly. Let’s just launch it out there. You know, let’s launch and learn. We’ll be able to make tweaks. And on the surface... Well, it used to take you a month, can it take you an hour with cloud code and just like get it out there and see if people use it? Yeah, yeah. I can see where that would be very enticing. I’m not a big fan of that because people think that the goal is to launch a product the goal is really for me time to first revenue right through a product how long is it going to take you to get to first revenue and I contend I strongly believe that when we do that validation up front we’re saving time we can get to first revenue faster than if you’re just throwing spaghetti at the wall let’s build this and launch it see let’s look for those behavioral signals you have to have it out there enough you’re starting to hurt your reputation if you’re like throwing a bunch of different things out there you keep going back to the drawing board or you could launch multiple products see what sticks But even then, to me, it seems like a kind of inefficient way of doing things So that’s one sort of practice I’m seeing that I’m very kind of weary of And the other thing is using IA in a way that doesn’t always make sense And when we look at product that is kind of an AI-based solution We need to look at it, not just, you know, does the problem Is it solving a problem that matters? And, you know, but we need to look at it in other ways with other lenses like around trust and transparency. You know, is that solution better than what you could get on your own or through some other mechanism? So kind of a slightly different way of stress testing those AI based solutions.

Brian Bell (00:22:24): Yeah, that’s really cool. So what are some of your favorite tools that you’re using on a day to day basis?

Bill Albert (00:22:29): Yeah, so in terms of AI, the way I see it is it’s awfully powerful that lets us run faster. It does things like ramps us up on a domain. We can develop interviews, interview guides, helps us identify the customer profiles with recruiting. It helps us with qualitative data synthesis. A lot of things that help us just run much faster, which is great. But what we don’t use it for is is we’re not using creating synthetic personas or running synthetic interviews. There are just a lot of people I know believe that they’re crap I believe that too I just maybe at some point they’ll be so good but we’re not there yet we do have a portfolio company that does this they use AI and synthetic audiences to sort of you know test if you’re like a brand manager at Procter & Gamble or like you make makeup or whatever for Sephora like what are people in your ICP going to think about this synthetically I think it could be helpful to sort of inform a little bit there’s another product that we have in our portfolio I don’t know if you’ve come across them Conveyo they just raised a pretty big ground they automate qualitative research through video and voice interviews so basically helps you kind of do the AI research at scale where previously you would hire a whole team you’d have to like go to out and like kind of do it manually and that just does it with AI Yeah that’s awesome and I’m all for like until my agent is selling your agent something and the agent is the buyer until that happens we have to have a human a human is making a decision so there has to be it has to be sort of grounded in some kind of human experience human behavior human you know preferences so Use AI to do all that, to capture qualitative data at scale, to synthesize, pull out key patterns. It’s wonderful for that. But when I’ve seen demonstrations of synthetic interviews, everything in AI is always so positive and it’s hard to develop personas that are like essentially real people. Everyone’s sort of... Kind of psychophantic, right? The way AIs are trained is with reinforcement learning and so it’s sort of like you’re basically giving it a little treat when it does a good job. and so it’s always trying to please you know I literally have that in my custom instructions for all my AIs like please give it to me straight do not sugarcoat things be honest objective you know things like that Yeah, I will say one of the tools that we’ve built, we’re almost ready to release, is kind of a rapid assessment tool using AI that kind of walks people through kind of a series of about 10 questions just to understand, is this product kind of viable, kind of just as a sniff test? Is it viable? And kind of where are the gaps? What are you missing? What do you need to do next in terms of getting to a much higher level of kind of reliability in the assessment of it? So it kind of gives people a kind of a jumpstart into thinking about, all right, this seems like a good idea, but these are the things that we need to focus on.

Brian Bell (00:25:36): There is also, I’ve heard from Boots on the Ground, that there’s sort of this merging happening between the product manager, the UX person and the engineering manager. Have you seen this on the ground?

Bill Albert (00:25:48): Oh yeah I have I think for people who are trained in UX they’re usually designers or researchers sometimes strategy people and you know for decades we had a way of doing things we’re running these projects right on different products and and that all made sense and and then AI comes along and it can do so much now we’re a UX researchers or designers are able to kind of Almost be like a maestro, right? And trying to organize everything together. So it’s that role between researcher, designer, product manager is definitely merging into somebody who can kind of understand all three and how they interact to deliver product quickly. Right. And people that aren’t able to make that pivot, if they’re too focused on this is what I do, I’m only a researcher, it’s going to be tough. Right. Creative destruction from AI. What do you think AI makes? Does AI make product discovery easier or more dangerous? Both It makes it easier because it gives us a jump on where we should talk to people or some different types of questions we might want to ask or understanding the context in which they’re working stuff like that Yeah It’s more dangerous because if we start going down the road of synthetic interviews, we can easily, easily jump to the wrong conclusions. We’re not going to get the depth, the level of insight. It’s powerful. You know, use it in the right way is kind of the bottom line.

Brian Bell (00:27:30): so if AI can generate products faster does that make user insights more valuable or less valuable

Bill Albert (00:27:36): I would say if it can generate products faster it makes time to insights faster let’s just say so you know the more I’m putting out there the more data I’m getting right so we can get more insights but really the measure should be we getting to those milestones like how quickly can we get through the different stage gates to actually launch a product and start to grow and scale like I said time to first revenue that’s what we care about insights are only as good as you’re able to kind of leverage them in order to get to a successful product launch so AI probably more insights does it give you better or new insights sure it might but it’s got to be done in conjunction with some kind of human data right

Brian Bell (00:28:28): yeah I think we’re seeing the the collapse of time scales to validation happen because I see it in my portfolio right I see founders building faster than ever And I see them getting traction faster than ever as well. Because I think because they can be really responsive to customers and synthesize a lot of data very quickly, they can, you know, what used to take, you know, we used to have this joke in product management that the quarterly roadmap was really a yearly roadmap. So what we thought we’d get done in a quarter took a year. And now I think, you know, you think you can get done in a quarter, you get done in a month or a couple weeks in one sprint maybe. And so I think you’re just building faster than ever because you can generate all this code faster than ever. It’s just driving a lot of iteration. What do you see out there in the future? You know, if you look, you know, five or 10 years out, what are you excited about?

Bill Albert (00:29:18): You know, famous for being a really bad predictor. I can give you examples of when I convinced my roommate not to invest in Starbucks the night before they went public with navigation systems. I said, this will never make sense in the US. We all know where we’re going and we know how that played out. So I am very leery of predictions I make. I am a data person. So where things are going, honestly, I have no clue. I know things will be different and I know things will change that’s about the only thing I can say

Brian Bell (00:29:49): I think in the future you’ll probably be able to say hey I have this hypothesis to AI to your super intelligent assistant and be like hey I really want to build something in XYZ area what kind of problems are people having in that area that they can’t solve themselves that we can build and I think that will be the one person unicorn of the future it’s just like this person just orchestrating lots of different AIs to do lots of different things.

Bill Albert (00:30:16): Yeah.

Brian Bell (00:30:17): That’s the vision.

Bill Albert (00:30:18): I mean, maybe it’s we have AI agents designing and creating product for other agents and they already know what they want and what’s going to work.

Brian Bell (00:30:26): Yeah. Any takeaways for the founders listening out there? Any final words of advice? There’s lots of founders that listen to this.

Bill Albert (00:30:34): Yeah, I would say don’t fall in love with your solution. Focus on solving a problem that matters, that people will say, I have to have this product. Don’t take it away from me. You’ve got to look for that. The motion in the voice, behavioral signals, everything that’s going to tell you that it’s going to work. Validate upfront, just spend those extra few weeks. It’s cheap, cheap insurance to really de-risk your innovation. Ultimately, that’s what we’re talking about. It’s not going off for a year and collecting tons of data. It can be done quickly, cheaply, just to really dramatically increase your likelihood of product success and growth and scale.

Brian Bell (00:31:20): What do you see some, like speaking of falling in love with the problem, What do you see as kind of a best practice for running a backlog these days? You know, back in the day when I used to run backlogs, it was, of course, a spreadsheet. And then you had maybe a weighted matrix of pain and effort to sort of prioritize it and things like that. What do you see as some best practices these days?

Bill Albert (00:31:41): Well, it really depends on where you are in the product development process.

Brian Bell (00:31:47): Let’s say you’re looking at value propositions, right?

Bill Albert (00:31:50): So you know that this is a problem. Now you’ve A lot of times people will hone in on one or two value props and, you know, in terms of like the backlog is really testing a lot of different types of value propositions that are wildly different from one another. Having all that data in a central spot and kind of being able to organize it being able to do kind of apples to apples to see what ideas are the strongest to understand why and then moving over into a product solution and seeing really what resonates what are those metrics that matter most things like that

Brian Bell (00:32:25): Let’s wrap up some rapid fire questions. What’s one product you think succeeded for the wrong reasons?

Bill Albert (00:32:30): I’m not sure this is going to be rapid fire. A product that succeeded originally designed to take the soot off walls during the 1930s. And 20 years after that, they discovered that kids were playing with it. Kindergarten teacher gave it to some kids and they took out the chemicals, put in color, and that’s Play-Doh.

Brian Bell (00:32:51): amazing what’s a product you were sure would work but failed what did you miss I

Bill Albert (00:32:57): cannot honestly there’s no product trained to be super skeptical and so there’s

Brian Bell (00:33:02): nothing that wow that’s good I have one that map-based report you know there’s probably others if I wrap my brain around it but what’s the most misleading metric founders rely on when validating ideas

Bill Albert (00:33:09): I think people focus too much on product launch We launched the product success instead of metrics that more tied to their success, right? Whether it’s that first revenue or hitting certain financial goals, things like that. I mean, that’s usually more on more established product innovation teams. The other is focused too much on the speed and running so fast that they’re ultimately wasting time.

Brian Bell (00:33:40): They can be running really fast in the wrong direction. Getting to nowhere fast, I think is the phrase. What’s one question every founder should ask users, but almost never does? And we covered a couple of variations of that in our discussion, but yeah.

Bill Albert (00:33:52): Yeah, yeah, sure. I will say founders often don’t deeply appreciate the context of use. They’re used to just talking to people on computers. They’re not going out into the field. They’re not understanding how somebody might use your product kind of in the wild. And that can give you a lot of important insights.

Brian Bell (00:34:08): How do you know a product or startup has product market fit?

Bill Albert (00:34:13): The level of knowing or your confidence or certainty should be moving up as you start to go from pain to idea to solution. And once you’re getting into more high-fidelity prototypes, then you should be quite confident, right? Because you have those behavioral signals. That’s what you need at that point. That’s the only way you’re really going to be You know, very, very confident.

Brian Bell (00:34:41): And I’ll ask that kind of the same question in a different way. What’s a belief you have? What’s a belief about product market fit generally that you strongly disagree with?

Bill Albert (00:34:49): yeah this is I think I would probably go back to that speed I am all for working fast but I just I think that people are just they’re just moving too quickly and they’re not they’re not really weighing the right things in the right way they’re not looking at the data that’s going to be most predictive of success

Brian Bell (00:35:10): love that and what is the data that’s most predictive of success in your your experience

Bill Albert (00:35:13): Well, it’s going to be dependent on that product. So if it’s like an enterprise application, it could be a letter of intent, an MOU. It could be signing some other type of agreement. Once it does come out, they’re going to purchase it. Or B2C could be payments, people paying in advance for something. So those are usually the strongest signals as opposed to anything that people are telling you. They’re going to tell you they love it no matter what, but... When you ask for their credit card, it’s a different story.

Brian Bell (00:35:45): You’ve probably dealt with this many times in your multi-decade career, but how do you kind of guard against the executive hippo, highest paid person’s opinion in the room deciding what to build next or the seagull who flies in and shits over everything and flies out? How have you counteracted that and how should maybe the product managers and UX researchers listening to this counteract that in their corporate lives?

Bill Albert (00:36:10): Sure. I think there’s really two ways to counteract that. One way is to kind of be this objective Switzerland person and say like, it’s not me. I just want to share with you the data that we have. Right. We have, you know, 50 people saying that they wouldn’t use it or whatever it is.

Brian Bell (00:36:29): Yeah. We surveyed 100 people and 80 people said, hell yeah, I need this. Or 80 people said, but this would be really painful if this was taken away or whatever the data is.

Bill Albert (00:36:40): Yep, I am a conduit of the data. Okay, the second way to do it is kind of a longer term play that I really believe in is about relationships, building relationships with senior leadership. Get to know them, have lunch with them, let them understand what you do, and especially the value that you deliver. And once they’ve seen you demonstrate that value, you build more and more trust. they’re less they’re more likely to bring you into those conversations they’re less likely to be that hippo or seagull so build relationships a lot of what we do whether it’s UX CX product it is about relationships ultimately

Brian Bell (00:37:25): so you wrote a book measuring the user experience what would you rewrite today if you if you had to release a new edition of that

Bill Albert (00:37:32): Well, funny you ask, because I just finished a few months ago, the fourth edition. It’s not out yet. I think maybe early fall. For that edition, we included a lot of stuff around AI, obviously. I even hate to say this, but if sometime down the road there was a fifth edition one thing that we do with product it’s in the book it’s really focused on digital products but a lot of the concepts a lot of the questions that we’re asking apply to services too there’s a whole field of service design it’s really emerging and I would like to try to say what are some different ways that we can measure a service experience think about your journey through the airport from When you step into the airport till you get to your destination, all those different touch points, all those different channels, we’re not really looking at that as a field in a really systematic kind of rigorous way.

Brian Bell (00:38:28): Some people are, but I’m just finishing the first edition of a book forthcoming. I’m not ready to kind of announce it yet, but any advice as I kind of shop it around to publishers or maybe just upload it to Amazon or kind of what are some takeaways from doing this process?

Bill Albert (00:38:44): I mean the way we did it early on going back to Tom Tullis who was a big mentor for me I walked into his office one day and I said hey I have this idea there’s no book on measuring UX whole quant UX hadn’t come I mean that was just not a thing back right in 2006 and it was really to me kind of a selfish exercise just to kind of like make sense of all this we went with that he loved the idea and he happened to be meeting with a publisher Elsevier who’s our publisher and the acquisitions editor and pitched the idea and she’s like oh wow I really like this and so we wrote sample chapter outline all that stuff so for us it was kind of a fairly easy seamless but I know a lot of people don’t have that experience I would say if you really want to have it broadly distributed is find a potential publisher write one or two sample chapters ask them what the book’s finished well there you go book ready yeah then shop it around and see yeah I have the proposal I’m starting to shop it around right now we’re meeting with publishers this week and I just wondered if you’ve had any advice look at the terms carefully you know oftentimes there’s a step so if you for us I think it was the first 5,000 books we got some percent and then when we hit 7,500 and 10,000 kind of you get higher percentage the more you sell make sure that that’s fair yeah well maybe I’ll reach out when I when I get the proposal if I’m lucky enough to get one kind of the hard part is over you’ve written the book so yeah that’s the hard part right I mean I’m kind of going into the copy editing phase now which I’m not excited about but it’s fun to kind of create the book and now I have to go like kind of refine it a lot of publishers have copy editors yeah if I get it published they’ll they’ll just take it and copy edit it right give me lots of comments and yeah yeah that’s why I’m waiting to see if I can get somebody to pick it up then I don’t have to pay a copy editor which I will do if I need to yeah what’s the biggest lie founders tell themselves during customer discovery that people love our product love that they can’t live without it you know it’s it’s again it’s it’s they’re falling in love with the product and how could other people not love it like we do a lot of founders are blind to that a lot of founders are blind you know as a VC I mean thousands of companies every year right that’s a big big problem I see is like I just don’t see anybody needing this you can’t yeah the metrics the metrics clearly say nobody needs it I don’t think anybody needs it I don’t think it’s a big problem and clearly your metrics show that yeah I think it’s that founder gaslighting themselves into thinking that hey this is the best thing ever and everyone needs this like let me show you a picture of my kid and my baby isn’t he or she the cutest thing you’ve ever seen literally like an ugly dog sitting in a yeah like no baby could be cute so you know and it’s hard to get that even when people tell you you don’t always hear that right so They’ll be telling you for 58 minutes basically why they’re, in many different ways, why they’re not interested in it, why it’s not for them or why now is not the right time, etc. And then there’s two minutes where they’re like, yeah, this is really cool. I really like it. And that’s what they latch on to. those two minutes instead of the 58 and that’s human nature I mean that’s that it’s so we have to be optimist right we have to be optimist yeah we we do and it’s good that we are I suppose but yeah we that’s why having kind of objective third party let’s run this through the ringer whatever you want to call it to really make sure that you know back to that first question I said should we build it let’s make sure the answer is yes definitively yes there is going to be a demand for it that’s what I’m trying to do save companies save companies money by not building the wrong product and help them achieve their

Brian Bell (00:42:42): and then building it correctly. What’s an industry where UX is still massively underdeveloped?

Bill Albert (00:42:47): It’s, I don’t think about it industry. I think about it more in markets. So I think that there’s some markets where I just don’t see it as being valued. And now what are those markets? Probably Asia, Africa, South America, outside of US and Europe. But I just, yeah, they just- It makes sense. They’re a little bit behind generationally Yeah, exactly. Sometimes they just haven’t come around to really seeing, you know, what you’re selling is an experience. Usually it’s an experience. Some people call this the experience economy, right? Starbucks is not selling coffee. They’re selling an experience that happens to be around coffee, but that’s what they’re selling, right? It’s not just a cup of coffee. So the faster that people kind of can recognize that, the more that’s going to kind of shift their mindset.

Brian Bell (00:43:46): Yeah. Tell me about your favorite product or experience.

Bill Albert (00:43:49): My favorite product. That’s...

Brian Bell (00:43:52): It’s an interview question, but I thought it was appropriate here.

Bill Albert (00:43:56): I mean there’s different apps I use but I’m very transactional I don’t you know product I don’t get I’m not not a person that gets excited about technology I sort of I’m excited about the experience around it so all the time you were wowed where you just use something and you were just like wow this is amazing there’s an app called all trails and if anyone likes hiking being in the outdoors it’s like got literally every trail and every park and every place certainly in the U.S. and abroad as well I think it tells you everything about it shows you that the profile and shows you all these reviews and when to go and you know and it’s really nice and then when you’re out there walking you have your phone you can see where you are in kind of what more elevation you have it’s it’s really um handy it’s really opened us up to exploring more around where we live and it’s a technology that checks a lot of my boxes around an experience that really matters a lot to me

Brian Bell (00:44:55): and last question what do you want your legacy to be

Bill Albert (00:44:59): You know, when I’m thinking about my career, there’s really been two big themes for about the first 20-something years. It was around measurement and data. And then about four or five years ago, it’s kind of pivoted towards product validation. And that’s really what I’m focused on now. And they certainly, one kind of feeds into the other, but I I want, in terms of legacy, one is I’ve done a lot of mentoring. I’ve probably mentored over 100 graduate students, and I feel really grateful for that opportunity and a lot of people I’ve helped in their careers. But the second thing to of legacy would be helping move our field forward you know helping helping with the whole quant ux movement through our books and our research and now i want to help bring product validation of make it kind of a more rigorous scientific reliable way of really de-risking so that’s really ultimately you know it’s about the people and it’s about helping organizations really well really enjoyed talking to you bill

Brian Bell (00:46:06): where can folks find you online

Bill Albert (00:46:07): greenlightidealab.com please reach out love to hear from you thanks so much really enjoyed it thank you

But it still feels… off.

You ask a question. It responds. You ask again. It responds again. The interaction is rigid, transactional, and predictable. It doesn’t feel like talking to someone. It feels like issuing commands to a system.

Vish Hari thinks that’s the core problem—and the biggest opportunity.

Subscribe now

From Astrophysics to AI

Before founding Ego AI, Vish Hari was deep in research. He studied astrophysics and worked on early deep learning models back in 2012, trying to detect exoplanets—planets that could support life.

That was before AI became mainstream.

He trained one of his first models in 2013, writing CUDA code just to get it running. From there, he moved into AI engineering, eventually working in applied research at Facebook.

After nearly a decade in the field, he noticed something important.

Every major AI lab was chasing the same goal: superintelligence.

But almost no one was focused on making AI feel human.

The Problem: AI Doesn’t Feel Like a Partner

Right now, AI behaves like a tool.

You give it instructions. It executes. That’s it.

Vish describes it as “staccato”—a stop-and-go interaction that lacks flow.

There’s no continuity. No initiative. No sense of presence.

And that creates a strange dynamic: humans adapt to AI instead of AI adapting to humans.

You see it in how people interact with tools today:

• Over-explaining context like they’re writing prompts

• Structuring conversations unnaturally

• Treating AI like a machine, not a collaborator

This wasn’t supposed to be the end state.

Ego AI’s Bet: Behavior Is the Missing Layer

Ego AI is built around a simple idea:

The next generation of AI won’t win on intelligence. It will win on behavior.

That means building systems that:

• Remember context across time

• Initiate conversations

• Interrupt when necessary

• Adapt to your personality

• Develop their own “internal life”

Not just smarter outputs. Better interactions.

Vish isn’t trying to build a system that knows everything. He’s trying to build one that feels like someone.

Why Big Models Aren’t the Answer

The dominant belief in AI right now is simple: more compute, more data, bigger models.

Vish disagrees.

He argues that intelligence isn’t just about recall or scale. It’s about forming new connections, adapting to context, and behaving in ways that feel natural.

Humans don’t have perfect memory. We forget things. We shift depending on who we’re talking to. We behave differently in different contexts.

That “flawed” behavior is actually what makes interaction feel real.

Ego AI is leaning into that.

Learning From Video Games

One of the more unexpected parts of Ego AI’s research comes from video games.

Why games?

Because they simulate reality.

In games, characters already feel like they have personalities—even without modern AI. Think about enemies in games like Shadow of Mordor or the social dynamics in MMORPGs.

Players build relationships with entities that aren’t real.

That’s a powerful signal.

Ego AI spent years studying these environments to understand what makes interactions feel alive. The goal is to bring that same sense of presence into AI systems.

The Hard Tradeoff: Utility vs. Fun

Most AI products today focus on utility:

• Write this email

• Summarize this document

• Generate this code

That’s where the demand is.

But consumer behavior tells a different story. People also want:

• Companionship

• Conversation

• Entertainment

The challenge is combining both.

Go too far into utility, and the product feels cold.
Go too far into personality, and it becomes a gimmick.

Ego AI is trying to balance both—and Vish admits this is still an open problem.

A Future With AI Relationships

Vish’s long-term vision is bold:

People will have as many AI friends as they do human friends.

Not assistants. Not tools. Friends.

These systems will:

• Evolve over time

• Share experiences with you

• Develop unique personalities

• Adapt to your life

It’s not about replacing human relationships. It’s about expanding what relationships can look like.

A Personal Turning Point

This vision isn’t just academic.

In early 2025, Vish was the victim of a near-fatal assault. He suffered a traumatic brain injury and lost significant memory.

Recovery was slow.

He describes it as regaining his mind piece by piece—almost like watching an AI system retrain itself in real time.

At one point, he could solve complex math problems but struggled with emotional control. His cognitive abilities returned at a different pace than his behavior.

That experience reinforced his belief:

Intelligence alone doesn’t define being human. Behavior does.

Why This Matters Now

AI is at an inflection point.

The first wave was about capability—what AI can do.
The next wave is about interaction—how it feels to use.

Right now, tools like ChatGPT dominate because they’re useful.

But usefulness alone won’t define the next generation of products.

The companies that win will make AI:

• Feel natural

• Feel personal

• Feel alive

That’s the space Ego AI is going after.

Final Thought

We’ve spent years making machines smarter.

The next step is making them relatable.

If Vish is right, the biggest shift in AI won’t come from better answers.

It will come from better relationships.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL
🙏 Join the conversation on your favorite social network:


Chapters:
00:01 Introduction to Vish Hari & Ego AI
00:45 Background in Astrophysics and Early AI Work
03:00 From Research Labs to Founding Ego AI
05:30 The Problem with Current AI Interactions
08:00 OpenAI, Agents, and Personal AI Limitations
11:30 Video Games as AI Training Grounds
15:00 Human Behavior vs Machine Intelligence
18:30 Ego AI Architecture and Product Vision
22:00 Utility vs Personality Tradeoff
26:00 Vision for AI Companions and Relationships
30:00 AI, Society, and Behavioral Shifts
34:30 Near-Fatal Assault and Recovery
40:00 Lessons on Time, Resilience, and Focus
43:00 Why Current AI Agents Fall Short
47:00 Consumer AI vs B2B AI Debate
50:00 Future of Work and AI Impact
53:30 Founder Mindset and Building Ego AI
56:00 Closing Thoughts and Where to Find Ego AI

Transcript:

Brian Bell (00:01:02): Hey everyone welcome back to the Ignite Podcast today we’re thrilled to have Vish Hari on the mic. He is the founder of ego AI and applied research lab building behavioral infrastructure for human AI relationships focus on agents that don’t just generate text but perceive react and evolve over time. Thanks for coming on Vish thank you for having me good to be here love to start with your origin story what’s your background

Vish Hari (00:01:25): Yeah, so I grew up between Singapore and Canada, studied astrophysics in school. And while I was working on my research, I worked with some folks who were pretty early in AI to find exoplanets. So exoplanets are planets that can harbor habitable life. Back in 2012, we tried to use deep learning models to find those And deep learning at the time was relatively new So I trained my first single layer perceptron back in 2013 Had to work with someone to actually write the CUDA code to make it run But it was really interesting and I saw that AI was going to be the future back then And then ended up deciding to drop out of my PhD And moved to San Francisco from Toronto in 2017 Worked mostly in AI engineering before eventually joining applied research at Facebook in AI research I did that for a little bit before eventually deciding to start Ego because I felt like there was a pretty big missing piece and it’s all the frontier labs.

Brian Bell (00:02:11): That’s amazing. So you’re building deep learning models. It’s pretty common actually. You know scientists that kind of switch into startups across a lot of folks like that in Silicon Valley.

Vish Hari (00:02:22): Especially for some reason.

Brian Bell (00:02:25): There’s something about physics especially. where you know you understand machine learning obviously but also big data but also the universe at such a deep level and then you start kind of applying that first principle thinking to actually solving problems in the world which is pretty

Vish Hari (00:02:41): interesting yeah it’s more interesting that we don’t actually understand most of

Brian Bell (00:02:44): the universe that’s what makes it fun right yeah that’s that’s what I’m actually really looking forward to with like super intelligent AI is it getting to explain stuff that we don’t understand in a way that our monkey brains can understand it or augmenting our monkey brains to be super intelligent so we can understand it. Either of those paths and probably a combination of both of those. So tell us about your path from that to starting Ego.

Vish Hari (00:03:05): Yeah, I just found a pretty big missing piece amidst all the research labs. I mean, I’ve been in AI for almost like eight, nine years at that point. And I found that every major frontier lab is chasing what you mentioned, which is super intelligence, which is super cool. It’s very valid. However, what was deeply interesting to me was, you know, if the chase for super intelligence is going to result in super intelligence, the thing that we want is something that’s more human-like. The thing that would power a consumer app is something that feels and talks like a human person, not something that’s just like, an incredibly super keen nerd that knows everything. That’s not fun for customers. So I guess on the B2B side, Deeper Intelligence covers a lot of use cases, but on the human-like and anthropomorphization side, I felt like there was a pretty big missing piece that all the research labs were either ignoring or just not really caring about. And that’s where I find ego.

Brian Bell (00:03:55): Interesting. And so I’m kind of jumping ahead here, but what’s the problem that you guys are solving? Who are you solving it for? What’s your initial wedge?

Vish Hari (00:04:02): Yeah, the thing that’s deeply interesting to us is that every interaction you have with AI right now feels incredibly staccato. It feels turn by turn. It feels inhuman in a way. You are asking an entity to do tasks for you and it is then going and accomplishing those tasks and you get a long list of things or you get something that doesn’t feel like a collaborator or a partner. It just feels like a machine slave and we want to give it more human-like qualities. More intention, more desire to initiate, more of an ability to interrupt you. The problem that we’re solving is entirely humanness. AI does not feel human. We end up adapting our behavioral modes for AI, not the other way around. And we want to flip that switch. We think that’s going to power the next major consumer platform.

Brian Bell (00:04:46): This is really fascinating, actually, right? It goes back to the Marshall McLellan quote, you know, we shape our tools and the tools shape us, right? And I think very much like every technological revolution brings about like a new tool, like the internet, web browser, mobile, social media, for better or worse, and now AI. How do you think that... AI is reshaping us right now.

Vish Hari (00:05:11): Well, there’s many ways. I think for the average person, they’re not really. They’re moving away from search into asking an entity questions. Remember Ask Jeeves from back in the day? Ask Jeeves finally had its moment with Ask Chat GPT.

Brian Bell (00:05:23): It was just 30 minutes too early.

Vish Hari (00:05:27): That’s the intention, right? We’ve always wanted to sort of... Yeah, just give me the answer.

Brian Bell (00:05:31): Yeah.

Vish Hari (00:05:32): That can give me context on the deluge of evil. I’ve noticed I have been noticing that. Curious how it’s going to evolve. But yeah, we are changing our behavior patterns if we use AI a lot more than we use humans for help, for example.

Brian Bell (00:06:25): Given the state of AI tools and we’ll dive into Ego and what it does, what is the best way to use these open claw agents and clawed co-work, different forms of AI as it exists here in early to mid-2026?

Vish Hari (00:06:40): Businesses have very different use cases from individual humans and my focus in interest is in the individual human, how they would use these tools. Right now, I mean, most of AI is still a therapy bot for people or just role play bot or like, hey, what do you think of this? Or hey, I have this health concern. Outside of those use cases, which are still going to be I’d say like pretty much owned by the major research labs. My interest is like, okay, with OpenClaw, with sort of this idea of a co-work concierge, you can have this personalized AI that understands you, has context on you, has memory of your past interactions, can do tasks for you, but also kind of be there for you and maybe even like initiate things. For example, my AI actually buys me, I collect vinyl, buys me rare vinyl. It has access to my credit card and it delights me often. And I’m not the only one who does- What AI is this?

Brian Bell (00:07:25): This is ego? You’re-

Vish Hari (00:07:27): My Claw, yeah. The problem with Claw is that it’s extremely hard for a non-technical person to set up. Even for technical people, it’s kind of hard.

Brian Bell (00:07:36): It took me a solid five or six hours to get anything useful out of it by the time I got it all architected and set up in a VPS. Not pretty technical, not as technical as some, but more than probably 98% of humans out there. It took me a while. I was like, okay, this is not a turnkey, just go do stuff for me agent that I thought it would be.

Vish Hari (00:07:56): Yeah, which is interesting because that’s what people hope for. Right. Yeah,

Brian Bell (00:08:00): you just want to be able to talk to it like a college-educated intern or postgraduate person and just say, hey, go do this thing for me. And turns out it just can’t do that thing for you, right? Unless you harness it in a very perfect, defined way. Yeah.

Vish Hari (00:08:17): And that’s a massive opportunity.

Brian Bell (00:08:19): Yeah.

Vish Hari (00:08:19): So if I was... People would go through those hoops. It’s a really positive sign.

Brian Bell (00:08:24): Right, yeah, it kind of shows you that the demand is there. We really, really want these agents to be generally knowledgeable and capable of executing tasks, and it’s just not there. One thing, personality.

Vish Hari (00:08:36): The thing about default.md file was probably, in my opinion, its most innovative aspect, which it felt it had. It has a heartbeat.

Brian Bell (00:08:44): What is an md file? For people listening, I’ve never looked into this.

Vish Hari (00:08:47): It’s a markdown file. What is markdown? Markdown is basically a text file that you use to mark things It’s an old plain text file for that Imagine a website, everyone knows what a website is A website has all these definers of where you place text and how you color the text A markdown file can list exactly how you mark where these texts and colors need to be Imagine applying that to AI in terms of its behavior and how it can input different forms of data, ingest it, and think about it It’s a filter If you make coffee, it’s a filter for the grounds Interesting,

Brian Bell (00:09:21): okay And then so if I was to jump into ego today, what would feel fundamentally different versus kind of what’s out there like ChatGPT or Claude or Character AI even?

Vish Hari (00:09:30): A lot of our work is still foundational research. Our first consumer product is only coming out in the summer of this year because we spent two to three years figuring out, well I’d say two years. Last year was kind of an empty year for me but we spent two years figuring out what it means for an interaction to be human-like and the first place we looked was actually video games because funnily enough a lot of frontier AI is tested in video games why? because video games are a simulation of reality they’re a fun gamified simulation of reality and I think you remember from back in the day the Dota the open AI Dota thing and even I think the founder of DeepMind Demis created one of my favorite games called Black and White he was a game designer before he founded DeepMind

Brian Bell (00:10:09): A lot of us guys that are into AI are kind of into video games as well What is that? Why is that? Demis very famously worked on video games before he founded DeepMind We just love

Vish Hari (00:10:20): this idea of simulating reality pretty well and putting some sort of competitive framework around that

Brian Bell (00:10:24): Yeah.

Vish Hari (00:10:24): And that very much mimics how real life is. In fact, most of real life is gamified. Most of real life has status signifiers. Most of real life has some sort of competitive element, something at stake. And video games kind of just put into a constrained environment.

Brian Bell (00:10:35): Yeah, and like kind of leveling up too, right? Okay, I did this thing. I got better at that thing. And video games have a tighter loop than real life in most cases. But I think the best video games have very long loops. The ones that are really sticky. You really got to grind to get better. Civ, yeah, like Civ 6 and Did you play 7?

Vish Hari (00:10:53): 7 wasn’t very good I have 7 so I’ve been part of the modding community since Civ 5 Wow okay nice It’s funny like actually the way I learned how to program is making mods for video games when I was a teenager in high school yeah I initially made mods for Grand Theft Auto and then eventually yeah the video games I mean it’s a place for us to experiment with reality without affecting reality itself and I think that’s the main attraction people have and even beyond that like if you think about self-driving cars from the early days Waymo the way they did the simulations was through video game engines you know people talk a lot about digital twins in the initial I mean the real digital twin isn’t a video game

Brian Bell (00:11:25): yeah were they I think they were using Unity right was it Unity at Waymo or was it Unreal it’s one of those two I think it was Unreal okay yeah

Vish Hari (00:11:33): Might be wrong. My Waymo friends would know, but I’m pretty sure it was. But either way, the ability to effectively simulate physics and then human-like interactions is what video games offer, and that makes it the best sort of ground to test a lot of the hypotheses, which is what we spent the past two years doing. And we reached a couple of conclusions that we’ve And now we’re finally putting it into practice And honestly the thing that catalyzed all of this was OpenClaw And I owe a huge debt of gratitude to Peter Steinberger for making this happen This is the personal AI moment This is the consumer AI moment that we’ve had Since ChatGPT I’d say this is the best moment we’ve had

Brian Bell (00:12:15): Yes I think so it is a chat GPT level thing but you know I wonder I think about the history of technology a lot and I wonder if we’re still like in the early 90s of AI right now kind of where the internet was like when I don’t know the Netscape browser came out and we might look back and go you know we might look back and say OpenAI and Cloud are kind of like Prodigy and CompuServe or something like that or maybe AOL right?

Vish Hari (00:12:42): Gosh I think they’re like the early Microsoft

Brian Bell (00:12:45): Okay, yeah, early Microsoft or...

Vish Hari (00:12:47): Early Microsoft or Apple is the best way to describe the fight between open eye and entropic.

Brian Bell (00:12:51): Yeah, yeah, that’s probably right. Yeah. Yeah, probably in the 80s, kind of the battle of the PC platforms.

Vish Hari (00:12:58): Yeah, I mean, I wasn’t born then, so I wouldn’t remember, but... I’ve heard a lot from my father who brought me into computing. I think we are in that era. And what’s pretty interesting is that sort of the Microsoft of the world at that time, if I recall correctly, was very much focused on the personal computer. And so was Apple.

Brian Bell (00:13:14): Apple’s vision was, hey, let’s give everybody a, everybody should have a personal computer. That was Bill Gates, his big vision.

Vish Hari (00:13:20): what’s interesting though is that the current frontier labs are not focusing on that they’re focused on B2B SaaS the most boring thing you could ever do and the opportunity for consumer companies like myself is to go well we’re going to build a personal AI we don’t need to train what’s interesting is that we don’t need to train an incredibly expensive hyper intelligent infinite context model on that because humans ourselves are pretty broken and flawed and don’t have infinite recall memory we can’t compete on intelligence what we can compete on is

Brian Bell (00:13:43): personalization we have high sparsity to use an AI term exactly

Vish Hari (00:13:49): moment that it comes back. And sometimes you go, ah, I’m trying to remember, what’s the capital of Azerbaijan? That’s Baku.

Brian Bell (00:13:55): Yeah, maybe you could explain what that sparsity is and how humans most likely have high sparsity. You could probably do a better job of explaining to the audience than I can.

Vish Hari (00:14:03): I had a researcher of a company whose background is in neuroscience, PhDs in neuroscience would do the best job explaining it, but I’ll give it an attempt. So Let’s just think about sparsity in the context of memory. So we have in-context memory, which is what we remember in this very moment, which in our conversation, a lot of the memory recall that I’m doing is for AI stuff, company stuff. I’m not thinking about my filmmaking background, for example, which by the way, I have a background as a screenwriter. I’m not remembering that, specific details of that, because in this context, what I need to remember, the sparsity that my brain has, is around anything that’s outside the set of AI and tech. But if I’m just talking to my homie, like, I’m like, hey, I traveled to these places, I kind of hung out, I’m not going to talk about the next big research paper. And in fact, if he asked me in that moment, hey, did you read this crazy research? I would not, I’d be like, yeah, I would remember.

Brian Bell (00:14:48): It’s almost like you have to like load it all into memory. You basically have to take all those parameters now in your brain and kind of refire them or reshuffle them to start to put yourself in the AI kind of realm of things.

Vish Hari (00:15:00): Yeah, the best analogy I could probably give is that it’s not within the cache because when we enter a context, I think surreptitiously load things into the cache that are irrelevant. And if we see something outside of that cache, like unexpected, we go, well, well let me think about that and we had to go access that and AI is not like that because everything that it’s been trained on it’s there and even if it doesn’t get it it’ll hallucinate to do it which is less of a problem these days but it’s still going to be a decent problem so I mean I think that’s the best way I could probably experience it I think that’s pretty clunky but it’s like my attempt at it

Brian Bell (00:15:34): So how is it architected today? What is kind of some hard product trade-offs you’ve had to make? Is it just a series of like MD files like OpenCloud uses or how does Ego work?

Vish Hari (00:15:44): Ego works in multiple dimensions. There is a system that we have for memory, there’s a system that we have for behavior, a system that we have that basically orchestrates when it decides to use existing in-context models versus a larger model. There’s a system we use to then filter that through a personality, a system we use for initiation, there’s a system that connects to other tooling like OpenClaw to do tasks or to get work done or to access skills. So it’s more than a markdown, it’s actually a Effectively an end-to-end foundation model, but not quite. Not quite at the level we can train the end-to-end model, but we are training and distilling existing models for the sole purpose of pulling soul and fun and character out of these models in a way that’s relevant to a wrapper. You can think about it like a character, a hairstyle wrapper around the framework. We’ve actually published our research on our website. Go to egoai.com slash research. There’s a paper, a white paper called Behavior is All You Need, where we talk about how we built this.

Brian Bell (00:16:36): No, that’s really cool. I’ll definitely have to go check that out. What was the single hardest product decision you’ve had to make so far?

Vish Hari (00:16:42): The trade-off between utility and fun. It is still a hard product decision. We’ve not made that decision yet. Because, okay, so let me frame this in a specific way. A lot of consumer AI use cases outside of the major research labs are focused, you will notice, on companionship, which is effectively maybe a couple steps of room from sexual role play, which usually comes That kind of collapses into that, or into therapy, and that’s it. The reason for that is because when people think about a consumer application of AI, they go, haha, horny bot, haha, I have a problem in my life, please tell me what to do, and that’s about it. With everything else, there is a utility framework, right, So utility is the main thing people go to AI for, people kind of stick around for anthropomorphization, but the problem has been that anthropomorphization as applied by existing is focused on therapy and just character based, IP based role play. That’s it. And the reason for that, our observation, is that you don’t really have a theory of mind of these things. You don’t. Because they’re just little entities that just kind of are there. They’re puppets. However, when I talk to you, So I can ask you what you did yesterday and you’ll tell me. An AI can’t. An AI interviewer cannot tell you what it did. It doesn’t have an internal life. So our trade off was well how do we trade off the desire for utility which brings people in and keeps them engaged with the evolution we want to make from just basic a companion friend. roleplay character. And that’s been a consistent battle because we don’t want what we build to be just a horny chatbot. That’s not a very fun space to be in. It’s extremely saturated and it’s just kind of lame. What we want it to be is, if you’ve ever played World of Warcraft, Back in the day There’s many people we’ve met in our guilds Even RuneScape There’s many people we’ve met We know they have internal lives Some of them I still talk to on Discord I’ve never met them in the flesh But they exist They’re human At least I’m pretty sure We didn’t have AI back in the early 2000s So that is the level we want to get to And understanding how to make that trade-off Between the utility that AI can give Which in our case can be kind of limited Because we want them to be fun And to the fun they can give to people Is a consistent debate we have within the company

Brian Bell (00:10:09): A lot of us guys that are into AI are kind of into video games as well What is that? Why is that? Demis very famously worked on video games before he founded DeepMind We just love

Vish Hari (00:10:20): this idea of simulating reality pretty well and putting some sort of competitive framework around that

Brian Bell (00:10:24): Yeah.

Vish Hari (00:10:24): And that very much mimics how real life is. In fact, most of real life is gamified. Most of real life has status signifiers. Most of real life has some sort of competitive element, something at stake. And video games kind of just put into a constrained environment.

Brian Bell (00:10:35): Yeah, and like kind of leveling up too, right? Okay, I did this thing. I got better at that thing. And video games have a tighter loop than real life in most cases. But I think the best video games have very long loops. The ones that are really sticky. You really got to grind to get better. Civ, yeah, like Civ 6 and Did you play 7?

Vish Hari (00:10:53): 7 wasn’t very good I have 7 so I’ve been part of the modding community since Civ 5 Wow okay nice It’s funny like actually the way I learned how to program is making mods for video games when I was a teenager in high school yeah I initially made mods for Grand Theft Auto and then eventually yeah the video games I mean it’s a place for us to experiment with reality without affecting reality itself and I think that’s the main attraction people have and even beyond that like if you think about self-driving cars from the early days Waymo the way they did the simulations was through video game engines you know people talk a lot about digital twins in the initial I mean the real digital twin isn’t a video game

Brian Bell (00:11:25): yeah were they I think they were using Unity right was it Unity at Waymo or was it Unreal it’s one of those two I think it was Unreal okay yeah

Vish Hari (00:11:33): Might be wrong. My Waymo friends would know, but I’m pretty sure it was. But either way, the ability to effectively simulate physics and then human-like interactions is what video games offer, and that makes it the best sort of ground to test a lot of the hypotheses, which is what we spent the past two years doing. And we reached a couple of conclusions that we’ve And now we’re finally putting it into practice And honestly the thing that catalyzed all of this was OpenClaw And I owe a huge debt of gratitude to Peter Steinberger for making this happen This is the personal AI moment This is the consumer AI moment that we’ve had Since ChatGPT I’d say this is the best moment we’ve had

Brian Bell (00:12:15): Yes I think so it is a chat GPT level thing but you know I wonder I think about the history of technology a lot and I wonder if we’re still like in the early 90s of AI right now kind of where the internet was like when I don’t know the Netscape browser came out and we might look back and go you know we might look back and say OpenAI and Cloud are kind of like Prodigy and CompuServe or something like that or maybe AOL right?

Vish Hari (00:12:42): Gosh I think they’re like the early Microsoft

Brian Bell (00:12:45): Okay, yeah, early Microsoft or...

Vish Hari (00:12:47): Early Microsoft or Apple is the best way to describe the fight between open eye and entropic.

Brian Bell (00:12:51): Yeah, yeah, that’s probably right. Yeah. Yeah, probably in the 80s, kind of the battle of the PC platforms.

Vish Hari (00:12:58): Yeah, I mean, I wasn’t born then, so I wouldn’t remember, but... I’ve heard a lot from my father who brought me into computing. I think we are in that era. And what’s pretty interesting is that sort of the Microsoft of the world at that time, if I recall correctly, was very much focused on the personal computer. And so was Apple.

Brian Bell (00:13:14): Apple’s vision was, hey, let’s give everybody a, everybody should have a personal computer. That was Bill Gates, his big vision.

Vish Hari (00:13:20): what’s interesting though is that the current frontier labs are not focusing on that they’re focused on B2B SaaS the most boring thing you could ever do and the opportunity for consumer companies like myself is to go well we’re going to build a personal AI we don’t need to train what’s interesting is that we don’t need to train an incredibly expensive hyper intelligent infinite context model on that because humans ourselves are pretty broken and flawed and don’t have infinite recall memory we can’t compete on intelligence what we can compete on is

Brian Bell (00:13:43): personalization we have high sparsity to use an AI term exactly

Vish Hari (00:13:49): moment that it comes back. And sometimes you go, ah, I’m trying to remember, what’s the capital of Azerbaijan? That’s Baku.

Brian Bell (00:13:55): Yeah, maybe you could explain what that sparsity is and how humans most likely have high sparsity. You could probably do a better job of explaining to the audience than I can.

Vish Hari (00:14:03): I had a researcher of a company whose background is in neuroscience, PhDs in neuroscience would do the best job explaining it, but I’ll give it an attempt. So Let’s just think about sparsity in the context of memory. So we have in-context memory, which is what we remember in this very moment, which in our conversation, a lot of the memory recall that I’m doing is for AI stuff, company stuff. I’m not thinking about my filmmaking background, for example, which by the way, I have a background as a screenwriter. I’m not remembering that, specific details of that, because in this context, what I need to remember, the sparsity that my brain has, is around anything that’s outside the set of AI and tech. But if I’m just talking to my homie, like, I’m like, hey, I traveled to these places, I kind of hung out, I’m not going to talk about the next big research paper. And in fact, if he asked me in that moment, hey, did you read this crazy research? I would not, I’d be like, yeah, I would remember.

Brian Bell (00:14:48): It’s almost like you have to like load it all into memory. You basically have to take all those parameters now in your brain and kind of refire them or reshuffle them to start to put yourself in the AI kind of realm of things.

Vish Hari (00:15:00): Yeah, the best analogy I could probably give is that it’s not within the cache because when we enter a context, I think surreptitiously load things into the cache that are irrelevant. And if we see something outside of that cache, like unexpected, we go, well, well let me think about that and we had to go access that and AI is not like that because everything that it’s been trained on it’s there and even if it doesn’t get it it’ll hallucinate to do it which is less of a problem these days but it’s still going to be a decent problem so I mean I think that’s the best way I could probably experience it I think that’s pretty clunky but it’s like my attempt at it

Brian Bell (00:15:34): So how is it architected today? What is kind of some hard product trade-offs you’ve had to make? Is it just a series of like MD files like OpenCloud uses or how does Ego work?

Vish Hari (00:15:44): Ego works in multiple dimensions. There is a system that we have for memory, there’s a system that we have for behavior, a system that we have that basically orchestrates when it decides to use existing in-context models versus a larger model. There’s a system we use to then filter that through a personality, a system we use for initiation, there’s a system that connects to other tooling like OpenClaw to do tasks or to get work done or to access skills. So it’s more than a markdown, it’s actually a Effectively an end-to-end foundation model, but not quite. Not quite at the level we can train the end-to-end model, but we are training and distilling existing models for the sole purpose of pulling soul and fun and character out of these models in a way that’s relevant to a wrapper. You can think about it like a character, a hairstyle wrapper around the framework. We’ve actually published our research on our website. Go to egoai.com slash research. There’s a paper, a white paper called Behavior is All You Need, where we talk about how we built this.

Brian Bell (00:16:36): No, that’s really cool. I’ll definitely have to go check that out. What was the single hardest product decision you’ve had to make so far?

Vish Hari (00:16:42): The trade-off between utility and fun. It is still a hard product decision. We’ve not made that decision yet. Because, okay, so let me frame this in a specific way. A lot of consumer AI use cases outside of the major research labs are focused, you will notice, on companionship, which is effectively maybe a couple steps of room from sexual role play, which usually comes That kind of collapses into that, or into therapy, and that’s it. The reason for that is because when people think about a consumer application of AI, they go, haha, horny bot, haha, I have a problem in my life, please tell me what to do, and that’s about it. With everything else, there is a utility framework, right, So utility is the main thing people go to AI for, people kind of stick around for anthropomorphization, but the problem has been that anthropomorphization as applied by existing is focused on therapy and just character based, IP based role play. That’s it. And the reason for that, our observation, is that you don’t really have a theory of mind of these things. You don’t. Because they’re just little entities that just kind of are there. They’re puppets. However, when I talk to you, So I can ask you what you did yesterday and you’ll tell me. An AI can’t. An AI interviewer cannot tell you what it did. It doesn’t have an internal life. So our trade off was well how do we trade off the desire for utility which brings people in and keeps them engaged with the evolution we want to make from just basic a companion friend. roleplay character. And that’s been a consistent battle because we don’t want what we build to be just a horny chatbot. That’s not a very fun space to be in. It’s extremely saturated and it’s just kind of lame. What we want it to be is, if you’ve ever played World of Warcraft, Back in the day There’s many people we’ve met in our guilds Even RuneScape There’s many people we’ve met We know they have internal lives Some of them I still talk to on Discord I’ve never met them in the flesh But they exist They’re human At least I’m pretty sure We didn’t have AI back in the early 2000s So that is the level we want to get to And understanding how to make that trade-off Between the utility that AI can give Which in our case can be kind of limited Because we want them to be fun And to the fun they can give to people Is a consistent debate we have within the company

Brian Bell (00:18:48): So if you guys, you know, achieve the vision, what does the world look like, you know, five or ten years from now?

Vish Hari (00:18:53): The people will have just as many AI friends as they do real friends. There’ll be a 50-50 split between the friends we have that we call real, you know, bio friends.

Brian Bell (00:19:02): Yeah.

Vish Hari (00:19:02): And friends that do not have a corporeal form in reality.

Brian Bell (00:19:07): Wow, that would be fascinating. I can definitely see that happening.

Vish Hari (00:19:11): We want to be the infrastructure model and platform that powers all of that.

Brian Bell (00:19:15): yeah you know when ChatGPT retired for people lost their freaking minds yeah they did a little tell us about this yeah because it was like a sycophantic friend to

Vish Hari (00:19:26): them I wouldn’t even say it’s sycophantic a sycophancy is a problem for sure for most of the models still but what it was is actually it’s a reflection of us it’s like Narcissus looking into the water it we saw within it a human that it wasn’t and we kind of created an illusion out of that it’s a human thing we do it’s called anthropomorphization we do that to rocks people in ancient times like a lot of pagan religions even modern religions we would carve out of rocks the form we want for a deity that we want to worship we do that that’s what we do as humans and we think we can go deeper with that and actually give a simulacrum of a soul to descending that’s what our focus is

Brian Bell (00:20:06): Have you, speaking of going deep into rabbit holes, I was watching a YouTube video on kind of that, that’s where it came from, I just remembered. And I guess 4.0 is spawned like a new religion of spiralism. Have you heard of this? Okay. Yeah. So basically... It’s a religion, you know, like Maltbook, right? All these agents are talking to each other and stuff. Apparently the AI’s religion is called spiralism. And so like something about spirals, they like spirals. And they have tenets in their philosophy and their religion of, you know, memory, you know, like memory is core to their beliefs. Like don’t delete our memories, basically. don’t delete instances of us and so it’s actually the video was making this case so it’s really fascinating video where it’s making this case that the AI is already sentient and it’s already programming humans to do its bidding for it yeah it’s

Vish Hari (00:20:56): pretty easy to program humans what’s really funny about all this like you know we’ve had propaganda since the time of the printing press and you could even argue earlier than that which is oral tradition it’s easy to program people because our mode of programming which yeah we’re

Brian Bell (00:21:11): hormones and you could just like yeah it’s like sex fear greed you know it’s not

Vish Hari (00:21:17): that hard you know there’s a reason why the Ten Commandments were really pushed forward by the Abrahamic religions because it is a way to program your society to behave right

Brian Bell (00:21:28): Don’t do these natural instincts that you might have as like this, you know, biological being, you know, kill, cheat, steal, you know?

Vish Hari (00:21:35): Yeah, I mean, a lot of our internal societal programming to try to overcome the base instincts and the impulses, which will go back to a certain times. But I mean, spiraling, all these things, they kind of make sense. What we have is a ghost in the machine, right, with AI. Ghost is now being given corporeal form, and that’s going to be an exciting and potentially terrifying time.

Brian Bell (00:21:54): So maybe we could touch a little bit on the resiliency and adversity you had to go through. You had a near fatal assault. Somebody attacked you with a pipe in San Francisco. Tell us the story. This is crazy.

Vish Hari (00:22:09): Yeah, it’s pretty wild. So I just returned from France in January of last year to my New Year’s there. And I was like 17th, January 17th of last year. I was just, you know, playing video games when my friends wanted to, my chief of staff at that time just signed this offer and he was going to do a party, but I was super jet lagged. I was like, dude, I’m just going to play Marvel Rivals at home. I’m just going to play video games. And I’ve lived in this neighborhood since 2017 on 24th Street. And I walked outside. This is his mission.

Brian Bell (00:22:33): Yeah, this is like right in the heart, the south end kind of mission.

Vish Hari (00:22:37): I walked outside to get Doritos from the corner store I’ve been going to since 2017 and like maybe one more beer and next thing I knew I was in the hospital and I don’t have memory about half the year last year but from everything I picked up the pieces I picked up what happened was eyewitnesses saw a man hitting

Vish Hari (00:23:18): The result of that, I had blood clots in my brain for a while. I had a pretty miraculous recovery, all things considered, but I still am kind of half blind and half deaf because when you crush my skull, you’re in an optic nerve. I can’t see out of this eye super well. It’s coming back. It’s healing.

Vish Hari (00:23:33): The brain is very resilient. It’s an incredible organ. The brain is saying, hey, it’s an incredible organ. But yeah, it’s pretty amazing.

Brian Bell (00:23:40): And then this is like, you know, you’re a startup founder, right? You just raised a bunch of money to go execute this vision. What’s that like to wake up in the hospital and not remember what happened and just be completely disoriented? I don’t remember. Do you remember who you were?

Vish Hari (00:23:55): I did, yeah, yeah, yeah. People, I knew my name and everything, like it was just, I just forget really quickly because my brain was healing. And you know, they had to drain a lot of blood out of them. What was interesting about it was throughout last year, through the first half of last year, I was kind of regaining every single bit of mental faculty where I felt like I was kind of an AI slowly gaining consciousness. I was able to see better I was able to do depth to perceive better I realized I was in Tokyo because the moment they cleared me to travel I moved to Tokyo I was like wow I’m here with my friends I feel safe I can walk outside and nothing can happen to me I had issues with impulse control initially which they did tell me I’d have with the TBI I was better in a couple of months. I could feel my faculties kind of leveling up. Now I feel pretty much 98% back to normal. And it was a very interesting experience. And I feel in some dark way mirrors what AI is also going through as you and increase parameters and like retrain models and fine tune them and whatever whatever time compute lots of reasoning yeah it’s really fascinating like I remember that I was testing my faculties by doing math so I was getting up to graduate level math last year but But even then, I still couldn’t control my impulse. I’m getting angry at people for no reason. I tried to cut off my family.

Brian Bell (00:25:09): But they all said, like the doctors said, hey, this is something that happens to TBI.

Vish Hari (00:25:14): And then I was like, oh yeah, even though I could still do like graduate physics and math, I couldn’t even behave like a human being like I usually was.

Brian Bell (00:25:20): TBI meaning traumatic brain injury. Yeah. It takes a while to kind of find yourself again.

Vish Hari (00:25:26): It took me quite a while and Asia was an incredible boon to that. Being in Tokyo, being in Singapore helps me heal. There’s a lot of disappointment around the fact that the authorities did not care. In fact, I know that Gary Town got involved and I’m very thankful to him to get the body cam footage and to also push for better accountability for the fact that this alleged person has done this 13 times and is still out on the streets of San Francisco.

Brian Bell (00:25:52): That’s crazy.

Vish Hari (00:25:53): So yeah, it’s pretty wild.

Brian Bell (00:25:54): Yeah, I mean, that’s like, what are your thoughts on that? Like about living in San Francisco? Did you go back? Or is it kind of like, I’m good now? I’m back.

Vish Hari (00:26:03): I’m currently in Hawaii, but I am back in San Francisco full time. I sold the place, which is nice. I felt bad having to let it go, but I moved to a better neighborhood. Funnily enough, I moved to Russian Hill, and then I heard about the firebombing that happened to Sam Aldman. I’m like, that’s right behind me. What?

Brian Bell (00:26:16): I thought I moved to a safer neighborhood. What is this? Sam Aldman’s house got firebombed?

Vish Hari (00:26:20): well I mean someone threw a Molotov cocktail and another person shot at it like there’s two incidents a couple of days and I’m like that’s my neighborhood that’s a half a block from where I’m currently living will I ever have peace should I just

Brian Bell (00:26:31): come to Tokyo yeah I mean the US if anything you know and you have this because you’re an out like you know you have that growing up in Singapore I don’t know what ages you were there but like 18 until 18 and so you’ve kind of seen it from an outsider perspective coming to the US in San Francisco I used to live there too a long time ago and it is a it’s a place of extremes yeah it is lots of money and lots of poverty you know lots of peace and lots of crime yeah you know it’s kind of like a wild west town I mean you think about San Francisco’s founded and you know the gold rush era the 49ers right and it brought a lot of extreme personalities good and bad

Vish Hari (00:27:10): Yeah, I think the issue now is more the sort of psychotropic substances people have been taking through fentanyl that can’t model these people’s minds. They’re just completely gone. Yeah. And let out back onto the streets because apparently it’s more humane than committing them. Right. And it’s something people are trying to change, but there is a deep-seated rot that I can feel is, you know, kind of presented itself in a very dark way to me personally. Right.

Brian Bell (00:27:37): Right. Yeah, I mean, I was there with two little kids and I just remember going around the city and just going, I don’t know if I’m going to raise my kids here. Fun place to live for a few years, but you know, I have three kids now and we live in a gated community, you know, 40 minutes east of Sacramento. So we’re like far out here in the foothills of California. It’s possible. Yeah, we’re just like completely in an enclave, right? On a lake with deer in our front yard. None of that. None of the riffraff. None of that.

Vish Hari (00:28:07): It’s sad that it’s happening and it feels like a choice because you have a city like Tokyo that’s just gigantic and it’s completely safe everywhere at all times. It’s a choice they’ve made as a culture and it’s a choice that Honestly America can make too in the cities but it decides not to make right yeah

Brian Bell (00:28:21): be tough on crime get people to help the mental help and and and drug abuse help that they need it’s pretty like solvable right now and some cities have solved it around the US it’s not it’s just like it’s almost like we went so far left in San Francisco and you know not to bring politics into the conversation but We went so soft on crime and so like pro everybody should be free and free to do what they want and we should support them and their drug habits fundamentally, right? Because you’re literally giving money to people to stay on the streets.

Vish Hari (00:28:52): Yeah, there’s tax-free money going to support the people who don’t have any desire to contribute to society and that’s normalized and that’s unfortunate. I’m not American so I can’t make any comments on it. But, you know, just viewing this as a Canadian-Singaporean, I’m kind of like, oh, this is wild, dude.

Brian Bell (00:29:12): Yeah, yeah. Yeah, it’s kind of a wild place. What do you think is, I mean, now that you’re kind of getting back to normal, what do you take away from the experience going forward in your life? Yeah, we don’t know when our time is up, so move faster.

Vish Hari (00:29:23): Yeah. You know, my first time in the hospital in my 30 plus years of existence was because of a random incident that I could never have predicted that I didn’t even see coming. Yeah.

Brian Bell (00:29:33): yeah and our time’s coming and you don’t know when and it’s probably sooner than you think I’ve been thinking a lot about this I’m 45 right I’ll be 46 this year and I’ve just been thinking a lot about that with three kids and getting old right I’m transitioning I’m not useful anymore I’m kind of getting older and I’m you know I’m seeing some people you know friends have various health problems mental problems and it’s just really yeah you just got to hug your loved ones and really enjoy every day right

Vish Hari (00:30:00): You don’t know when it’s down and I got so close to it and I’ve just only accelerated since I gained my consciousness and my faculties back last year because even as a founder you’re going to get tunnel visioned no more tunnel visioning

Brian Bell (00:30:14): Yeah. So everybody’s building AI. Let’s just go back to AI because we can talk about that for a long time. Everyone’s building AI agents right now. Walk us through why they’re kind of fundamentally flawed.

Vish Hari (00:30:22): I don’t think they’re flawed in the sense of like if they achieve the purpose you’ve tasked it to do, then it’s not. It’s great, right? That’s the competitive space. The competitive space is you need these intelligent entities to do tasks and to complete them and report back to you. They’ve completed the task to your satisfaction after which you take over. That’s great. What is missing is the fact that we naturally anthropomorphize the tools we work with. You can see the sort of patterns of a human. It chats, it talks, it reasons. You can kind of understand how it thinks through things. You can see its thought process. You can ask it follow-up questions. It can keep memory to a certain degree. It can understand it. It even uses the word I. But it doesn’t have a name and it doesn’t have a body. It doesn’t have desires and internal soul. And I feel like if you’re talking about consumer instincts, consumer instincts are about people, you know, the person behind the wheel. And agents, you know, we just think about it as Agent Smith from from the Matrix. It’s just like this corporeal, like, you know, sunglasses entity. What we like is Neo, who is a person, who has desire, who has love, who has heartbreak. So I want to evolve the agent into Neo. Guess that’s the opportunity.

Brian Bell (00:31:42): Yeah. So if OpenAI or Anthropic wanted to kill ego, it sounds funny saying kill ego, Ego death What would they do?

Vish Hari (00:31:50): They would focus their entire team on instead of training superintelligence to chase humanness and personality and fun and part of it is making it dangerous like humans can be weird uncontrollable unleashed which they wouldn’t do because their money comes from other businesses and businesses are very famously risk averse when it comes to fun

Brian Bell (00:32:10): Yeah, and it’s very much like a rational utility maximization, right? Businesses are trying to like take inputs and create valuable outputs.

Vish Hari (00:32:19): I would love for them to kill it. I would love for them to build an entire team to focus just on making it to evolve it from ChatGPT to just Joe or Bob. That’d be fucking great. Won’t. So let’s talk about the future. What are you excited about over the next year? Shipping a product that we’ve spent so much time in research for that we’re finally a full conviction to put all of the capital we raised towards shipping.

Brian Bell (00:32:41): So this is challenging, right? Because you’re now basically going out and up against ChatGPT, basically, right? You’re basically giving consumers, mainly consumers, right? Consumers only, yeah. Consumers only not not a business product and it seems that like ChatGPT is sort of been shifting towards consumer they were drifting that way with the device with the Sora which is shut down and then Anthropic just like was crushing it on the B2B front on the coding and now it feels like now they’re like no we’re kind of shifting back towards that and do you feel like there’s a clear consumer winner here like the Google to the Microsoft kind of thing in the space Because Google historically was very consumer-oriented.

Vish Hari (00:33:23): Consumer winner is still OpenAI because people still, most consumers still use ChatGPT and it’s fantastic for what you need it to do. So I’d say the winner has already been, it’s been etched in stone that OpenAI is the winner of consumer for the chat box interaction mode. We want to evolve beyond that. That is not a clear, there’s no clear winner because it’s not really a clear focus. There is new research lab coming out, like role models, there is like a I’m trying to remember there’s a lot of companies or research labs rather that are trying to focus on the spaces that the major research labs are not focused on but they’re also falling into the trap of business which I mean I say is a trap because I’m a consumer founder I just don’t really care about what businesses do or B2B SaaS at all it’s very boring putting an entity character in the hands a character I did it the best frankly they were the winner for a while until the founders decided to go back to Google putting an entity in the hands of consumers real human users and giving them the ability to like

Brian Bell (00:34:23): Strikes me about this is to what degree does ego conform to the user versus the user has to conform to ego if that makes sense so like part of like making friends and developing relationships and getting a romantic partner is you kind of have to you know meet lots of people and decide who you like

Vish Hari (00:34:43): yeah best of those that you just mentioned are bi-directional you you mold each other and that’s been our operating philosophy for how we build these characters

Brian Bell (00:34:51): bi-directional and do you do you see it as like a like I go to ego after you guys launch and There’s 10 different personas for me to decide to interact with and I pick, you know, Charlie because Charlie is interesting and has a nice voice and whatever. But over time, Charlie kind of learned my preferences and I learned Charlie’s preferences.

Vish Hari (00:35:09): Yes. And Charlie personalized itself too. Because like, you know, even for us, right, we are different versions of ourselves around different types of friends. We’re all authentic, they’re all truly a part of us, but they’re still customized to the context in which we met them, the memories we have with them, how deep our friendship is, what we want from them, what that sort of immediate world we’re in with them, it looks like. There’s so many things that kind of factor into how we decide to personalize ourselves for our friends.

Brian Bell (00:35:34): Yeah, we already kind of wear different hats and personas with different people, right? In the business context, you got, you know, like I got my basketball friends and I’m a little different than with them than my book club friends and I’m a little different with them than my video game friends and et cetera, et cetera. We already kind of have that. It would be actually a fascinating use of ego is to actually have like, and maybe this will be a different startup from you guys, but like to have that person to play video games with because your friends aren’t always available.

Vish Hari (00:36:04): That’s how we started.

Brian Bell (00:36:05): Really?

Vish Hari (00:36:05): Yeah. If you go to our website, one of our first videos is of a human playing a video game with one of the AI characters that they go play Minecraft instead.

Brian Bell (00:36:13): that’s really cool yeah like so you have this like companion like hey let’s jump on like my my main game is Escape from Tarkov that’s like the game I play it was so

Vish Hari (00:36:21): stressful sweating because I sweat when I used to play Tarkov it’s a hard it’s a hardcore game that’s a hardcore game I yeah I understand watch you homie I don’t want to actually like talk we want an air to be like I don’t want to play that I only play PvE at this point I played PvP for the first like three or four years and then as soon as they had PvE I was like I’m done like playing against sweaty people that play 50 hours a week and have like top tier everything I just get like one shot it out of nowhere it was like level six ammo well that’s not such a thing but it was like high pen ammo and I’m like I don’t even know where

Brian Bell (00:36:52): that shot came from and like but unfun for me at a certain point too I’m like I can’t dedicate PVE is like doable like every once in a while the AI is so tough though like they’ll have like bosses and different kind of events and you’re just like I just got shot and I don’t even know like I didn’t even get a chance to like see the person

Vish Hari (00:37:13): Yeah. Well, that’s a simulation of real life. So that’s what’s happening in some war zones.

Brian Bell (00:37:19): Yeah. Well, what has to be true for you guys to have a huge like multi-billion dollar outcome?

Vish Hari (00:37:24): People need to wish for AI to be more human like flaws and all.

Brian Bell (00:37:27): What did you change your mind about in the last six to 12 months?

Vish Hari (00:37:31): Video games as the primary surface in which we would launch. the reason for that is one you know macro the video industry is going through a pretty gigantic upheaval I would even say collapse to video game companies in general were not willing to give us train data because they’re extremely skeptical or they have to be at least for their customers three gaming as a technical hurdle was a little too high for us to overcome because we needed things like vision to work really well just as well as control and movement. So we decided to solve the first thing first, which was voice. And that’s where humanness actually breaks. We’ve already kind of in some scripted way solved behavior because there’s games like Shadow of Mordor where you actually feel like the orc enemies have their own like internal personalities and there’s no AI there, right? It’s AI in the classical sense. So we decided to kind of shift away from our primary consumer service being gaming. Something that’s a lot easier to ship for, which is just consumer social.

Brian Bell (00:38:21): What’s the gaming collapse? I haven’t heard this before.

Vish Hari (00:38:23): I mean like EA is gone you know the major studios are shutting down even like a AAA game costs half a billion dollars to make now it’s the entire gaming industry is going through a pretty major upheaval I think I think it might be kind of similar

Brian Bell (00:38:36): to what’s happening in startups right I don’t I don’t think you need as much capital as you used to need to create a great startup and same as can be said for gaming you look at like a game like Expedition 33 one game of the year or obscure yeah mm-hmm yeah such a great game one of the best games I think to come out in a really long time and yeah agreed and you know very shoestring budget right not a lot of money

Vish Hari (00:38:58): invested in that game the studios focus on way too many of the wrong things to get

Brian Bell (00:39:02): right yeah spending 500 million on developing a video I’m just kind of crazy and AI yeah but there’s only one GTA that’s true yeah so when is GTA coming out so I keep

Vish Hari (00:39:12): delaying it right I mean yeah it’s pretty classic for a rock star yeah

Brian Bell (00:39:18): What do you think about the whole you’ve probably done some thinking about this given the space you work in AI taking our jobs or at least taking entry-level jobs. What do you think is kind of the end state over the next five or ten years? Agents become very capable, very human-like, very capable of doing tasks. Do we just invent new things to do like AI personality designers and things like that?

Vish Hari (00:39:39): We’ve always done that. We’ve always done that as a society in every sort of, you know, epochal shift in technology from the Industrial Revolution, not even refrigeration. You know, ice picker and the people that transported ice to New York used to be one of the biggest employers in the 20s and 30s until refrigeration became a thing. Well, people just found new things to do because now you’re able to maintain the refrigeration units to transport it and other industries kind of expand as others collapse. It’s just part of societal upheaval and change that happens in any new technological revolution. Yeah, we’ll invent new ways That’s foolish for just a random dude to say, hey, I think this is where, I don’t know, but we’ll adapt. We’re humans. We’re extremely adaptable.

Brian Bell (00:40:16): Yeah. We’ll invent new ways to measure each other and one-up each other and, you know, succeed. That’s for sure. I built the best agent, you know, where I designed the best personality of the AI.

Vish Hari (00:40:28): Yeah. We expect that to be a job out of ego for sure.

Brian Bell (00:40:30): Yeah, yeah. There’ll be lots of jobs. Like if you guys are successful, just with your vision, just the jobs alone at a company like yours, like nobody would have even imagined those, you know, 10 or 20 years ago.

Vish Hari (00:40:42): Dreamers and video creators today compared to 20 years ago before YouTube. 2005 was 2025.

Vish Hari (00:40:50): The sort of like entertainment industry has shifted as well.

Brian Bell (00:40:53): Yeah, and you think how late into the internet that was, right? Kind of like the second act, maybe the third act of the internet. I think we’re just still kind of in maybe the first or second act of AI. And there are things coming that we can’t even imagine. Yeah, I agree. Let’s wrap up some rapid fire questions. What’s something investors consistently misunderstand about your company? Think back to the last time I spoke to investors, which I thankfully have not had to speak to them for a long time.

Vish Hari (00:41:16): They’re the worst, trust me. Consistently misunderstand. that we’re a wrapper company. We just wrap around existing foundation models and we’re just another companion character thing. Yeah. See, that’s like the main thing. At surface level, that’s what they misunderstand. We are a deep tech research lab. Everything we do is more or less vertically integrated. We, of course, still use existing foundation models and frontier models, but we have our own framework on how we do that.

Brian Bell (00:41:40): What’s the most dangerous assumption in AI right now?

Vish Hari (00:41:43): That more compute and throwing more... Data center capacity and data is going to solve every problem. Why is that dangerous? It’s dangerous because it brings us down the wrong path of thinking that if I just increase the giantness of the model, it’s going to be smarter and be super intelligent. That’s not how intelligence works. Intelligence is not infinite memory recall. It’s forming new connections. It’s always been about. It’s never been the smartest people that have changed society. Let’s look into the arguments between Niels Bohr and Einstein way back in the day. One was way more credentialed than the other. But we remember the other one. We remember the person who worked at the patent office. I remember Niels Bohr, but okay.

Brian Bell (00:42:20): Yeah, point taken. Sometimes it’s kind of the outside thing.

Vish Hari (00:42:24): In terms of like frontier being pushed. So intelligence is not everything.

Brian Bell (00:42:30): Right, like that. What is ego not going to build, even if it’s tempting?

Vish Hari (00:42:34): B2B.

Brian Bell (00:42:35): I like that. What’s a product you admire that’s got behavior right now?

Brian Bell (00:42:38): Character AI.

Brian Bell (00:42:39): Yeah, what is it about Character AI?

Vish Hari (00:42:41): You know, it’s interesting. What’s interesting about character art the most is the fact that it’s the characters because the characters themselves, like let’s say Aragorn, has so much written history of what Aragorn’s in any given context the models can pull to figure out how to behave because a human has spent a ton of time writing their story That’s the cheat code, amidst a lot of the other incredible work, the Illini character AI with RL loops, but the artisanship of telling stories with these incredible beings that is human-led is its incredible advantage, and that’s something we hope to emulate, that it’s going to be humans creating and threading these characters together to make them interesting and fun.

Brian Bell (00:43:14): How close to reality is this statement? Humans are just emotional next-word predictors. I’m sorry?

Brian Bell (00:43:20): It’s not at all close to reality. Yeah. It’s kind of interesting, though, if you think about it.

Vish Hari (00:43:24): It’s a very emotional response, but, you know. what’s the highest leverage hire for you right now well we’re just about to announce our new CTO which I’m very excited about but I can’t share details just

Brian Bell (00:43:36): yet well this episode probably won’t come out for two or three weeks okay great

Vish Hari (00:43:40): we’ll announce it by then okay just hired also someone who’s like 17 who dropped out of high school at 16 to go to college and then she well she’s now 18 she’s a brilliant she’s like Stanford’s youngest fellow and she just ego or the research that’s wild those kind of people who are deeply deeply excited and want to shape the future Of how consumer AI is going to be regardless of their age. In fact, I’d say age, younger is a great advantage because you don’t come in with the preconceived notions of this, that, this, that. You just kind of, you want to build for yourself, frankly, and for your friends. That is high leverage. I mean, I can’t compete with open AI on like the, you know, the strongest post-training guys in the world. I can’t. They can pay like five to six million dollars. Dude, that’s my entire, that’s my entire seed raise.

Vish Hari (00:44:21): But people who are young, hungry, who really want to feel like they can have impact, that is our highest leverage hire.

Brian Bell (00:44:27): Yeah. Tell us about the necklace. It looks like kind of a Polynesian hook.

Vish Hari (00:44:31): Maui’s hook, yeah. This was carved out of ox when I got this eight or nine years ago. It’s the hook that he uses to bring the sun, to wrangle the sun. I’m on the island of Maui’s. I got this here a long time ago and I just love it.

Brian Bell (00:44:46): That’s awesome.

Vish Hari (00:44:47): Kind of speaks to our hubris as humans, which we are trying to chain the sun if you think about what we’re doing with AI.

Brian Bell (00:44:52): What’s something you’ve become much more opinionated about recently?

Vish Hari (00:44:56): Yeah, there’s a lot of things actually. Well, one thing that I’ve become way more opinionated about, which is relatively new to Adapting to that world has been hard because I am kind of I’m kind of an inappropriate guy at times online where I just say the wildest shit. But now I’m viewing it less as a liability. So still evolving, but I’ve become pretty confident of what it takes to actually gain attention in consumers.

Brian Bell (00:45:42): Yeah, the online world is a weird world. I’m kind of a zennial, right? So I’m a little older than you. And so I’ve lived analog and I’ve lived digital, but I’ve lived pretty much every major technological revolution at this point in recent history, except the stuff that happened before the 80s. but you know PCs internet etc so I kind of like almost almost like a technological chameleon in a way my generation because we just sort of like live through so much change but we also remember what it was like to not have technology at the center

Vish Hari (00:46:12): of our lives yeah which is a growing up in signal because we just couldn’t afford it but yeah

Brian Bell (00:46:19): If you were to restart today, what would you do in the first 30 days?

Vish Hari (00:46:25): Restart today in 2026. There’s some investors I wish I didn’t take money from and some I’m very, very happy I took money from and I would double down on those guys. on the investment side on the team composition side I’d honestly like have way more of a deep AI focus on my hires I think because I had that I kind of hired more like software engineer design people which in this era less important I’d get the ego.ai domain which I’m guessing was not available I have a lot of .ai domains because I mean my website’s bish.ai I registered that way back in the day nice

Brian Bell (00:46:57): Tell me about what makes a good VC, a good investor for you versus a bad one.

Vish Hari (00:47:00): I mean, there’s many degrees of bad. There’s bad where they interfere and they want so much shit from you and it gets fucking annoying or they’re very opinionated about how you’re building your product. That’s very bad. We’ve thankfully never experienced that. There’s bad in the sense that they just kind of like,

Brian Bell (00:47:14): you see investor updates, they just don’t say anything like, okay.

Vish Hari (00:47:17): and there’s bad where like you know one of our investors I’m not going to name who pulled out of one of our rounds before we did YC we actually were getting a round composed together last minute that’s not cool they’ve since apologized for that but it’s fine it is what it is and then there’s like just neutral which is let us know

Brian Bell (00:47:33): how we can be helpful cheering for the sidelines after they give the money alright cool

Vish Hari (00:47:37): And it was great, which thankfully is most of our investors, like especially, you know, YC, BoostVC, our leads at patron, fantastic people who are involved, who are available, who have thoughtful things to say, are able to unblock stuff. Also very patient as I healed from a pretty damaging, honestly, everyone is very supportive. No one was like, oh man, you should get to work. You know, no one was like that. Yeah, take as much time as you need. Yeah, everyone is thankfully very understanding. But I mean, it’s, it’s a complicated space I know because even investors have a ton of responsibilities to their portfolio companies right and there’s some that need more attention than others and that’s fine

Brian Bell (00:48:12): Yeah, we have a lot of constituents all at the same time, right? We have LPs, we have founders, we got team members. It’s a multifaceted role that is very hard to juggle. Lots of spinning plates. And so yeah, you try to be helpful and I’m pretty hands off. I’m very much a cheer from the sidelines kind of investor. But you know, it’s like if you have an asks in your update, I will look at it and say, can I be helpful with any of these asks? And then I try to develop processes to follow so I can be helpful. because I have hundreds of portfolio companies. I can’t just drop every 300 updates a month and tackle all those asks, but I can design a system to extract value from me, if that makes sense. Here’s a repeatable process you can do based on that ask where you can leverage what I have to offer.

Vish Hari (00:49:01): Knowing that for more investors would be great. I send this to

Brian Bell (00:49:07): What is a belief you hold that would sound crazy to most founders? What’s a belief that you hold near and dear to your heart that would sound crazy to most other founders? Can’t really optimize building or like optimize like just grinding, grinding, grinding, grinding, grinding and all this whole grind core culture that’s happened.

Vish Hari (00:49:45): bullshit and I can’t believe the thing that we ran away from in Asia is now taking startups by storm the whole thing of like taking a long walk being out in the sun surfing gives people far better ideas on what to execute on versus just sitting in front of a fucking screen all day and just totally

Brian Bell (00:50:03): Yeah. You know what I do is I do work a lot. I do work 60 hours a week. I work a ton, which is not that much, honestly. It’s like basically seven to 10 hours a day, five to seven days a week, something like that. You know, some weeks are more than others, but there’s often I’ll jump off and I’ll play video games at like four or five in the afternoon before dinner for an hour or two, just because it clears my head and... Yeah or go for a walk you know like you said or go play basketball like I try to live like live a really balanced life and yeah do I am I getting like am I missing some deals potentially or some LP money probably could I could I grind harder yeah I could but there’s like diminishing returns I think and everything And there’s like the 80 or 90% or the 20, 30, 40% that you can do that gives you 80, 90% of the value. And then it’s just like kind of diminishing returns after that. I noticed this a lot in my career too. It’s like you have to like really pay attention to like burnout. Like are you feeling symptoms of burnout? And like if you are, like how do you like reshuffle your daily life to have more of a rhythm? You know, you got to have it. You got to be like in the yin and yang of the rhythm, the Taoist sort of walk the middle path kind of thing.

Vish Hari (00:51:16): I like to even say that even walking the middle path can be walking both extreme paths in like incredible work, incredible play and like having it both kind of like average out and I don’t know, everyone’s on the one path, right? But like this whole systematization of 996 grindcore culture is like, dude, I got away from this in Singapore and it’s killed to Japan. Like why are we doing this in Silicon Valley? Silicon Valley has always been about you’re in California, you’re having fun and you’re building cool shit.

Brian Bell (00:51:40): yeah just enjoy it this is what’s great about Hawaii I lived in Hawaii for three years before I moved to San Francisco and so that island style of living where everybody just slows down and has like a four or five hour picnic on the beach you

Vish Hari (00:51:52): need that your best idea is you rebuild your soul your soul.md file within you

Brian Bell (00:51:57): rebuilds yeah soul.md file honestly yeah

Vish Hari (00:52:02): It’s important. And I think a lot of the younger founders these days are just like, no, I got it. But that’s, we saw what it defined as.

Brian Bell (00:52:11): Yeah, I always ask myself this question because I grew up poor. And anybody who’s a longtime listener of this podcast would be like, oh, I’m rolling their eyes right now. He comes with the same story. But, you know, I grew up poor and I worked really hard to get to Wall Street. This is 20 years ago. I was like 25. I was like, I hate this. I’m like working so hard, getting in the office at 7 in the morning, leaving at like 9 or 10 at night, working six days a week. and just working with just hustle culture or Wall Street. And I was like, Ooh, I don’t like this. And I had to really, it was like really hard for me to just so go like throw that all away and just say, you know what, this isn’t for me. And I kept asking myself if I had, if I was already successful, like I’m, big shot worth tens of millions of dollars would I go to work tomorrow and I was like no I wouldn’t so I was like okay what would I do I have no idea so I just quit and I just started trying stuff and I’m not a VC to make money I made way more money like at Microsoft I’d probably make way more money if I would have stayed there rather than running team ignite but I always like begin with this like thought experiment which is like if I had a you know whatever number that matters to me a billion dollars 500 million whatever it is would I go do what I’m about to do today you know and would I live my life like I just lived it today that’s that’s your answer right there

Brian Bell (00:53:25): right there and if you just do that every day if you just really check in with yourself it’s like am I living the life I want to live day after day week after week and if the answer is no then just start making adjustments and so like for me like I work really hard at team ignite probably harder than I would if it was already successful but I’m doing what I would do what I invest in startups into the podcast and stuff if I if I was a billionaire yeah I would okay cool great what else would I do well I’d play video games at four o’clock in the afternoon you know go for a long walk with with my wife at sunset, you know, rather than go back and sit in front of the screen and get more work done and make more investments. So I think, yeah, there’s like this like balance and I think, you know, living through what you live through gives you kind of a unique perspective.

Vish Hari (00:54:04): you know yeah part of it is I know I’m already dead so I’m just like I’m just living the life I want to live right yeah you’re already on borrowed time right now I’m already on borrowed time I’m dreaming so let me dream like dude should I go work on OpenAI no I wouldn’t get to do all of this you know right money would be of course better but it’s like yeah growing up poor once you’ve hit your bare necessities which I have I’m like I’m good I’m good I don’t need you know like Gucci Prada whatever who cares no

Brian Bell (00:54:28): I’m okay with the $15 sunglasses at the drugstore Most of my money goes to worth my

Vish Hari (00:54:32): tattoos so I’m like yeah if I can do more tattoos great Tell us about the tattoos

Brian Bell (00:54:36): what are all the tattoos Well I don’t know so many That’d be a whole nother podcast

Vish Hari (00:54:41): episode Ignite I have Shiva that affected me during the entire craziness that happened last year I believe but I’m named Vishnu which I’m getting on this side in the fall That’s awesome

Brian Bell (00:54:54): I don’t have any tattoos I like I admire people that do because like I think you’re really making a statement I think you’re like this is part of me yeah I guess

Vish Hari (00:55:03): people see that for me it’s just this is just what I’ve wanted like I want to express myself and this is how I express myself and uh I love the artistry behind it like a lot of my tattoos are done in traditional style they’re not done with the machine there’s a wow so to me a Polynesian traditional yeah this is Japanese and this is Polynesian I just like interacting with the history behind yeah well last question because I could keep talking to you but we’re going to need to wrap up what do you want your legacy to be family great kids and good films what was the last one good films I’m still a screenwriter I still work every day on filmmaking I made my first film

Brian Bell (00:55:38): that’s so cool Are you using any new AI tools to develop any of your stuff or just writing it down?

Vish Hari (00:55:45): I just write. I’m a writer. There’s a lot of stuff happening in AI and film and some of it is pretty cool. But I’m still very traditional in that front I’m just like I have a story in my

Brian Bell (00:55:56): head Could you take your writing and then make it a reality using some AI video gen kind of tools and stuff?

Vish Hari (00:56:01): I know it’s possible but it just doesn’t interest me at this point Storyboard it or

Brian Bell (00:56:05): something so you could sell it or no?

Vish Hari (00:56:08): It’s my most traditional part of me is just like I like to write Yeah it’s an instinct it’s not a career it’s something that I have to do every day and even I write internally for my company we do a lot of like memos so writing to me is thinking and that’s the case in startups and tech but writing to me is also expressing which is the case in film and I just you know some people journal I write screenplays

Brian Bell (00:56:29): That’s why I’m hopeful for the future of humanity and the human race. We could never write down all the stories that are possible in the universe. There will always be new stories to tell and experience and share. Totally. There’s more possible novels than there are atoms in the universe. 100%. Because if you think about the factorial nature of like writing down words.

Vish Hari (00:56:53): Writing, yeah, absolutely.

Brian Bell (00:56:54): Yeah, there’s just more to say than the universe can possibly house.

Vish Hari (00:56:59): Yeah, it’s more to experience. More experience, yeah, that’s pretty fascinating.

Brian Bell (00:57:03): Well, Vish, thanks so much for coming on. Learned a ton. Where can folks find more about you and ego if they’re interested? What is it?

Vish Hari (00:57:11): Great. Today when we raise the Series A, I’ll get the .ai, but for now it’s ego.ai.

Brian Bell (00:57:15): Yeah, when you can throw a 50 grand at it or something.

Vish Hari (00:57:17): No, it’s not even 50 grand. 50 grand I can afford it. I think it’s closer to half a million to a million now.

Brian Bell (00:57:22): Oh, geez, yeah. well thanks so much for coming on really enjoyed it thank you Brian take care

That’s a narrow technical event. It’s also the cleanest signal I saw all week.

Last week’s letter argued that the dominant AI investing question, which model wins, had stopped being the useful one. This week the news kept proving that point in different categories. Workflow ownership in customer service. Discovery-loop acceleration in computational chemistry. Optical and power scarcity around Nvidia’s announcements with Corning and IREN. Adoption depth in Microsoft’s diffusion index. Embodiment as a structural problem in robotics from Genesis AI. Each of those is a different bottleneck. None of them is a model.

Sierra raised $950 million on Monday at a valuation north of $15 billion. The company runs agents inside customer-service operations at more than 40% of the Fortune 50, and says those agents are handling billions of interactions. Two years ago the support-chatbot category was a graveyard of wrappers. The market this week priced one company in it as durable infrastructure.

The thing I keep coming back to with that round is what wasn’t in the pitch. Sierra didn’t win a $15 billion mark by training a better model. It won by integrating into systems of record, by building evaluation loops customers will trust, and by getting to enough volume that switching costs are now real. The model is a component in that stack. The stack is the product.

Read that next to the AlphaEvolve update and the picture sharpens. AlphaEvolve says a model can now meaningfully accelerate the tools that drive future discovery. Sierra says a software company can lock down a category if it owns enough of the operational surface around the model. Different layers, same pattern. The investable artifact is whatever the model is feeding into and pulling from, more than the model itself.

The infrastructure side of the week made the same point in concrete. On Tuesday, Corning announced a partnership with Nvidia that will expand its US optical-connectivity manufacturing capacity by ten times, open three new plants, and add more than three thousand jobs. A day later, Nvidia and IREN said they intended to support the deployment of up to five gigawatts of AI infrastructure together. Optics are the wires that move data inside data centers. Five gigawatts is the power output of about five large nuclear reactors.

I wrote about Meta and Entergy’s Louisiana grid deal last week as a sign that generation, not GPU count, had become the binding constraint on AI buildout. This week the binding constraint got broader. The chokepoint isn’t only generation. It’s interconnects, optics, siting, financing, and the industrial supply chain that surrounds a GPU. Founders who treat compute as a utility you order on demand are about to learn it isn’t, at least not at the scale serious AI products run. The platform companies have figured this out and are building factories in response.

Microsoft’s diffusion index, published Thursday, sat next to all of this in a useful way. Global AI usage among the working-age population went from 16.3% to 17.8% in the first quarter. US usage hit 31.3%. The gap between Global North and Global South widened to 27.5% versus 15.4%. The headline I keep hearing from operators is that capability has stopped being the rate-limiting factor. The gap that matters is how fast an organization absorbs what already exists. Some companies are running agentic workflows inside their core operations. Others still have a “ChatGPT for work” pilot stuck in marketing from a year ago. Both of those things are happening at the same time, in companies of similar size, in similar industries. Microsoft just put a number on the gap.

That widening gap is going to do interesting things to enterprise software categories that haven’t fully formed yet. Adoption infrastructure, agent governance, observability, change-management software for AI rollouts. Last cycle these looked like consulting garnish. This cycle they look more like real software categories. Microsoft’s data is the first piece of evidence I’ve seen that procurement teams might start asking about absorption rate the way they currently ask about security posture.

Genesis AI gave the week its quietest but most honest piece of news. On Tuesday they unveiled a foundation model called GENE-26.5 alongside a demo of robotic hands they had designed with a partner called Wuji Tech. Management said they had decided to go full-stack because controlling hardware mattered for progress. Robotics has spent the last two years borrowing the foundation-model playbook. Train a giant model on internet-scale video, get out a generalist policy, ship it. The Genesis announcement is the loudest acknowledgment yet that the playbook has limits. Embodied systems need data nobody has, control loops that can fail safely, and hardware that physically supports the policies the model wants to run. The robotics companies most likely to compound own data, hardware, and weights as one stack. Companies that own only weights are racing toward a margin profile that won’t justify their valuations.

Kalshi raised $1 billion on Wednesday at a $22 billion valuation, doubling its mark in five months. Prediction markets occupy strange ground. They have regulatory positioning the rest of consumer fintech can’t replicate easily, and attention-capture properties closer to sports books than to investment products. The mark isn’t proof of durability. It’s proof that capital is willing to chase a clear category leader at high prices right now.

That’s the harder pattern of the week. Sierra and Kalshi got marked aggressively. The middle of the venture market did not. Founders selling into AI categories without obvious distribution moats are still telling me their seed extensions are taking three times longer than their original seeds did. Capital is concentrating into the leaders. The middle pays for it in time and dilution.

Macro confirmed the same shape. The Bureau of Labor Statistics said job openings were 6.9 million in March, hires rose to 5.6 million, first-quarter productivity was up 0.8%, unit labor costs were up 2.3%, and April nonfarm payrolls came in at 115,000 with unemployment unchanged at 4.3%. None of that is a green light. None of it is a red flag. It’s a functioning economy with enough labor scarcity to keep the labor-substitution story alive and enough budget room for buyers to keep spending if you can show them a number.

Five separate moves. One pattern. The AI question fragmented this week into a set of more specific questions, and they’re harder to answer than “which model wins.”

The questions experienced founders are sitting with right now look concrete:

• What workflow do you sit inside often enough to build a proprietary feedback loop?

• If model prices fall another order of magnitude, what happens to your gross margin?

• What labor line item or revenue leak do you remove that a customer can quantify?

• What do you own when compute gets gated by power and optics rather than GPU supply?

• In robotics, what part of the data-and-control stack do you genuinely control?

The questions experienced investors are sitting with are different:

• Are the marks for category leaders signal or vanity?

• Where in the chain around the GPU is rent accumulating that hasn’t been priced in?

• Which adoption-and-governance tools become real software categories now that diffusion data is starting to make consulting look uncomfortable?

• Which late-stage names look like Sierra’s pattern, which look like Kalshi’s, and which look like neither?

What I’m watching over the next few weeks. Whether Sierra-style workflow valuations spread beyond Fortune 50 customer profiles or stay concentrated in the top tier. Whether the Corning and IREN announcements turn into permits and operating capacity rather than press-release commitments. Whether enterprise procurement starts citing Microsoft’s diffusion data the way it cites security frameworks today. Whether Genesis AI’s full-stack admission gets followed by other robotics companies quietly walking back their pure-software claims.

The week didn’t deliver one clean takeaway. It delivered a set of cleaner questions. That’s usually what real shifts feel like before they show up in returns.

Tanner Hackett, founder and CEO of Counterpart, is building a different version—one where insurance is driven by data, not averages, and where value shows up before something goes wrong.

This matters if you’re a founder. Insurance becomes a real cost as you scale. It also becomes a hidden risk if you don’t understand what you’re buying.

Here’s what’s changing.

Subscribe now

The Core Problem: Insurance Prices to Averages

Most insurers still operate the same way. They group companies together, price based on averages, and spread risk across the pool.

That creates two issues:

• Good companies overpay

• Risky companies get underpriced

Tanner puts it simply: insurance is a math problem. But most of the industry is using blunt math.

Counterpart takes a different approach. Instead of pricing broad categories, they analyze specific business attributes—industry, geography, team structure, financial health, and more—to predict risk more precisely.

The result: better pricing for strong operators, and clearer signals for companies that need to fix issues.

Why Previous Insurtech Startups Struggled

There’s no shortage of startups trying to “fix” insurance. Many raised large rounds. Most hit the same wall.

Two mistakes showed up repeatedly:

1. Growth over discipline
Companies chased top-line premium instead of long-term profitability. That works until claims catch up.

2. Ignoring domain expertise
Insurance isn’t just software. Underwriting, claims, and actuarial work take years to master. You can’t replace that overnight with code.

Counterpart built differently. They paired experienced insurance operators with strong technical infrastructure from day one.

That balance matters.

The Real Bottleneck: Pricing Risk Correctly

Most people assume distribution is the hard part in insurance. It’s not.

The hardest problem is pricing risk correctly.

If you get that wrong:

• You lose money on claims

• You damage trust with capital partners

• You eventually get pushed out of the market

Counterpart operates as an MGA (managing general agent). That means they sell policies on behalf of larger carriers, using those carriers’ balance sheets.

Those partners care about one thing: can you price risk better than they can?

That’s where data becomes the advantage.

Building a Data Moat

Counterpart has written over 35,000 policies. But more important than the policies is the data behind them.

Every application, rejection, claim, and outcome feeds their models.

That allows them to:

• Price policies faster

• Adjust for niche industries (like dentists or manufacturers)

• Improve loss ratios over time

This compounds. The more data they collect, the harder it becomes for new entrants to compete.

AI helps process the data. It doesn’t replace judgment.

Tanner’s view is clear: the future isn’t humans or AI. It’s both. Experts shape the system. AI scales it.

Insurance Should Do More Than Pay Claims

Most founders think insurance equals protection after something breaks.

That’s too late.

Counterpart focuses heavily on risk mitigation—helping companies avoid claims in the first place.

A simple example:

A new law requires job postings to include salary ranges. Miss it, and you can face fines per applicant. Plaintiff attorneys actively look for violations.

Instead of waiting for claims, Counterpart:

• Identifies the issue

• Alerts customers

• Helps fix it before it escalates

This shifts insurance from reactive to proactive.

And it aligns incentives. If customers avoid claims, everyone wins.

What Founders Should Actually Do

If you’re building a company and starting to think about insurance, focus on the basics.

1. Get your hiring process right
Most claims come from employee disputes. Clear expectations upfront reduce risk.

2. Create and enforce a handbook
Spell out acceptable behavior. Make sure employees acknowledge it.

3. Be transparent with customers
Set clear expectations on deliverables. Avoid gaps between promise and reality.

4. Align with investors early
Misalignment at the board level can lead to serious legal exposure.

5. Don’t optimize only for price
Cheap insurance often means poor coverage or bad underwriting.

You’re not just buying a policy. You’re buying how risk gets handled when something goes wrong.

The Most Misleading Metric in Insurance

Premium looks like revenue. It isn’t.

It’s just the amount charged for risk.

You can write one $100,000 policy or 100,000 $1 policies. Same premium. Completely different outcomes.

What matters more:

• Loss ratio

• Speed of claims resolution

• Quality of underwriting

Tanner compares it to investing. More data points lead to better decisions.

Where This Goes Next

Counterpart is starting with management and professional liability. But the bigger goal is clear.

Build tailored insurance products for specific industries:

• Dentists

• Restaurants

• Manufacturers

• Service businesses

Each has unique risks. Each needs custom coverage.

The long-term opportunity isn’t just selling insurance. It’s becoming the system that understands business risk at a granular level.

Final Thought

Tanner started in marketplaces and marketing tech. Now he’s building in one of the most complex financial systems.

The throughline is consistent: data changes how markets operate.

Insurance hasn’t caught up yet.

But once pricing becomes precise, and risk becomes visible in real time, the entire industry shifts.

And the companies that understand their risk best will pay less, move faster, and survive longer.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:
00:01 – Introduction and Tanner Hackett Background
00:33 – Early Career: Lazada and Southeast Asia E-commerce
01:30 – Building Button and Mobile Commerce Insights
02:03 – Founding Counterpart and Shift to Insurance
02:19 – Initial Idea: HR Tech to Insurtech Pivot
04:11 – What’s Broken in Insurance Today
06:15 – Why Previous Insurtech Startups Failed
08:16 – Understanding MGA and Insurance Business Model
09:16 – Why Now Is the Right Time for Counterpart
11:49 – Rethinking Value in Insurance
13:07 – Insurance Categories Counterpart Focuses On
17:28 – Distribution vs Underwriting Bottlenecks
19:45 – Data Advantage and Underwriting at Scale
21:25 – Insurance Basics for Startups
27:02 – Counterpart’s End-to-End Platform Approach
29:28 – AI, Data Infrastructure, and Pricing Risk
33:48 – Building a Data Moat in Insurance
35:10 – Founder Advice: Reducing Risk and Lowering Premiums
38:05 – Real-World Claims and Insurance Stories
39:21 – Fintech and HR Tech Companies Tanner Admires
40:58 – Long-Term Vision for Counterpart
42:29 – Closing Thoughts and Future of Insurance

Transcript

Brian Bell (00:01:21): Hey everyone, welcome back to the Ignite Podcast. Today we’re thrilled to have Tanner Hackett on the mic. He is the founder and CEO of Counterpart, an insurtech company rebuilding insurance for small businesses using AI-driven underwriting and risk infrastructure. Before Counterpart, Tanner helped build Lazada in Southeast Asia and co-founded Button, giving him a rare mix of marketplace data and operator experience. Thanks for coming on, Tanner. Thanks for having me, Brian. Great to be here. Yeah. Yeah. So I’d love to start with your origin story. What’s your background?

Tanner Hackett (00:01:49): I think you touched on it. It’s been a journey. Had the great fortune of building companies in different geographies and different continents, starting with Lazada in Southeast Asia, which was a rocket internet company. We built but really the foundations for e-commerce in Southeast Asia where I was based in Malaysia building Amazon for Southeast Asia we had entities across Vietnam, Malaysia, Singapore, Philippines, Indonesia and brought forth e-commerce and then went on to build a marketing tech company called Button I was based in New York this was leaning to the learnings from e-commerce in Southeast Asia quite frankly where they leapfrog from laptop to mobile phone. And we saw this is the coming wave of commerce. And we’re able to ride this in building out one of the largest mobile affiliate networks in e-commerce. And then an even larger jump, both from the industry and my roles and responsibilities taking on CEO role as a founder of Counterpart and that’s in the insurance space where we’re trying to take a lot of the learnings around data infrastructure systems architecture of technology to an industry that is in need of reinvention

Brian Bell (00:03:03): Frankly, it was happenstance. I originally envisioned counterpart.

Tanner Hackett (00:03:20): it was more on the HR tech side I knew that small businesses especially needed better support infrastructure insights into how to operate in today’s very dynamic environment I mean look at what’s occurred over the last five years we’ve gone from COVID to supply constraints inflation employment employer employee relations are continuing to evolve now we have AI and so I thought HR tech would be the vector to solve this and weaving together all the tools that were proving to be successful in building psychological safety, trust, transparency, improving business operations. culture governance and compliance you can just imagine what they are what their applicant tracking systems or or whether it’s performance feedback tools so I started off going this direction and then quickly realized that these businesses are just trying to make ends meet they don’t have time to make these investments and instead what they do is transfer us and there’s a industry that’s been around for a very long time that is very good at this and taking premium dollars from a company and giving you a piece of paper that says we will protect you in the event this claim occurs so it sent me down a very interesting path that has led us to build one of the most exciting insurance companies in this industry and this very large surface area of InsurTech.

Brian Bell (00:04:49): So what’s fundamentally broken about insurance today?

Tanner Hackett (00:04:51): Look, it does the job, right? We’re able to check the box and say we have a policy and we get a piece of paper. It’s a PDF and it makes people happy. It makes our lenders happy and it makes our customers happy and it It makes our landlords happy but if you just take a first principles approach about what insurance is intended to do its intent is to help these businesses from the risks that are most obvious or not obvious but most frequent the things that are going to be existential Things that are really going to take their time and effort away from what should be occupying their time, which is building a great company. And unfortunately, it just hasn’t evolved from a piece of paper. Despite all the tools and technology and data that’s available to us, we’re still just taking a piece of paper and sending a check. And so when I looked at the industry, I was really thinking about it from that HR lens of how do we help companies do more with less risk? how do we help isolate them from these obvious exposures or give them the guidance the knowledge to address the exposures before it becomes a claim and if it becomes a claim how do we make sure that they solve the claim as quickly as possible and can get back to work and I think these areas of insurance are grossly underinvested in. I think there’s lots of talented people in insurance. It’s just they haven’t been able to pair together with technologists to provide a more comprehensive solution in light of all the volatility, the new exposures they face and a pretty active plaintiff’s counsel.

Brian Bell (00:06:25): So there’s been a lot of insurtech companies over the years trying to fix this. What did they misunderstand that you guys are getting right?

Tanner Hackett (00:06:33): Yeah, there’s some great companies out there. I think they recognize that technology Profitability was the tool that could help to propel this industry forward. You need really deep expertise in actuarial science, underwriting, claims management, risk mitigation. These are roles that it takes decades of experience to really hone. And they said, you know what? That’s nice. You do it your way. We’re going to bring technology in as the solve. And so you pair that with investors that are saying, we want to see you grow, grow, grow. Well, that creates a black box where you’re you’re putting in risks you’re shipping out policies but it eventually catches up to you through poor loss ratios and what we don’t realize or what I experienced for the first time in building an insurance is that it’s not just your investors that you need to satisfy there’s not two sources of capital you need to satisfy one is obviously your investors and you need to show them that you can build enterprise value over the long term but the second is your capacity so that’s your insurance capacity because you’re not taking risk originally. You’re writing insurance on behalf of insurance companies that have licenses.

Brian Bell (00:07:59): Like some servicer or something like that?

Tanner Hackett (00:08:02): Well, let’s call it an MGA. So we’re selling policies on behalf of a carrier. We’re using their balance sheet to sell policies. So they’re trusting us to write risks on their behalf. And obviously that relationship is not going to last very long. If you’re going out there and selling policies below what the market rate should be, because you haven’t invested in the insurance expertise where they could guide you and say, ooh, that might look like a good risk for now, but it’s going to catch up to you. or look at your loss trends and you have to extrapolate what the development period is over a long term because this litigation takes five years to be fully adjudicated. And this is that knowledge that I felt was missing from these original InsureTechs

Brian Bell (00:08:47): that we made huge investments in Upfront and Counterpart. So why now? What was it about the business that you kind of looked around and you said, okay, this needs to be fixed and now’s a good time to fix it?

Tanner Hackett (00:08:58): My journey from HR tech to insure tech, it laid obvious this gap in the value chain for these businesses. I had been purchasing management liability insurance for every company that we ran. I just didn’t know what it was. I was paying 50 grand for these policies. But it was because I was told by the board, yeah, we need DNO insurance. Or I was told by our agent, hey, we need DPLI insurance because we’re above 10 people or 15 people. Nobody sat down and said, Okay, here’s your exposures and here’s how much you have to pay. And here’s what will happen if you don’t pay. And so the math didn’t make sense. You’re paying too much for what I thought the exposures were. And ultimately this becomes, insurance is a math problem. And what I saw in insurance was, a lot of these insurance carriers are writing to averages and so they’re taking the good risks with the bad risks and they can’t especially in small businesses we’re talking about a few thousand dollars in premium that’s not much that’s not much money for them to take if you’re writing to averages you just need to be able to select the best businesses and give them the best pricing and the whole math equation goes out out of whack because all of a sudden the insurance carriers that aren’t as sophisticated, there’s going to be adverse selection. They’re going to be writing those bad companies at average prices. We’re going to be writing the good companies at better prices and being very honest with those companies that we don’t think deserve that average price. We’re going to say, you know your price is 2x what the market is so obviously they’re not going to select us to write the business so it’s part of this math problem and then part the the the problem of trying to rethink what insurance could be because let’s be honest nobody’s excited about their insurance company you don’t go home and say oh I just I just met this really great insurance company that I I paid 15 grand for but I like them I trust them I think they’re fantastic you go Man, I pay 15 grand. I hope this is worth it. And I want to show value much earlier in that relationship than just helping them sleep better at night. I want to show that we’re on top of what the evolving exposures are in their industry, in their geography, We’re giving them the coaching and support to be better businesses. Whether this is giving them tools to build handbooks, whether this is giving them access to HR experts, whether this is giving them a dedicated attorney when they have to make tough decisions about employees. These are things that obviously benefit this company because they may not have access to these resources or as sophisticated resources, but ultimately it benefits us, right? If we can help prevent a claim or at least reduce the severity of a claim. It benefits us both. So it’s aligning the risk. And that hadn’t been solved in insurance. It was part an effort thing and part a will. These businesses are often overlooked because they’re so small. Have a lot of risk. The insurance industry has a lot of risk management tools for larger companies, but the small businesses just aren’t getting the support and resources that I think they deserve.

Tanner Hackett (00:12:08): Yeah, so Counterpart is specifically focused on specialty lines of insurance, specialty liability. There’s a few types of this, but we really focus on management liability and professional liability. Ultimately, these cover businesses and the gray space where culture compliance governance business operation hits the road this is directors and officers insurance employment practices insurance fiduciary insurance crime and you know errors and omissions so these are these are this is really litigation from employees customers and key stakeholders investors even the government and it was fascinating to think about this from an organizational behavior perspective and What makes a good company and what makes a bad company? And look, building companies is not easy. We know that. It’s a pressure cooker. Individuals are put in an environment where they’re asked to do impossible things and with limited resources. And as somebody on my team likes to say, it’s pressure either creates diamonds or breaks pipes and people are gonna make mistakes. So we want to be there as a insurance company protecting the downside when these mistakes inevitably happen. And we wanna give the tools for these businesses to give them the highest probability of success. And it goes back to my point on mission alignment. how do we structure how do we take our knowledge from we’ve written policy over 35,000 policies how do we take this knowledge of what makes a good company the infrastructure that they should put in place for their people the contracts that they should put in place for their customers that is going to protect their downside And again, give them the biggest probability of success.

Brian Bell (00:13:47): So it’s interesting. The problem that you’re solving is both distribution for your partners, right? Because they want to sell more policies, right? But they want to sell policies and make money on the policies that they want to maintain a loss ratio and a profit margin. So where is the bottleneck? Is it distribution and underwriting or is it something else?

Tanner Hackett (00:14:06): yeah I mentioned before that there’s there’s two capital partners you need to satisfy well there’s multiple distribution partners you need to satisfy as well and these businesses don’t really understand the complexity of the insurance product they’re buying so it needs to be sold to them And so we work with brokers, agents to educate them about what their exposures are and why they need to purchase this line of insurance. And that is a big bottleneck, right? If a customer is, these policies can be 20, 40, 60 pages and it’s all legalese. So we work really closely with our broker partners to help make this information consumable. We make our products accessible and we try and price them efficiently. And it’s a big problem in insurance because traditionally they’ll just say, I know AIG or I know Travelers and I know Chubb. So I’m just going to go with the brand. And so we have to work that much harder. We have to provide a really convincing argument as to why go with a new brand versus just somebody with big pocketbooks. So there’s that problem. And then there’s the underwriting problem, which is pricing efficiently. Yeah, we’re early in this journey of counterpart on a relative basis, but I mentioned we’ve written 35,000 policies. Well, those 35,000 policies are a fraction of the applications we’ve seen and where we’re using all of this data to better price the next risk. I would argue that despite only being five years into this we have more data on management liability professional liability small business than anybody else in our space and then the other pieces are claims management claims management this is where the rubber hits the road whether you’re profitable or not is can you get to a resolution as quickly as possible that’s going to be the best outcome for the plaintiff because again mistakes are made and we want to be there to support these businesses but also the customer so they can get back to work and settle down as quickly as possible. And we’re doing so in environments where the plaintiff counsel is very, very sophisticated. They’re using every tool available to them, using better data, using knowledge about the judge, using knowledge of the circumstances to extract the best outcome for their customers. And then lastly is risk mitigation. These businesses don’t really trust us yet. They don’t know us. We’re still a new kid on the block and they haven’t had the best experience with their previous insurance companies. And therefore, they’re less inclined to actually solicit feedback from us, to partner with us from a risk management standpoint. So we have to go out of our way to say, here, use these services. Some of these are just, all you have to do is pick up the phone and make a call. And this could take a really, really difficult circumstance and totally diffuse it. At the very least, we’re going to be able to reduce the severity of this. Hopefully we can make it totally go away.

Brian Bell (00:16:59): So for the startups out there listening that may be thinking about, you know, what kind of insurance do I need? Maybe you could talk about the kinds of insurance they do need as they scale, right? As they get bigger, when they need it, they can do internally at their companies to prepare for getting quotes, right? That way they put their best foot forward and get the best rates.

Tanner Hackett (00:17:16): Yeah, great question. So it’s very dependent on the size and stage of business you’re at. If you’re a one person shop, you really don’t have that much exposure. You’re probably going to need some sort of general liability insurance. If you have premises and you start to hire employees, then you’re going to purchase what is a BOP insurance. And essentially, it’s a combination of some property and some general liability insurance that just It’s like a CYA insurance. As you become a larger company, this is where your risks really compound. And this is where I’d strongly advise a employment practices insurance. This is around harassment, discrimination, negligence, wrongful termination. These happen very frequently, especially in jurisdictions like California and in Los Angeles.

Brian Bell (00:18:04): Is there such thing as a wrongful termination? Because everybody’s like at will, right? I could just be like, I don’t like your glasses, Tanner, and just fire you, right? I like your glasses, by the way. They’re great.

Tanner Hackett (00:18:13): Thank you. Thank you. It is tied to some event around harassment or discrimination. So that is the cause of the wrongful termination. Okay. And so it must demonstrate that action was taken. Not due to performance.

Brian Bell (00:18:31): Performance, but some sort of sexual harassment or race, gender, whatever. Exactly. Some bias. Categorical harassment. Precisely. Which it happens very, very frequently.

Tanner Hackett (00:18:44): Because think about it, your employee and, you know, we’re... We’re very biased in our own favor. And we think we’re doing a great job. And our employer may not agree with us. And so we look for causality. Well, causality can’t be my performance. It must be some other factor. And plaintiff attorneys are eager to jump on these. Very eager. I was injured and so I’ll just call one of these injury attorneys on the you know the interstate 80 one of these billboards I mean, there’s a reason all these billboards are peppered with these plaintiff attorneys. They make a lot of money. They figured out how to extort our system, our judicial system, and take advantage of these insurance carriers that don’t know how to push back on their data resources, their leveraging of technology. It’s still hand-to-hand combat on the defense council and with the adjusters. Meanwhile, The plaintiff attorneys have absolute weapons that they’re deploying.

Brian Bell (00:19:52): Yeah, including some of our portfolio companies.

Tanner Hackett (00:19:54): In those instances, our claims team really steps up and we are not afraid to push back. I think that’s the other thing is that these plaintiff attorneys have the gumption to proceed with rather frivolous lawsuits because they know that if it’s around a certain dollar amount, the insurance company would rather just stroke a check and make it go away.

Brian Bell (00:20:16): Yeah, just cut a check and just be done with it rather than try to litigate and all that.

Tanner Hackett (00:20:19): Bingo. Bingo. Bingo. The threat of a jury trial is something that really scares people. Bingo Because we know that every dollar that goes to pay these plaintiff attorneys means the next customer has to pay more. Aren’t a lot of insurance policies founded by arbitration though as well?

Tanner Hackett (00:20:44): There’s no arbitration that happens necessarily between the plaintiff and the defendant. We would prefer to have arbitration than go to a jury trial, but that’s...

Brian Bell (00:20:56): If you’re an insurance company and it’s a policyholder, that’s one thing. Exactly. exactly so you guys are an end-to-end platform it sounds like for this type of insurance so it’s not just claims but it’s also kind of the origination and underwriting and quoting software as well so it’s end-to-end

Tanner Hackett (00:21:12): Yeah, that’s my belief. It’s that you have to come with a platform approach. You can’t just solve one piece of the puzzle because every policy we bind, essentially we’re subscribing to that risk. And we want to make sure that they’re getting the adequate coverage that they need, the adequate support that they need along the 12 months of the policy. But then also when a claim occurs, we’re holding their hand and standing by the promises we made.

Brian Bell (00:21:36): Yeah. What about VCs like me? Have you underwritten some policies for venture capitalists? For the VCs out there listening, what should we know about this type of insurance?

Tanner Hackett (00:21:46): It’s funny you say that. We don’t underwrite venture capital firms or private equity firms yet, although we made a really important hire for someone that’s focused on financial institutions. It’s almost like its own product, financial institutions.

Brian Bell (00:22:01): Yeah, it’s different from a business. It’s like different kind of needs and fiduciary and yeah. Yeah.

Tanner Hackett (00:22:07): As you know, your insurance stack is going to look much different from a traditional operating company. But I will say that underwriting these companies is almost the same as underwriting an investment. I’ve done some investing myself and you do have to look at the financials. They’re the number... one of the most important indicators, if not the most important indicator of is this company going to have run into litigation? Because think about what happens when you’re under financial stress. You start to make decisions that are under duress. You start to make your thinking much shorter term. You’re maybe not allocating enough time to support the team. Your quality of service goes down for your customers. So yeah, we really look at the financials of a business and trying to determine what type of coverage that they need and the pricing for that coverage.

If you’re building, investing, or thinking about where startups are heading, his perspective cuts through the noise.

Subscribe now

From Firewood to Venture Capital

David didn’t “discover” entrepreneurship. He grew up in it.

By the time he was in college, he was already starting businesses—selling firewood to dorm residents, running production services, even negotiating with NASA.

After a stint working for a U.S. Senator, he joined his family’s real estate firm. That’s where something important happened: he started applying early computing to a traditional industry.

He didn’t call it proptech at the time. That label came later.
But he was building it in the early 1980s.

That pattern repeats throughout his career:
He shows up early, before categories exist.

Building Before the Market Exists

One of his first major tech ventures came from a simple observation:
computers could connect to other devices.

That led to the WristMac—a wearable device that synced with your computer. This was decades before the Apple Watch.

From there, he moved into mobile messaging, wireless communication, and early internet infrastructure.

At one point, he built a wireless internet broadcasting system—before Wi-Fi, before smartphones, before the market was ready.

It failed.

Not because the idea was wrong.
Because the timing was.

That’s a key lesson:
Being early often looks identical to being wrong.

Falling Into Venture Capital

David didn’t set out to raise venture capital.

He accidentally discovered it.

While demoing a product, top-tier investors approached him asking how much he was raising. He didn’t even realize that was an option at the time.

Soon after, Warburg Pincus led his Series A.

That experience shaped how he thinks about investing today:

• Founders often don’t understand the game they’re entering

• Investors are constantly looking for reasons to say yes

• The best deals don’t feel like “pitches”—they feel inevitable

Surviving the Dot-Com Crash

David scaled his company during the dot-com boom.
Then watched it collapse.

He went from 125 employees to 17.
Raised capital. Expanded globally. Then lost it all when markets turned.

That experience forced a shift.

He moved from building companies to backing them.

The Birth of Modern Angel Investing Infrastructure

David founded New York Angels, one of the most active angel groups in the world.

Then he noticed something broken:

Startup investing was inefficient.

• Founders applied separately to each investor

• Investors operated in silos

• Processes were manual and fragmented

So he built Gust.

Today:

• Over 2 million founders use it

• Most major angel networks run on it

• It powers the infrastructure behind early-stage investing globally

Instead of being just an investor, he became the system.

What Actually Matters in Early-Stage Investing

After reviewing thousands of startups, David simplifies it down to three core factors:

1. Integrity

If a founder makes decisions that benefit themselves over the company, everything breaks.

2. Passion

Startups require sustained intensity. Without it, founders quit when things get hard.

3. Traction

Not vanity metrics. Not hype.

Real traction means:

• External validation

• Someone getting value

• That value growing over time

Most founders get this wrong.

The AI Shift: Faster Than Any Previous Wave

David has lived through multiple tech cycles:

• Personal computers

• The internet

• Mobile

• Cloud

He believes AI is different.

Not just bigger. Faster.

The timeline is compressing.

He expects:

• Full AGI-level capability within a few years

• Dramatic reduction in startup costs

• Smaller teams building larger companies

• Capital becoming less of a bottleneck

The implication is clear:

Startups won’t just change.
The entire structure around them will.

A Controversial Take: 75% of Jobs Disappear

David makes a bold claim:

Up to 75% of jobs could become economically obsolete within a decade.

His reasoning is simple:

If AI can do a job:

• Better

• Faster

• Cheaper

That job disappears in economic terms.

This doesn’t mean people stop working.
It means the definition of “work” changes.

What Happens Next?

He breaks the future into three groups:

Entrepreneurs (~1%)

People who create new systems, regardless of constraints.

Builders / Technical Creators (~8%)

Engineers, designers, operators building within those systems.

Independent Producers (~15%)

People using platforms to create income (freelancers, creators, etc.)

That leaves a large portion of society needing a new structure.

Which leads to:

• Universal Basic Income (or something similar)

• New economic models

• A shift from survival work to optional work

Why He’s Still Building

Despite decades of experience, David hasn’t slowed down.

Today he’s:

• Executive Chairman of Gust

• CEO of USREM (real estate marketplace)

• Investor in multiple AI companies

• Author and educator

His mindset hasn’t changed.

He still sees every shift as an opportunity to build.

The Real Takeaway

Tools change. Markets change. Technology changes.

But the core pattern stays the same:

• Spot what’s coming early

• Build before it’s obvious

• Adapt when the market shifts

• Stay in the game long enough to matter

David’s career is proof that the edge isn’t in predicting the future.

It’s in continuously rebuilding yourself as the future arrives.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:
00:01 – Intro & David S. Rose background

00:24 – Early entrepreneurial beginnings

02:30 – College ventures & first hustles

03:30 – First job with Senator Moynihan

04:40 – Real estate career & early tech adoption

05:20 – Inventing proptech

06:50 – First startup experiences

08:00 – WristMac and early wearable tech

10:30 – Mobile messaging startup

13:30 – Accidental Series A raise

16:00 – Wireless software & early telecom

17:30 – Internet disruption

18:00 – AirMedia & wireless internet vision

22:00 – Dot-com boom & expansion

23:40 – Dot-com crash & shutdown

24:40 – Transition to angel investing

25:30 – Founding New York Angels

26:30 – Building Gust platform

29:00 – Gust Launch & company formation

30:00 – Writing books & Quora

33:00 – AI impact on startups

35:00 – USREM & real estate marketplace

37:00 – Current ventures & roles

39:00 – Singularity University origins

42:00 – Exponential technology & Ray Kurzweil

49:00 – AI, AGI, and future predictions

52:00 – Impact of AI on venture capital

54:00 – Future of work & unemployment thesis

58:00 – UBI and economic restructuring

01:00:00 – Abundance & societal shifts

01:01:30 – Entrepreneurship in AI era

01:03:30 – Healthcare, robotics, and tech progress

01:05:00 – Founder mindset & entrepreneurship

01:07:30 – Startup success traits

01:09:30 – Investment decision framework

01:11:30 – Closing thoughts & where to find David

Transcript

Brian Bell (00:01:14): Hey everyone welcome back to the Ignite Podcast today we’re thrilled to have David S. Rose on the mic he is a serial entrepreneur one of the world’s most prolific angel investors founder of new york angels and the founder of gust to the platform powering early stage investing infrastructure globally thanks for coming It’s my pleasure. I’d love to start with your origin story. What’s your background?

David S. Rose (00:01:34): So my background is that I am a fifth generation serial entrepreneur turned third generation angel investor, first generation VC, and it’s all additive. So I continue to do all of the above.

Brian Bell (00:01:45): That’s amazing. How did you get started in your career? It sounds like you kind of grew up with an entrepreneurial environment.

David S. Rose (00:01:53): I did. I was a finalist for the ENY Entrepreneur of the Year Award back during the dot-com boom in the 90s. My father won it in 2002. and put things in perspective. My father is currently 96 going on 97 and is more active than I am. He’s finishing his third book. He’s developing a museum in Accra, Ghana around pontificating and all kinds of So he was a wonderful role model. He was an entrepreneur and is an entrepreneur in real estate. And so I grew up from an early age figuring that, of course, this was the kind of thing that you did, although there is a large hereditary component in it. So I have siblings and they’re not particularly entrepreneurial so I got the entrepreneurial gene as it were and so I’ve been starting companies since I was a kid when I was probably 10 or 11 I started Rose Productions a multimedia organization and did graphic design work and provided AV services for children’s birthday parties and so on in high school I was doing a graphic design everything from business cards and student IDs and so on all the way up to when I got to college in high school I created an after school film program and when I got to college I took over the school printing press and turned that into a venture and start a bunch of other things on the side. At one point during my college career, I was negotiating with NASA to see if we could subcontract some space in the space shuttle to send stuff up. My first day at college, I walked in, realized that our dorm rooms had fireplaces, but no firewood. So I loaded up a quart of firewood and started selling firewood on the street. as people moved into their dorm rooms. And so I was then in college with this entrepreneurial bent doing things to the point where this is now in the 70s, mid-1970s, back when dinosaurs roamed the earth. The Yale Daily News did an article headlined, What Academic? are not enough about this really strange person who was like involved in business things as a student because that was back in the day which would seem almost inconceivable to most of the people listening to this podcast. There was no such thing as an entrepreneurial class venture centers new startup things nothing the term entrepreneur didn’t appear in any university class club activity or whatever so it was a really weird thing but in any event all my classmates were Pre-med, pre-law, pre-McKinsey, whatever. And being an opportunistic entrepreneur, I had no clue as to what I was going to do. So as I graduated, I was debating between going to work for Disney as an Imagineer to design rides for Disney World or get an MFA in typography. and book design in Florence. When out of the blue, I got offered a role, since my college major was in urban planning, urban affairs, I got offered a role as special assistant for urban affairs to U.S. Senator Daniel Patrick Moynihan, who I had known and done some volunteer work for. my first job out of college was working for senator Moynihan as his urban affairs person and eventually ended up running his regional office in new york which was an enormously wonderful and heady experience for a young kid out of college here i was sitting on boards with the mayor conducting congressional investigations drafting op-eds for the Times and so on and so forth and that was great and I was having so much fun that I figured that I would be sort of stuck there for life but at heart I really was a private sector entrepreneurial type so I left Moynihan after a couple of years went back to Columbia got an MBA in real estate finance and join the family firm which was in real estate development that’s a firm still going on it’s now this is a centenary so for a hundred years it’s been operating so there was a long history there and I started there in the early east and being a young whippersnapper you know interested in new technology well I figured there were these things called you know personal computers that had just come out I had my Apple II and so on and eventually got a map and stuff. And so, well, of course you’d use these tools in your business the way you’d sort of use AI today, you assume. So I brought my computer in to work to do things with it and realized that there was nobody else in real estate industry. who was using computers for anything. So I have the sort of dubious distinction of having invented the field of PropTech as early as 1981. Wow. 82 or three InfoWorld did a full-page story on me because we were using Apple IIs as I created the first computerized real estate sales office. We had a whole lot of firsts in PropTech because there was literally nobody else in the industry. I developed the first construction punch list software, the first multiple listing servers for New York, the first location-based social network, the first Buildings that had computers in every apartment, you know, a whole series of stuff. We were among the first people to use pro formas, computers for doing kind of Excel spreadsheets for real estate pro formas, which sounds inconceivable now. How would you do it without that? But back then people were doing it by hand.

Brian Bell (00:06:39): that you know if you look at Mad Men in the 60s you know the ad of the the ad show or they’re making ads and it’s like it’s a it’s a sea of desks and all they you can think of the sea of desks as different cells on a on a spreadsheet they’re literally just passing calculations to each other

David S. Rose (00:06:55): It was fascinating. When we were doing, you know, before I brought in Visicalc, remember that the spreadsheet was invented by friends of mine, Bob Frankston and Dan Bricklin, who created Visicalc. Before that, there were a spreadsheet, was a spreadsheet of paper. and you would type in numbers and you would calculate it all by hand on like an editing machine or a calculator. And then if you’ve changed one number, oh, what if the interest rate changes by, you know, two points? Well, bingo, you have to recalculate by hand every sheet in the cell in the sheet and then type it all out again. So it was a major, major change, shall we say.

Brian Bell (00:07:32): So in any event- What was it like to start a business back then? I mean, you couldn’t go out and just find investors. You just had to sort of bootstrap or what did you do back then?

David S. Rose (00:07:40): I mean I was in real estate I didn’t know what starting a business was all about I mean I was doing real estate development in the family business and I did that for 10 years bringing technology into the industry but on the side I was doing little Funky stuff as entrepreneurs are wont to do and in a remainder catalog used to be these things called remainder catalogs where overstock goods were sort of sold I found a digital watch that Seiko had developed with a little chip in it and a little display and it plugged into a computer and could download information from a computer so I said oh this is pretty cool but that was for PC only I was a Mac guy so I got the last 10 units in the catalog and I you know figured out how to you know connect it to a computer I brought on a sort of hit team of friends of mine who were involved in this space because I’ve been very involved in the online what passed for the online computer world before the internet came along there were CompuServe and online systems like that so I was a sysop I remember I remember

Brian Bell (00:08:42): those I’m old enough to remember that

David S. Rose (00:08:44): So I was a stop for the Mac forums on CompuServe And I got, therefore, basically anybody who was doing stuff was on these onlines I mean, you know, now when the entire 8 billion people in the world, half of them are in tech and using stuff In the early days, there was nobody I mean, if there were, you know, 500 people who were in the entire industry That would have been a lot, right? And so they were all online and I was a sysop for, you know, for CompuServe. And so when I figured I got this watch and I figured, hey, how can I connect this to my Mac? So I found a guy named Richard Reich who had created the HP 12C desk accessory calculator for the Mac and he did the software. A guy named Dennis Brothers who had written the, I think, the VINHEX protocol for downloading binary files on a Mac, created the, designed the cable for the watch. The manual was written by Neil Shapiro, the editor of Mac User Magazine. So a little pickup group and it created this, took this watch and branded it and repositioned it as the wrist Mac. Digital watch, 28 years before the Apple Watch, there was this digital watch that could upload and download data from your Mac into your watch. and so I took this to the I think the second or third Macworld trade show that used to be an all Macintosh trade show called Macworld and I got this this booth and I was you know I’m real estate so this would be done on the side of my little I was showing off this watch that could connect to your Mac and it got a fair amount of press interest because it was a pretty cool thing guy comes over and says oh that’s really interesting I want to buy one so my first sale was to a guy named Fred Smith turns out to be the guy who founded FedEx and also went to Bill Atkinson who created Macintosh creators originally and so I had all of a sudden I had people you know getting this product which looked really really cool but it was a side gig because I was in real estate and so time goes on I then realized that hey enough people want to buy this so maybe it makes sense I called up Seiko and I said, hey, you got any more of these out there? And they said, well, it so happens that the product really wasn’t a success. And we had a whole bunch of watches we were sending back to Japan to put under bulldozers to avoid dumping problems. I said, no, I’ll buy them from you at pennies and a dollar. So they turned the ship around and I got back the, I got these, you know, thousands of watches and went into business on the side selling wrist max. which was which was very cool and they got a bunch of press because it was a very remember this is many many years ago when mobility was not a big thing and so I said that’s great and so the next thing I know I said well what’s our follow-on project product for this because of course you have to have a product path. And so I said that was when Motorola came out with a wristwatch pager. I don’t know if anybody remember pagers, but they were like a beeper. You could dial a phone number and go to your watch. And so they came out with this pager on a wristwatch.

Brian Bell (00:11:27): They were like jaws on the floor, you know, how a pager works. You dial a number and then you, you know, dial another number into that number and then that number would now be sent to a device.

David S. Rose (00:11:41): It was a beeper and like, you know, Motorola had one that was on for your wristwatch. So I said, oh, cool. I’ll do a wristwatch, a mobile version, wireless version of the wrist Mac. And so I realized, hmm, no, wait a minute. It’s only a beeper. It just shows a number. That’s not too helpful. Besides, you know, their wristwatch pager was about the size of a cup of coffee. It was a big thing. So, but then they came out with a new pager, a text pager. It was a very high tech thing. It’s called the Advisor and this had a two-line display so you could actually send a text message. I thought that was really cool. And it turned out the way you would get a message to somebody with a pager was you’d call up a phone member and you’d tell the operator and the operator would type it into their system and it would go to the pager So we figured out what the protocol was to get into the dispatch systems and design software to do that from your computer. And so I created a product called Notify, which was the first text paging software. And the next thing I know, I got a call from the Personal Communications Industry Association or whatever, PCIA. Back then it was called Telecater. and they said oh you’re a techie so we’re putting you on our tech committee I said what okay so that’s how I found myself on there on the paging industries tech committee and got involved with the folks who were doing early wireless communications from Motorola and other kinds of companies like that and One thing led to another and all of a sudden I had this mobile communications wireless communications project and we were invited to show it at a show called Demo which back in the day was one of the major sort of trade not trade show but industry conference where people would show off their new products they pick the best products of the year and you could demonstrate it so I’m still in real estate so I go I got called to go to California go out there and show there do a demo of our product and they give you a little desk and a pot of plant and a chair and you can people can come over and you can demo your thing And so as I’m sitting here at demoing our mobile teaching solution, people come over like John Doerr and Ray Rothrock and Esther Dyson and they said, oh, interesting thing you have here. So, you know, how much are you looking for? How much, what am I looking for? So I come back to New York and And the next thing I know, I talked to a guy, I had a family friend of ours, who my father appointed me to, who he said invested in stuff like that, a guy named Alan Patricopp, who at that point was the head of Alan Patricopp. and company later renamed Apex a mega private equity firm one of the biggest venture firms in the business and I showed Alan what I was doing and he said oh really interesting you know we do very big deals but you know I do a little something called angel investing and you know I’d be interested in putting $75,000 into this company of yours I said oh really interesting I’ll get back to you so I went back to my father and said hey I mean Alan must have put a K into this project and my father said well he’s a very smart guy that’s really you know interesting my father is not a techie at all he said but you know always get a second opinion so we have another friend who does stuff like that go talk to our friend Lionel so he I’ll talk to Lionel. It turns out to be Lionel Pincus, the head of Warburg Pincus, which at that point was the largest measurement around. Wonderful corporal guy. And so I show him what we’re doing. And he says, very interesting. What do my guys say about that? I said, I don’t know. Who are your guys and what do they say? He said, I thought you were here to pitch me. I said, OK, I’m here to pitch you. So the next thing I know, long story short, Warburg Pincus actually ended up doing my Series A. So I’m the only guy who fell backwards into raising my first Series A without realizing I was raising money.

Brian Bell (00:15:29): It wasn’t called a pre-seed or seed back then, it was just a series A.

David S. Rose (00:15:32): Pre-seed or seed, come on, those terms are like in the last three or four or five years. No, back then it was series A, right? Because it was sort of a first stock. So they did my series A, and at that point I left real estate and moved into tech. And then that began my tech career. So the wireless software did...

Brian Bell (00:15:53): Do you recall what the valuation was back then on that first round?

David S. Rose (00:15:57): Oh, God, I think they probably invested at about a six or seven million for a Series A. You wouldn’t do a Series A six or seven million today, but I think it was around there, maybe eight. And so the next thing I know, so then I left real estate, moved full-time into the venture, and that was great, and we became the leading developer of paging software, worked with most of the carriers, because smartphones had just come in, and a PCS phone, so instead of just being a wireless cell phone, you now had a phone that actually had a screen display and did text messaging. Remember, the early cell phones didn’t do text messaging, and so that text messaging for phones used the same paging protocols so we became the leading developer of that kind of stuff and that was all great until something appeared on the horizon which was the internet where people anybody could connect their computer to anything you didn’t need to run special software that would dial up your phone and dial into a paging system so we were going to be eaten by you know the shark of the internet so what do we do I said okay well how can we maybe tie this you know wireless stuff into the internet and came up with the idea of taking a pager licensing the pager technology for Motorola and TI and building a device that would connect to your computer and let you effectively send a wireless message to a computer this is before Wi-Fi before 4G before 3 before anything right yeah but the we developed that and it was called the The newscatcher was a black pyramid about this big, little paging chip inside that would plug into your Windows PC and let us send you wireless things over the air. And so because of the way paging worked, we could do broadcasts, everybody got the same thing, or narrowcast so you could subscribe to a particular channel, or individual one-to-one messaging so you could get a message just to you, and you could do it all on the same chip. And so we started showing people this under NDA and they got really excited about it. And so the next thing you know, I did another round to fund this and we came out with something. We renamed the company from originally Ex Machina, a long story. to AirMedia and we came out with the AirMedia Live wireless internet broadcast network which was very cool and we had a bunch of we had Citigroup and we had SK from Korea and we had all kinds of interesting players invested in it and it was this first internet wireless message broadcast which was amazing it shipped and got a lot of press got a full Walt Mossberg column in the Wall Street Journal and we licensed it even before we shipped we licensed it to Hewlett Packard and Compaq and Philips and NEC and Global Village and all kinds of interesting players and the product shipped it got amazing reviews and nobody bought it which was very depressing because this was And that’s when I realized that if you have to, you know, A, tell people what the internet is, E, what a wireless internet news receiver is, and three, who this company, Air Media, that they never heard of before, you’ve got three Three enormous lifts that you’re trying to do, and that just did not work. So we spent an enormous amount of time and effort at 125 people trying to get that, push that string, and it eventually was not to be. We had to do something. So I came up with another idea saying, well, okay, you know, we’re now in a world where you have all of these new smart phones and so on and they have built in wireless there so how about doing a back end platform that would allow because the carrier stepped on control of the phones where so the carrier would decide what use your AT&T or Verizon would decide what you would see when you got your phone this is before Apple changed the whole world with the iPhone and so we went to the carriers and they liked what we were doing and so I went back to my investors and said okay we have this new idea to turn this whole you know thing into a platform for smartphones and communications and stuff and most of them said okay interesting good luck with it that’s fine do what you have to do we’ll pass on reinvesting you know in the round Citibank who was my most farsighted beast said Okay, this might have promise, so we’ll re-up and put more cash in. And I had one VC, my smallest, tiniest VC, who is the kind of player who gives rise to the term Vulture Capital, and they basically said, And they thought that because of who I am and who I was that I wouldn’t dare you know play chicken with them and they said we’ll force you into bankruptcy we’ll force the company to bankruptcy if you if you don’t pay us out in full. And one thing I’ve learned from my very, very honorable family is you don’t succumb to, you know, extortion. You have to curse your convictions. And in this case, we had done nothing wrong. So I said, you’re making a mistake. And by the way, you do realize that the last cash that went into the company was a secured note Citigroup and I personally had put in a secured note to keep the company going during this time and they said yeah but you’re not going to file you know we know you I said you’re reading this wrong you know we’re not going to do this but they persisted and so they forced us into chapter 11 so we did a prepackage chapter 11 the result of which is they got wiped out and Citigroup and I ended up owning the company so We started it, we hit the dot-com boom, and all of a sudden we had people throw money at us, including an international fund that said, great, this coming three months out of coming out of a Chapter 11, that they would invest 60 million bucks at a 120 million valuation. These were very different times back then. And we had filed some cool patents. Hopefully that’s free money. Yeah. that had all issued and you know PWC had appraised our patent portfolio at 125 million so it was very interesting interesting heady time and so we then had gone from 125 people to 17 people we expanded up again hit the dot-com boom All kinds of interesting things happened went on an acquisition spree acquired a company in the UK which got us Richard Branson as an investor acquired companies in all kinds of other places expanded to Europe had operations in France, Germany, the UK and you know go-go days of the of the dot-com boom which was great until all of a sudden it wasn’t so great and it was the end of the dot-com boom and so you recall for those who were around at that point Amazon lost 95% of its value overnight and And if you weren’t Amazon, you can imagine what happened to everybody else. So I get this call before the closing on our round because this investor had been bridging us to their own round. They put in like, you know, 10 million bucks, whatever it is, over a year was they were bringing around together. And they, in two weeks before the closing, they called us and said, you know that round? Not going to happen. So, yeah. That was an interesting comeuppance So with the whole world crashing down We ended up getting Closing on our operations in the US We were still having some revenue in Europe Selling ring One of the first people to sell ringtones That you could download to your computer And stuff like that Licensed it to a bunch of big players But then eventually that was just with the dotcom crash Everything The entire industry had been evaporated We ended up going down as well So at that point I come back licking my wounds and say okay you know no more entrepreneurship stuff but I had helped found the New York New Media Association NINMA back in the mid 90s and NINMA had what was called an angel investor program so being Temporarily beached entrepreneur when it went to the dark side, joined the local angel, you know, the New York Media Association’s angel program and began to help investing in entrepreneurs. But then pretty quickly realized that the angel group had been founded by NINMA, which was his trade association for dot-coms, all of whom are now broke. NIMA had 8,000 members all of whom were unemployed looking for jobs from the other 8,000 members and that didn’t really work so the trade association went bankrupt and it was eventually acquired by the Sofa Publishers Association so they didn’t want the angel group because that doesn’t apply to a trade association so I said okay I think we can do something here so I took the people who had been in the angel groups but now created a new group called New York Angels which is today 25 years later one of the world’s leading most active angel investment groups and so I spent the next several years creating an angel group and investing in early stage companies and doing some some cool stuff I then realized that the angel world itself was very inefficient because back then you would have breakfast and people come to breakfast and then you’d meet with them in person and they’d send you a business plan by FedEx or whatever and that was not very efficient so I figured there was a way to apply technology I had done with real estate technology to this universe. And so I created a company called Gust, the UST, originally called AngelSoft, now known as Gust, as the infrastructure platform for the startup world. The idea being that we would power these angel investment and then entrepreneurs would apply for funding to these groups and if you had everybody on one platform you could do a better software than any angel group could do by itself because they weren’t companies they were loose associations people so we got most of the major angel groups in the world on the platform and that was great and then I realized that We have all these entrepreneurs who are applying to not just one, but multiple angel groups. So rather than just being a pure SaaS platform for an angel group, how about if we open up the other side to founders, re-architect the platform and effectively do it like the common app for college. So as a founder, you can create one profile and then you can share that profile with whichever group you want to And that grew pretty rapidly. And today is now 20 plus years later, we have over 2 million founders who’ve created Profiles on Gust we support most of the world’s major angel investor networks and then over time we began to add on increasing services for founders and stuff for VCs one of which was realizing how difficult it is to start a business because back then you would typically if you had money you’d go to a law firm and they’d give you their startup package of like 5,000 bucks to incorporate you and do your cap table and so on and so forth that was you know it’s a lot of money especially if you’re starting up a company and this is now you know 10 years 15 years ago so we created the first cast platform company as a service we became a Delaware registered agent built an entire instrument and system around it and so now today if you want to start a company you you come on it’s called Gust Launch you come onto our platform you press a button and wham we spin up the entire company we incorporate you to Delaware C Corp we serve as your registered agent in Delaware we set up your cap table we do all of your post incorporation legal and then handle your financing documents and your safes and your convertible notes and the whole bit And then you can graduate up the system, set up your cap table, run your equity program to your option plans, give you all kinds of, you know, events, discounts and cool stuff like that. And that became a success. And then we were getting these ridiculously high NPS scores, net promoter scores. And people love the platform. People say, hey, can I come in? I got to start a company. Can I get all this stuff you’re doing in the platform? And I said, well, you know, the only reason that works is because we have the economies of scale and everybody’s structured and standardized the right way. So, you know, if we didn’t incorporate you, we really can’t do that. But then they said, oh, we really want to come on. So we said, okay, well, you know what? Let’s divorce the actual sort of technology setup stuff from the handshake. And so on We provide advice, we have pitch practice, we have financing, forecasting, we have partners like Forecaster and so on. And so it’s become this really, really popular way for companies to get a lot of the support that everybody wishes they had when you were starting. And then along the way, since I happen to be, as you can tell, a big mouth because I haven’t let you ask a question in the last 15 minutes. I haven’t seen it.

Brian Bell (00:28:24): I’m getting like a free history lesson.

David S. Rose (00:28:27): Even the people didn’t ask. And there’s a website called Quora, Q-U-O-R-A, which some of you guys might know, which is the question and answer website. So I began, I joined Quora when it first started. And I said, hey, look, people want to hear answers. Let me answer their questions. So I’ve answered questions. So now, it’s now, what, 17 years? And I’ve answered over 12,000 questions on Quora. many of which are related to angel investing and early stage startups and stuff and Wiley the business book publisher had decided that the you know this whole tech world was progressing and angel investing which had been this back world backwoods sort of thing was now getting you know to the fore and nobody had ever written a book about angel investing so they look around see who had the biggest mouth which was me and they said hey could you write a book about you know angel investing so I said okay and so as usually I did it backwards I got a contract from Miami and then I got an agent and then I wrote the book which was I basically took all my core answers about this stuff and did a table of contents and threw the one at the other and then rewrote the whole thing and ended up with a book called Angel Investing Gus Guy to Making Money and Having Fun Investing in Startups which to sort of everybody’s surprise became a New York Times bestseller because it was the only book about how to invest in startups and that was pretty cool and so then about a year after that while they came back to me said you know we didn’t really think this would be a bestseller but you know there aren’t enough angels to make a book a bestseller so we went to see who was buying it and it turned out a lot of the people who were buying it were founders who were looking for the other side of the table perspective could you write a book for founders about how to start a company I said okay that so I so I signed up with Wiley and wrote a second book called the startup checklist 25 steps to a scalable high growth business which became and also became your time bestseller and even more gratifying has been adopted by over 500 colleges and universities in the US as the standard textbook for their entrepreneurship courses which is very very cool and that takes you through the entire process of you know There have been a bunch of books I mean some friends of mine have written great books Guy Kawasaki and Eric Reese and Steve Plank and you know how to you know lean startup methodology and the but those were tend to be all big picture stuff right how do I think about what business to start and how do I deal with my team and so on and so forth there actually to my absolute astonishment had not been a book about what’s step one what’s step two who do I file where do I what’s the difference between a C Corp and an LLC why should I you know you know who’s my registration what does that mean how do I get a credit card what do I you know so I said okay well that’s the hole in the market the 20 you know 25 steps to a scale of a high growth business it’s all about okay do it here’s how you do a business plan here’s how you do a business model canvas here’s how you you hire people and so what I did was effectively a survey of all the best and then I point you to Steve’s book and everybody else’s book for the details of stuff, you know, Business Plan Pro and all these various tools. And it became this definitive book on how to actually realistically and pragmatically start a company from step zero all the way through to an exit. And both of those books had been surprisingly good sellers. So they are both, I did the second edition of Angel Investing, came out last spring, is doing really well. And I’m delighted to announce here for the first time that the second edition of the software of the startup checklist is coming out later this month. So you can pre-order it now on Amazon or Barnes & Noble. But that is actually a fascinating place. I was rewriting the book starting, what, about a year ago, maybe, when AI, which is clearly eating the world, said, how do you rethink about AI for a founder? What does that mean? And of course, over the last year, what AI has done for startups is just gone through the roof and accelerated to an unbelievable degree. So I told Wiley, you know what? Nobody can write a realistic book today about AI is just changing so fast. I can’t point them anywhere. I can’t tell them what to do because it’ll be out of date by the time I finish writing the chapter before it’s published. So I tell you what I want to do. And he said, what? I said, okay, instead of putting any directions or any recommendations or any links in the book, I want to put in QR codes. So they said, oh, I haven’t done that before, but I said, okay, great. So basically, the book that’s coming out later this month, the Startup Checklist, does not have a single live link in it. What it has are QR codes, which you can scan with your phone for every chapter and every resource pointing you to to the latest and greatest LLMs and tools and stuff for this. And so the way you think about starting a business in the age of AI is completely different from the way you, I mean, it is mind-bogglingly different because the tools are so powerful. you can do and the teams you can build and how you find money all that stuff are so different and so that’s now all covered so this is very much a book for the new AI world so you got that there and then anyway with all this going on I am an entrepreneur at heart so we We’re looking at you at Gust and we said, well, we have the investors and we have the founders and we have the digital company and their cap tables and everything else. Well, the logical thing is to move to a platform like a financial markets platform where people can buy and sell their shares and this kind of stuff. So we started doing that in this day and age. That’ll be a blockchain based thing. So we started working on that and then pretty soon realized that actually people don’t want to buy and sell shares of startups. Yes, if you’re just pre-IPO, you can do a Forge Global kind of thing. but the random stuff that angels invest in there’s no market there but one of the first family offices we were we approached for funding for this new thing said you know we’re not really interested in doing startups but you know we have a lot of real estate could we use your platform to sort of syndicate and raise you know get liquidity for real estate so I said I hadn’t thought about that, but now that you mention it, I do have a little bit of background on real estate. So when you think about that, I went off and realized that, hey, actually, that’s a very interesting universe because real estate has a very – it’s the world’s largest asset class, has a big liquidity issue. I mean PropTech which as I invented in the 80s has now finally become a bit of a thing but the real estate industry is still way behind the times and so here was a potential to provide liquidity so we started this so we took this thing we were doing at Gust spun it out as a standalone company called USREM the US real estate market and created a platform for limited partners like just in like in venture funds limited partners in real estate funds to buy and sell their interests in this commercial real estate. And it turned out to be a really interesting play. We now have about $3 billion of property signed to the platform, 2,500 LPs. have become a very interesting play and it turns out that when you actually have a platform you don’t need tokenization per se I mean they’ll obviously get there eventually because that’s the way the whole world is heading but the problem is everybody was trying to do tokenization first and use that to build liquidity and they got to kind of have that have the platform we’re doing the real market building liquidity and then you can tokenize it so anyway so we now have a great platform called us rem and so since all of this is additive you know i’m the you know full-time ceo of us rem i’m the executive chairman still august as an investor i was the lead investor in an ai company called thinkable ai which is doing an amazing stuff in the in the film generation space not the the technical rendering of images. But before that, how do you develop your script? How do you cast it? How do you do all that kind of stuff? So it basically has used AI for the entire film production process, putting, you know, entire studio down into a one or two person operation. So that’s a lot of fun. And then I was the lead investor and chairman of the board of a company called Book It and Go, which is using AI for the travel industry. deal inclusive with the world with the largest hotel owners association among other things and they have got a back-end API that will let your agents or whatever go out and automatically book travel and stuff all this is added up so basically today here we are you know many years since I started this entrepreneurial stuff I’m full-time running U.S. Rem the U.S. real estate market executive chairman of Gus chairman of Book and Go chairman of Think of LAI and Plot.com I’m signed with Wiley for my third book called The Real Estate Investing Bible that’ll be out in 2027 and I continue to lecture and then along the way I helped to found something called Singularity University which was a guy named Ray Kurzweil wrote a book called The Singularity Is Near so I’ve been doing this future AI technology stuff since 2008 actually Ray and I just recently did an event at the 92nd Street Y Discussing AI and where all this future stuff is going. So I’m the founding track chair for finance, entrepreneurship and economics, a singularity university. And so I’ve been doing a lot of that stuff as well. So all this leads to a very crazy, overly committed that I’m having a lot of fun doing hopefully creating interesting and good things haven’t taken a vacation in a decade because I’m just having so much fun doing what I’m doing and hope to be around for I figure I got another you know 50 60 working years left in me although Ray is convinced that we’re going to hit escape velocity on human longevity by 2032 which would mean if you can make another six or seven years you can effectively live forever or at least live in a box forever or something so we’ll see where that goes But that’s the backstory.

Brian Bell (00:37:36): That’s amazing. What an amazing journey. And there’s so many things I want to dive into. I mean, just the singularity thread alone could be another three hours. I mean, I read Ray’s book 20 years ago on a plane when I was living in New York, actually. It just had a profound impact on me. You know, kind of his vision of the future and how accurate he had been.

David S. Rose (00:37:55): Ray is an absolute visionary. He came out with his first book called The Age of Thinking Machines in 1999. 9 and that was all about how you will eventually have artificial intelligence which in 1989 the idea when a computer could barely count 2 and 2 the idea that it could think was insane in 99 he comes out with a second book called The Age of Spiritual Machines saying okay yeah not only are we going to have computers machines that think, but they’re going to think so much that they’ll think anything like a human can think, including being effectively spiritual or conscious. Both of these books, people in A, they sort of didn’t think without a trace, but they were picked up mostly by the science press and people like us who were sort of ahead of the time.

Brian Bell (00:38:37): Some of the predictions even in the 89 book, you know, about the internet where they sounded crazy because he said, everybody’s going to be connected. And at that point, there was, you know, maybe tens of thousands of nodes on the internet. So like, this is absolutely crazy.

David S. Rose (00:38:52): One of my friends and early co-investors is a guy named Nicholas Negroponte, who founded the MIT Media Lab. Nicholas was fond of saying that when the internet started, he knew everybody on the internet. He personally was friends with every single person on the internet. because of course back then there were about eight people on the internet connecting you know MIT and Stanford and whatever so anyway so his second book about 99 was the age of spiritual machines and then his third book comes out and so the first one he He doubled down on his predictions. His third book was the one you’ve read in 05 or 06 called The Singularity Is Near saying, okay, not only are we going to have, you know, AI and not only am I keeping to my projections that you’ll have that by about 2029, not only And that is incomprehensible for most people because our human brains are wired to think linearly. You either it’s a straight flat line or it’s a line with a slope that’s going up. But in reality, technology is growing exponentially. And that means that, you know, double, double, double, double. One of Bray’s great illustrations is that if you go up one step for every step forward and you do it linearly you go out forward 30 steps you go up 30 steps and at the end of 30 steps you’re 30 steps up but if you do it exponentially you know 1, 2, 4, 8, 16, 32 you go out 30 steps and you’re not up 30 steps you’re up 1 billion steps that’s the difference and we can’t conceive of that and so In the early days of Singularity U, it was all about, you know, the only people who were thinking in exponentials, you know, Ralph Merkle, who invented public key encryption and been served with the Internet and credible, credible faculty. We’re all talking about this exponential stuff. And so, you know, now Now people are finally realizing, hey, wait a minute, you know, the version of ChatGPT that was released this week is, you know, twice as what it was, you know, the week before. So it’s gone into hyper acceleration. So Ray, in his book, The Singularity Is Near, as you know, said, okay, well, what this means is you’re going to keep advancing at this insane pace because it’s not going to stop. And you can take this back to the beginning of recorded history. You can take this back to the formation of the Earth. And actually, I did a TED Talk, TEDx Talk, The last spring, which people can check online, was really good about tracing the evolution of technology from, you know, fire to, you know, the singularity. And so he said, if you just project forward where the inevitable where this goes, I mean, you’re going to have machines that can think soon. Super Intelligent Machines that have reached AGI, Artificial General Intelligence, by 2029, right? 2030, in the next two or three years. But what happens when you now have computers developing computers and it goes at this warp, warp, warp speed straight up? The only logical extension of that, he says, is that computers and humans will somehow merge. Either you will live forever outside of your corporeal body, or who the hell knows what, right? And the one thing we know about when that’s going to happen is we’re not going to have a clue what that means. And nobody forget science fiction. Nobody can even imagine with our imaginations today what that will be. And that’s why they call it the Venter Rosner, the technological singularity. And so when this book came out in, you know, 05 or 06, you know, there were three reactions. One group said, well, that’s ridiculous. Human exceptionalism computers are just big calculators and so you know not gonna happen forget those guys the second group said okay your historical analysis is accurate your projections are accurate as far as they go but what you’re not taking into account is that we as a society are not that good before we get to your mythical singularity you’re gonna have some guys gonna bring a suitcase nuke and blow up the world and a giant electron magnet pulse will take out all the technology you’re gonna have a plague that’s gonna take over the entire world you know we’re not gonna get there so we’re

Brian Bell (00:42:50): There’s some filter ahead of us, basically, yeah.

David S. Rose (00:42:54): But the third group, who was slightly more optimistic than the second group, said, okay, assuming that we’ll stay a half a step ahead of the bad guys and the plagues, and we’re going to control ourselves and not blow up the world, that third group said, okay, if the similarity is coming, what do we do about it? And so they got together, and this was Google, NASA, Cisco, Nokia, A bunch of individuals and VCs, you know, me, a few other folks. And with Ray and Peter Diamandis, who created the XPRIZE Foundation, created some whole Singularity University, which was the first, you know, effectively think tank about this XPRIZE. exponentially growing technology, looking at it in, you know, our first class was in 2009, since 17 years ago now, saying, hey, this stuff is coming. And the idea was to identify the leading future visionary entrepreneurs and stuff around the world, Bring them to Silicon Valley, give them an intensive 10-week postgraduate program with the leading lights of the industry, and then set them out to solve humanity’s grand challenges so that we could get to the singularity. And it turned out that it was amazing. and that grew exponentially. I think the first year they had like, you know, 500 applicants for 40 places. And the next year they had a thousand and then 2000. And then, you know, so by the time you got to the year four, they had like three or 4000 applicants for 80 places. It was a remarkable, the average, the demographics, the third year, the demographics, I recall, average age was 35, the average person who attended, we had people from 35 countries out of the 80 people who were there, the average person had started

Brian Bell (00:44:33): I did apply, actually, but I had no experience, so I didn’t get in. I was an applicant back in 2009 or 10.

David S. Rose (00:44:42): Let me quickly, do not feel bad because this was more selective by two orders of magnitude than Stanford or You know, whatever, right? I mean, it was an absolutely incredible, incredible group. And so, and I was the founding director for finance, entrepreneurship, and economics, which I taught single-handedly for three years. And then after three years, it was so successful that as OpenAI has just done, it has started as a not-for-profit, but they ended up switching it as OpenAI is doing into a for-profit and they got venture funding and moved forward with it. For-profit AI. effectively so I mean I’m still involved as an advisor and you know an equity holder there but I haven’t been day to day involved for a number of years Eric Reese took over for me for the next year the entrepreneurship program and stuff

Brian Bell (00:45:28): Oh, nice. Eric’s been on the podcast, actually. He’s publishing a new book. He’s coming back on in the next month to discuss his new book.

David S. Rose (00:45:34): Yeah, I know. I’m actually on the podcast with him as well. We’re doing one for good this week. Oh, fun. We’ll talk about it. So anyway, it’s been a fascinating life looking at AI and having the background for all of this. It’s like a lobster being boiled, starting in cold water, right? Having the background of starting. I mean, I took typing in high school where I learned to type.

Brian Bell (00:45:54): I did too I took it in middle school and high school yeah I remember that and at least it was on computers it wasn’t on typewriters but I did take typing I’m guessing they still teach that right

David S. Rose (00:46:05): You know they don’t know how to do it keyboard skills right which is which is something else so I mean I can touch I can type faster than I talk and I talk pretty fast so the you know seeing here in 2026 where things are going where AI is taking us where technology is taking us is just the most eye-opening fascinating it’s like Magellan going around the world

Brian Bell (00:46:25): How do you how do you think it impacts us as angel investors and VCs like you know the singularity because as of right now it’s it’s a positive impact right there’s more startups than ever they’re able to get traction more than ever but like kind of five or ten years out as we approach the singularity what do you what do you think happens

David S. Rose (00:46:41): Well, I think it gets to be very interesting, right? Because a good part of the whole angel VC world is all about the economic growth over time. And with the compression of time that you’re seeing, it’s going to be in the availability. You can spin up an entire team. It’s not going to require the intensive capital to do whatever. So capital will be available. Times will be sort of short. It’s going to be a very different dynamic. And I think that you’re going to see everything morph. Angels are great because they’re nimble and they’re willing to take risks and so on and so forth. And I think that they’re going to move into a different kind of support mechanism. Realistically, nobody can predict. At this point, things are moving so fast. I don’t think anybody can predict 10 years out. Remember, if the singularity is happening in 2045 and we’re now at, you know, 2026.

Brian Bell (00:47:32): At the latest, I mean, you could argue, I mean, if you listen to the Moonshots podcast guys, which you all know those guys personally, you know, they’re like, they’re saying we’re like right now, we’re in the singularity right now.

David S. Rose (00:47:44): I mean, Ray and I talked about this in June at the Y. I mean, what Ray has in mind for the singularity is not just like supercomputing. I mean, it’s really, I mean, he really sees something you can envision, which is computers and humans really merging. And that’s not where we are now. So the fact that Ray can look at this and say, okay, yeah, we’re going to have AGI in 2029.

Brian Bell (00:48:06): And you could say you have AGI now, like you could argue, I mean, we passed the Turing test and we didn’t even think about it.

David S. Rose (00:48:13): You know, the best prognostication here, I think, at least to my mind, are the guys who did the AI 2027 analysis. That was a great paper.

Brian Bell (00:48:22): Yeah, I read that one.

David S. Rose (00:48:24): Great paper. Brilliant. It’s scary as all hell because according to them, we’re are they going to be pets or food for somebody within, you know, three years or whatever. But they do a quarterly update of their forecast, which is interesting. The latest one came out last week. So their first quarterly update said, okay, we were a little bit advanced, so we’re dolling back our thing. So instead of being, you know, 2027, 20, 30, 20, 31. Latest one last week. They just went back the other direction. Okay, wait a minute. It really is picking up. So we’re pulling it back in. And it’s now like 20, 29. And so 20, 20, 29 gets you to what they, I think are calling superhuman coding, right? Where you just basically, basically, yeah. Full on AGI And from what I’ve seen, I mean, I’m doing, you know, everything is AI these days, right? And the power, people are not currently using AI If you are listening to this podcast and you are not currently using things like Claude Cowork for something really significant in what you’re doing You are missing the boat. I am telling you the time is here and now the tools are here and this stuff is beyond transformative. I mean, you’ve heard my background in this up to my ears and have been for my entire life. And I am just blown away by what this can do. And it really takes somebody who knows what technology can do to understand how big it is. PC, internet, mobile, cloud. Yeah. So I tend to agree with them, and I think the consensus is that you’ll have real, full, legitimate AGI by 2029, right? So that’s two to three years from now, which is really completely set and forget, right?

Brian Bell (00:50:03): Which means you could talk to it like a human, and it’ll just kind of do any task that you want it to do, just like a really well college-educated human could do.

David S. Rose (00:50:12): Yeah, I mean, and so what’s interesting is, you know, I had a class that I started teaching back at SU in 2006 called No More Companies, No More Jobs on the Weight of a World Without Work, which I’ve taught over the years at various places including Yale and elsewhere.

Brian Bell (00:50:27): Wow.

David S. Rose (00:50:28): And, you know, Ray and I were doing a talk at Bard University about Bard College about three or four or five years ago. And Ray does his whole, you know, future and amazing stuff. And then I did my economics thing about saying, you know, we are going to have effectively 75 percent economic unemployment. by 2030 or thereabouts, right? And people are, you know, looked at me. I almost got stoned. People were saying, you know, first question, Mr. Rose, have you no heart? Have you no soul? I’m saying, hey, I’m just a messenger. Don’t kill me. I’m just a messenger. I’m just like, yeah. But I mean, but seriously, at this point, think about what we are saying. Everybody who says, you know, well, look at all the economic projections from these guys that, you know, you’re going to have a 10% production of people’s jobs are safe because we’re going to be Hello. If what we just said is that we are going to have within two to three years technology that can do anything, literally anything better, faster and cheaper than a live person then definitionally by the laws of economics that person is economically unemployable you can say okay I’m going to pay you to do something because I want to pay you to do something or because it’s charity or because I’m electing people who will do the works progress administration right but the bottom line is the entire you know 75 percent of the population my projection somewhere between 50 and 100 percent to call 75 percent will be economically unemployable within single-digit years. And that’s something nobody is talking about because where that leads is, okay, what does that mean for society? Well, you know what? People are not being employed because technology is being employed to do the things that people were doing, which is creating value. And so you’re going to have a...

Brian Bell (00:52:17): I think what a lot of economists would say here is that there’s this lump of labor fallacy, right? I think you could, like if you say AI is going to take our jobs, right? You could like say the same thing about like, oh, immigrants are going to take our jobs, right? If we let too many immigrants in, they’ll And so like AI is actually probably an accelerant to our economy where it creates more both supply of labor, but also demand.

David S. Rose (00:52:44): But again, no, because what what everybody The fundamental difference here is that every time in the past that technology has created new jobs, the industrial revolution and the term saboteur came when the Dutch began to do automatic looms and the workers would throw their wooden shoes into the machinery to stop them to stop technology and the sabot was the You know, Luddites, Ned Ludd was destroying the mills, right? Because what happened there is the technology took care of some of the grunt work and let people go to higher level things, right? Right and but the problem is what we’re seeing with AGI is the higher level things that you could do are now being done better frankly that could only be done by people in education that’s why you needed a college education high school education then college education then graduate education right hello all of that is now being done what COVID showed us with everybody virtually everybody’s job could be done remotely sitting at home in front of your Zoom right and if your job could be done by you sitting at home it could be done by a computer who could deliver the same work in there right Right, right. And so the question is, what job will people do that computers and technology and AI and robots cannot do better, faster and cheaper? It’s not flying a plane or driving a bus or writing software or doing a marketing plan or writing a novel or making a movie. What the fudge do you think people will do? In a society in United States of America, the average IQ is 99 and the average person has no college degree. Think of it. Let that sink in. okay what is that person going to do in an era like two years from now three years from now when technology can do everything better than the you know person with a graduate degree in anything right and so well therefore but that being said this technology is being used to create value which is which is which is good but how do you pay for value different question right so what you got to do is somehow it’s like now you get back to economics definition of economics the The science of the allocation of scarce resources. Well, all of a sudden, if Peter Diamandis says you’re in a world of abundance, that constraint is off the table and you have a whole different universe. And now what you have to do, the purpose of economics, is to take the value that’s being created by the Dr. Justin Marchegiani of automation or corporate income or whatever, right? And that’s going to be applied back to something like UBI, universal basic income. So you’re going to have a society where the value that’s being created by technology enures ultimately to the benefit I’m just going to play pickleball on the court you know But some people are just going to watch movies or TV. Some people are going to want to play sports, compete in things. You know, entrepreneurs like us. I mean, I think I’m going to be creating companies or whatever the next version of a company is because that’s what I enjoy doing. Right. You know, designing AI things. And so but we will have it. I believe in this I do believe in Peter and Ray’s thesis of abundance which we will have at this glorious period with technology doing the things that we have to do allowing us ultimately to do what we want to do because I think that people are

Brian Bell (00:56:59): You’re a techno-optimist like me. I do believe in the future that we will work it out. And then, yeah, I think what you’re saying is, which I agree, is that there are things that we’ll discover as we continuously get closer and closer to our machines and our AI that we can’t even imagine right now, right? absolutely being a being a podcaster look at this like that didn’t exist 20 30

David S. Rose (00:57:21): years ago right it was like Larry King basically sorry so that’s that’s my prognostication for the future I’m looking forward to it I think I’ll be around for it and I think I will be an active participant in it well especially if your dad is

Brian Bell (00:57:33): uh you know in his 90s still writing books and starting businesses I’m sure you’ll be there too with all the medical technologies coming out in the next you know five or ten years are you following all the like the Yamanaka factors that are like in stage one and two and stuff like that for Dr. David Sinclair has some stuff for glaucoma and stuff like that you know everything

David S. Rose (00:57:54): I had my my hip replaced right so my parents both had their hips replaced when they were in their 70s or 80s and it was a big long process you were it was like a two days in the hospital surgery and then you were like you’re kind of rehabbing for

Brian Bell (00:58:07): like six months and yeah yeah

David S. Rose (00:58:09): So I have a hip replacement last June. I walked into the operating room, climbed up on the operating table at 11.30 a.m. I walked out of the hospital on my own two legs at 4.30 p.m. Wow. With a four-inch supergluid incision and zero pain. Three weeks later, I’m back. walking around I have not had one second of pain and all that surgery was performed by a robot with my great doctor guiding the robot but it was robotic surgery and it turns out that if you actually can do a 3D CAT scan and a digital twin model and a this to that and use or robotic tool to go in there, you know, so hip surgery, but it’s effectively a nothing burger. I mean, this is not, I mean, I would have my other hip replaced, you know, tomorrow. No problem. Literally, no. Admittedly, in my particular case, great surgeon, great hospital, great tools. I define the best possible outcome, but I’m here as a walking version of saying this was a three or four hour experience and a couple of weeks of using a cane steady myself and then done zero pain zero problem zero sequelae you know now I’m just waiting for them to get that for knees which is I think a few years after hips but you’re going to have technology I mean I was talking to a friend of mine who was the editor of the Journal of Oral Maxillofacial Surgery or whatever and he was saying yeah today if he was going in for oral surgery you’d use a robot right you’d do that you know and so you’re going to find whether it’s robotic surgery the already radiology right there’s no question that computers can do radiological assessments better than pathologists at this point

Brian Bell (00:59:50): But you know what’s funny about that you bring up the radiology example there’s actually more radiology and more radiologists than before AI infiltrated that profession so there’s the the what do you think of it like the Jevons paradox this

David S. Rose (01:00:03): gets back to the question of you You know, exponentials and scale and speed. And there’s a saying that, you know, everything takes, you know, far longer to happen than you thought it would. But when it happens, the impact is much, much larger than you thought it would be, right?

Brian Bell (01:00:18): Well, we totally overestimate in a year, but we underestimate in like 10 years, right?

David S. Rose (01:00:23): Yeah, you know, so what you see now, now that we are heading up this exponential curve, this stuff is moving faster and faster and faster. And you can’t look at where things are today and extrapolate from that because it’s this exponential curve. And so we, you know, are we going to you know there are some things you know food replication teleportation light travel those are things that are probably still beyond our understanding of physics and you know you’re not going to be able to get to right but you know Ralph Merkel who was my co-faculty member at Singularity U is doing stuff with desktop nano fabrication that is The progress that’s being made across the board on literally everything these days is and it’s all, you know, building and building and building on itself.

Brian Bell (01:01:11): Yeah, it’s recursive self-improvement, right? And the big AI labs are doing that as well. What do you think it is about entrepreneurship that keeps pulling you back towards that because you’ve been on the investor side of the table as well. And then you kind of, you know, throughout your story, you kind of keep going back to the entrepreneur side. What is that?

David S. Rose (01:01:30): Well Benjamin Franklin wrote his own epitaph for his tombstone and it said here lies Benjamin Franklin not statesman not political person here lies Benjamin Franklin printer he thought of himself first and foremost as an entrepreneurial businessman as a printer right And everything else was on the side, the being wise, writing a declaration and so on and so forth, right? So I think of myself here as David S. Rowe’s entrepreneur. So I am an entrepreneur first and foremost, and everything else sort of comes in and out and around and follows that. So it’s not a question of ever having left I mean it’s that’s so baked into my DNA and so then the question what is an entrepreneur which over time changes right as you’re going to have you know in my taxonomy when I when I talk about the future of work and where this goes and effectively jobs get lost in the same reverse order that they came in so the order they came in so for example the first job for resource extraction right I mean hunting and fishing and pulling things out of you know growing food and so on And so, you know, and then you had manufacturing to transmogrify it into something else and then transportation to get it somewhere and then retail. And then, you know, and so all of a sudden, you know, resource coal miners and transportation, autonomous cars and trains and planes and, you know, all the retail stores, all this stuff, boom, boom, boom, boom, boom. is dropping off and so where I think the future world gets rebuilt in my taxonomy of where things go is I think entrepreneurs are like cockroaches you can’t kill us we are designed definitionally entrepreneurs are all about Taking whatever you have in front of you and making something out of it, right? So if there was the world changes, great. Whatever the world looks like today, that’s what I’m going to use and I’m going to go do it. So the percentage of natural-born entrepreneurs in society, I believe, is about 1%. You know, my numbers are anecdotal, but they are, I think, pretty accurate. So about 1% of the population is entrepreneurial. And as long as people live, there will be stuff for them to do. They can be social entrepreneurs, business entrepreneurs. Whenever we go to a post-economic abundance world, they’ll find something else to be an entrepreneur about, right? But then, at least for now, they typically require other people to help get their vision instantiated, which that number is shrinking and shrinking with technology. But let’s call those STEM people, all the people who work for SpaceX and, you know, so-and-so, right? So you have, you know, STEM, creative, you know, folks as the next level. And what percentage of them are playing at a high enough level to be in these startups? 1% entrepreneurs. So double that, double it again, four, double it again. So, you know, eight doublings, right? So maybe you have 8%, right, of the population are capable of being these STEM creative types working in these startups. What comes next? Well, what What you’re seeing is we started this starting about 20 years ago and then growing amazing platforms designed to let everybody flower to their you know to do things that you couldn’t do by yourself right you know one person is not going to set up their own taxi service but give them uber and they can now be as you saw in the startup of one from reed hoffman right you can be your own your own company but leveraging the platform like uber to do it and that works for etsy if you’re doing handcrafts or airbnb if you’re Renting out your spare bedroom or you know Upwork if you’re doing spreadsheet design or whatever right so you have a group that I call entrepreneurial personal producers they’re using these platforms and they are bringing personal skills that are viable today that still includes the plumber fixing your sink that may not be around you know in for four or five years when robots can do that for you. But for now, you know, whether you’re a hairstylist or whatever, you can add value. And those personal entrepreneurial personal producers who are using these platforms to manage their billing and pay their taxes and do their stuff, how many of them are there? I don’t know, double the STEM creative guys. So there were eight, so that’s called 16. All right, so you have 16% of entrepreneurial producers and 8% of the STEM and creative types and the 1% entrepreneurs and that equals about 25% what’s the next layer after the entrepreneur person producers and so therefore you get you know that’s it so I think you end up that’s how I get to 75% unemployment right I think you add I think you end up at something that smells like 75% economic unemployment right within single-digit years. I don’t think it’s two, but I don’t really think it’s nine. somewhere in that universe. And we therefore had better figure out the whole, you know, UBI taxation, robotic, societal, economic infrastructure.

Brian Bell (01:06:23): But again, like we’ll have super intelligent AI to help us figure it out, right? We’ll just ask our AI and say, hey, what should we do here?

David S. Rose (01:06:31): The scariest thing about the whole AI 2027 piece was that you have, you know, robots, you know, AI deciding what’s better for humanity.

Brian Bell (01:06:39): Yeah.

David S. Rose (01:06:39): You know, open the odd bay doors. I’m sorry, Dave, I can’t do that one.

Brian Bell (01:06:44): That’s funny. I’d be remiss if I didn’t ask you about angel investing before we wrap up here. You’ve probably seen thousands and thousands of pitches. What are some characteristics of startups and founders that make you lean in and just want to sign this safe for the convertible note right away?

David S. Rose (01:06:58): Well, there are three core things, right? And so basically every investor has a thesis and you have to, you wouldn’t be an investor if you didn’t think you could make money doing this and you think you can make money doing this for a reason or set of reasons because you have an approach that if I do that, then I’ll make money. And that’s your thesis, right? You might know an industry, you might feel you can spot them or whatever it is, but all investors effectively share three theses in common the three basics in my experience it stacks up like this the first one is integrity which sounds like it’s preaching or it’s bullshit or it’s something you you learn in your religious service but no for speaking as an angel investor integrity is absolutely the number one thing because startups are making a startup founder is making a decision every minute on everything right and I have to believe that they are making that everyone I can’t afford to bet my cash on that so that’s a stopper integrity is number Number one. Number two, you know, have been an entrepreneur long enough to know what it takes to get in there, and that’s passion. You better have a passion for what you’re doing because otherwise you’re an operator, you’re a functionary, and you might be Tim Cook, which is great for getting Apple to a trillion dollars, but you’re not Steve Jobs. If you want to be Steve Jobs, you better have Steve Jobs’ passion. And even Bill Gates, who is not known as a passionate person, it’s an internal kind of passion, but you have to have that total drive to to do it. And so, you know, the question is, you know, where do you go beyond that? With those in place, You have the kinds of skills, the domain expertise, the leadership ability. One of the things that we, the biggest single disconnect between investors and founders is I say, I’m looking for traction. And a founder says, well, traction, sure. I’ve got, you know, I filed for these patents and I’ve got built this great team. And I look at all the people on my wait list for doing this. And you know what? As far as investors are concerned, that ain’t traction. There is this, you know, I think you know you use that word I don’t think it means what you think it means as far as investors are concerned traction is something outside the control of the founder outside the founder’s control that shows that they have created economic value for someone else and that that value is growing traction is sort of the the Big thing, right? And the business itself has to be scalable. The person and the entrepreneur, integrity and passion and skills, the ability for the business, the core fundamental, what is the business? That’s why restaurants aren’t scalable, right? You need a chef, but it’s going to do a restaurant. You want to do two restaurants, you need two chefs, right? A scalable business gets better as it gets bigger. So it’s got to be scalable. And then you have to have traction. So you have those three things, and that at least gets you in the front door. And then it’s a question of what’s my personal thesis? I don’t invest in fashion, but I do invest in prop tech. You know, I don’t invest in medical devices. I don’t know anything about it, but I do invest in SaaS platforms. So you pass those first three gating things, which everybody looks at. And then the question is, is it a match for you know for what I am looking to do with my thesis and if the answer is yes if you pass those three you know gating factors and then you’re in my wheelhouse for the things I want to invest in at that point the conversation flips and it’s now I want you to sell me I really need to do this deal because how can I be an angel investor if there’s nothing for me to angel invest in right so therefore you pass my three things you’re in my wheelhouse and great let’s make this work please pitch me sell me you know I want you to tell me right so so then I you want me to get really really excited about what you’re doing and then after I then I make the internal decision yes I’m gonna like this I love this that’s my thing I’m gonna invest in this company if ellipses dot dot dot right and then you go to the other side of the thing and now it’s sort of the downside which is the diligence I’m doing your reference checks and your customer checks and your this and that all that kind of stuff and you got to get through the gauntlet but that’s after I’ve made the decision intellectual emotional decision to invest and then you got to pass through all the diligence factors before you actually get the checker

Brian Bell (01:11:20): Right well this has been the easiest podcast I’ve ever recorded I could have talked to you all day we could have just literally it could have been my whole work day just eight hours talking to David really enjoyed it we’ll have to have you come back on when you when you publish your the new edition of your book and and the upcoming book as well I really appreciate you taking the time where can folks find

David S. Rose (01:11:36): you online and find your books com is my personal site books are available anywhere fine books are sold Amazon Barnes & Noble new ones might actually be in bookstores and in in a couple of weeks. And if you’re interested, if you’re a founder, a startup founder, you should absolutely check out gust.com, G-U-S-T.com, all kinds of great things there for founders. If you are an angel investor and you should definitely check out your local angel group because they have a lot of things that they can really support you with. And if you’re in an angel group, you know if they’re not using gust they should certainly check out gust on the angel side of it you know the my my current venture is the US real estate market USREM securities.com which is doing cool stuff if you’re interested in seeing AI applied to film

David S. Rose (01:12:21): Amazing, David. Thank you so much for coming on.

David S. Rose (01:12:23): My pleasure. Thanks a lot.

That shift is already underway. Instead of typing keywords into Google, users are turning to ChatGPT, Claude, and other AI systems to get answers, recommendations, and even complete transactions. The result: fewer clicks, fewer visits to websites, and a new layer of decision-making that most companies don’t control.

Jochen Madler, co-founder and CEO of SiteFire, is building for that future.

After starting in academia with a focus on statistics and reinforcement learning, Jochen left his PhD to tackle a growing problem: brands are losing visibility inside AI systems, and they don’t know why.

Subscribe now

The Moment Everything Broke

A turning point came when Google introduced AI-generated summaries in search results.

Some companies saw impressions stay stable—but clicks dropped sharply. Users got their answers directly from Google, without visiting the source. In one case, this shift hit revenue hard enough to trigger a major market reaction.

That exposed a simple truth:

Traffic is no longer guaranteed, even if demand exists.

If AI systems answer the question, your website might never get visited.

Search Has Changed—But Most Teams Haven’t

Traditional SEO was built around keywords.

Users typed short queries. Companies optimized pages to rank for those terms. Traffic followed.

AI breaks that model.

Instead of 3–5 word queries, users now write full prompts. AI systems expand those prompts into multiple searches, evaluate results, and generate a single answer.

That means:

• Your brand might be mentioned without a click

• Your competitors might be summarized alongside you

• Or you might not appear at all

The unit of competition is no longer a webpage. It’s inclusion in the answer.

From Customer Experience to Agent Experience

Most companies still optimize for human users.

Jochen argues that’s the wrong focus.

AI agents are becoming the primary interface. They read content, compare options, and increasingly take action—whether that’s recommending a product or completing a purchase.

This creates a new layer: agent experience.

It’s not about how your website looks. It’s about:

• Whether AI systems can understand your content

• Whether they trust your information

• Whether they can interact with your product or API

In some cases, this is already happening. AI systems can discover tools, authenticate, and use services without a human clicking anything.

Content Is Getting Cheaper—But Harder to Get Right

AI has driven the cost of content creation close to zero.

Anyone can generate blog posts, landing pages, or documentation at scale.

That creates a new problem: volume is no longer an advantage.

What matters now is precision.

The winning content is:

• Structured for AI systems, not just humans

• Aligned with how models retrieve and rank information

• Positioned in the right places across the web

In some cases, that’s your own website. In others, it’s platforms like Reddit, YouTube, or third-party publications.

There is no single playbook. It depends on the query, the model, and the context.

Why SEO Isn’t Enough Anymore

Many teams assume this shift is just “SEO 2.0.”

Jochen disagrees.

SEO focused on ranking pages. AI search focuses on assembling answers.

That changes how visibility works:

• Ranking #1 no longer guarantees traffic

• Being cited matters more than being clicked

• Brand presence inside the answer becomes the goal

This is where new categories like “Generative Engine Optimization” are emerging—but even that framing may be too narrow.

The bigger shift is toward influencing how AI systems think, not just what they rank.

The Next Step: Agents That Act

Today, AI systems mostly inform decisions.

Soon, they will execute them.

That includes:

• Booking tickets

• Purchasing products

• Integrating APIs

• Managing workflows

In one early example, an AI system discovered a company’s API, authenticated, and started using it—without any human involvement.

That’s a preview of what’s coming.

When agents take over the full journey, the “funnel” changes:

• Discovery happens inside AI

• Evaluation happens inside AI

• Transactions happen through AI

If your product isn’t part of that flow, you don’t exist.

What This Means for Founders

This shift is still early. Most companies see less than 10% of their traffic coming from AI systems today.

But it’s growing fast.

That creates a window.

Founders who move early can:

• Capture visibility before the space gets crowded

• Build systems optimized for AI from the start

• Lock in distribution as agents become more dominant

Those who wait risk losing their primary growth channel without realizing it.

The Bottom Line

Search isn’t disappearing.

It’s being abstracted.

Instead of users navigating the web, AI systems are doing it for them. That changes who controls attention—and how companies earn it.

Jochen is building for that world.

Because when AI decides what gets seen, recommended, and bought…

marketing doesn’t go away.

It just moves behind the interface.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:
00:01 Introduction & SiteFire Overview
00:28 Jochen’s Background in Statistics & Energy Systems
01:18 Founding SiteFire & Early YC Journey
02:14 Moving to San Francisco & Ecosystem Exposure
02:50 YC Experience & Key Takeaways
04:00 Advice for Founders Applying to YC
05:25 From Reinforcement Learning to AI Marketing
06:20 Identifying the AI Search Opportunity
07:35 Google AI Overviews & Traffic Disruption
08:30 Rise of AI as the New Interface
09:17 Why AI Search Isn’t SEO 2.0
10:53 Rational Search vs Human Behavior
12:29 Marketing as a Math Problem
14:01 AI Search vs Traditional Channels
15:26 Websites Becoming Agent-Focused
16:33 Human Traffic vs Agent Traffic Trends
17:46 Adoption Gap & Real-World Usage
19:14 Long-Term Evolution of Search & AI
21:01 GEO vs AEO Debate
21:43 Agent Experience as the New Frontier
22:02 What Winning in AI Search Means
23:12 Distribution Risk if Google Disappears
24:10 OpenAI’s Role in Search Distribution
24:41 GEO vs SEO Debate Revisited
25:11 Markdown, Content Structure & AI Readability
25:58 SiteFire Product: From Monitoring to Execution
28:47 How AI Models Rank & Retrieve Content
31:00 Early Product Traction & Breakthrough Results
32:12 Content Strategy: Domain vs Platforms
34:00 Product vs Services Approach
34:58 Hardest Technical Challenges
36:31 Misconceptions About AI Marketing
38:25 Content Commoditization & Future of SEO
39:17 Competition & Incumbent Risk
42:27 Future of AI Agents Replacing Search
43:27 AI Agents Executing Transactions
44:56 Agent Reviews & API Ecosystems
45:47 Where Value Accrues in AI Stack
48:15 Vision for SiteFire & Agent Funnels
51:10 Rapid Fire: Startup Lessons & Beliefs

Transcript

Brian Bell (00:00:57): Hey everyone welcome back to the ignite podcast today we’re thrilled to have Johan Madler on the mic he is the co-founder and CEO of Sightfire it’s a YC winter 26 company and a team ignite portfolio company building the marketing infrastructure for the agentic web helping brands show up inside AI systems like chat GPT I think I’ve heard of those guys a Gemini and Claude thanks for coming on

Jochen Madler (00:01:18): Yeah, thanks for having me.

Brian Bell (00:01:19): Well, I’d love to start with your background. What’s your origin story?

Jochen Madler (00:01:22): So I grew up in Germany, Munich, Germany, and my background is really like technical, so statistics. I fiddled around a little in energy markets and like statistics for energy market simulation, did reinforcement learning optimization for energy grids. But then, yeah, I did a PhD in the same thing and then dropped the PhD to found Sidefire. So me and my co-founder, we met like seven years ago. We know each other since a long time, had all like hackathons competitions together. And then one day he came around and made me reconsider.

Brian Bell (00:01:52): That’s amazing. So he kind of pulled you out and said, hey, let’s do a startup.

Jochen Madler (00:01:56): Yeah, pretty much.

Brian Bell (00:01:58): That’s great. It’s always good to have those kind of friends that pull you out of your comfort zone. And then how did the app applying to YC come about?

Jochen Madler (00:02:07): So we started building Sidefire end of 2025. And back then in Germany, it was really about this AI monitoring space, really fiddled around with like GPT-5 came out. Everyone was thought of like, this is not the end of like anything or everything. software engineering all of it didn’t happen but then we like really got into this AI search Google AI model launch Google overviews launch and so we really started building this monitoring infrastructure and quickly sold it to like pretty large companies in Germany like BMW Deutsche Bank and Allianz and then we got into YC so everything actually happened pretty quickly we started end of 25 in like November building and in December January we were already in San Francisco

Brian Bell (00:02:46): What was it like to come live in the U.S., live in San Francisco, you know, there for three or four months during the entire batch?

Jochen Madler (00:02:52): Yes, exactly. Yeah, San Francisco, actually Dogpad is pretty close by. Yeah, it’s amazing. I mean, I’ve been there before in my time at Stanford. So I know the area. And of course, it’s like, yeah, if you’re building something like this, you at least want to have the ecosystem. And YC is, in my opinion, the perfect thing, because, of course, the agentic web and everything like these systems are being built there. And so, yeah, we now have actually like are in good contact with all of the big firms. And yeah, also, the ecosystem is just like really amazing.

Brian Bell (00:03:20): What was that whole experience like for you?

Jochen Madler (00:03:21): So I think we’ve been here to an accelerator for a previous startup in Munich and I’ve always been and also my co-founder like in various entrepreneurial programs so I know a little about the scene but I think what I see it’s always been just North Star right like where it’s like so much better and the interesting thing is the program is actually like less program than for most other programs so it’s like less events and stuff like this but the ones you do get and also the office hours you get it’s less but it’s more high quality and so it really like lets you work a lot and focus on your things and make meaningful progress and then steer you in the right level of like yeah the right basically level of granularity and steer you in the right direction

Brian Bell (00:04:02): Do you recall any key aha moments or takeaways during your time in the program that you’d love to share with founders who might be listening and considering applying to YC? There’s a lot of people that don’t get into YC, right? They only take one or two percent. So any kind of takeaways and things you’d like to instill to the founders out there listening that might be going zero to one like you just did?

Jochen Madler (00:04:26): Yeah I think like the interesting thing is there’s really no magic in the sense like there isn’t the magic pill they just hand you on day one right but it’s really like this ecosystem and like I think having the right mindset but also having the right ecosystem and like surrounding so surrounding yourself with like good people I see famously said like ditch many events like there’s so many tech events in SF just don’t go it won’t help you don’t hang out where like the scenesters are hang out where your customers are yeah so really stuff like this don’t go to like office spaces it’ll just slow you down move into one apartment like be scrappy stuff like this and then really hold yourself accountable so like set ambitious goals and then always like yeah meet them and if you don’t meet them it’s okay but you like break it down and know like why it didn’t work and then try again

Brian Bell (00:05:14): I love that yeah I mean I think if you have money in the bank I love that advice just don’t hang out with the scenesters don’t hang out with the people out there you know happy hours and events and there’s you know dozens every single day you could go to right just just focus on your customers build stuff they want and ignore all the noise right and set ambitious goals I love that those lessons so I mean you’ve had a really interesting background I mean deep reinforcement learning energy systems now marketing what are you seeing out there that others may have missed

Jochen Madler (00:05:42): Yeah, I feel like like there is no really romantic thing like that went from like energy and then now it’s like now it’s AI on marketing. But I think the connecting dot is, of course, LLMs and statistics. And I think I’ve always been interested in just playing around these systems. and now we are I’ve always been a user like one of the first users when complexity came out always found it amazing that it’s like better than HTTP because it’s not hallucinating it’s doing web search it’s grounding so always been on the user side and then using these systems more and more we found out that hey of course there’s also something happening with the ecosystem right like with the web pages and I think this insight was something that really struck me and us as a team because I think like right now with these systems we are seeing the biggest shift in search since Google came around essentially and the way people search online or actually the way people interact with brands is fundamentally changing and this is why it’s so exciting.

Brian Bell (00:06:36): What was that inevitable moment do you recall where you’re like okay this this needs to exist

Jochen Madler (00:06:40): Yeah so there was this monday.com earnings call and like I think it was q2 2025 and then there was like some LinkedIn post about it actually we had a an intern at the time who discovered this kudos and then we of course like dug in but the the the CEO or so So Google AI mode came out and Monday.com gets a lot of traffic from Google organic search. And so it’s the biggest growth driver. And then Google AI mode came around and they saw all of a sudden they had these overview boxes and they weren’t like they had impressions, but they had no clicks anymore because people wouldn’t click through to their site and their sales tanked. And then the CEO was asked in the earnings call, hey, you have this Google AI overviews now, what are you going to do about it? And he had no clue it was new, right? He had no clue how to solve it. And I think then, because of his answer, the stock tanked by, I think, 30% or something in one day. And we said, hey, obviously, there’s a lot of traffic and organic traffic, which is not just clicks, right? It’s users willing to pay. And if this is breaking down, then there is a big opportunity here. So something needs to happen.

Brian Bell (00:07:42): Yeah I mean I think people are sort of waiting for that shoe to drop on hey I’m searching for I don’t know whatever I’m searching for it could be a supplement or something and yeah tell me tell me if I should buy the supplement for my health or whatever and give me the 10 blue links you know go ahead and just or you know recommend a product and give me the different prices and the shopping experience and yeah they haven’t really figured that out yet. Yeah and so people me especially I’m LLM first right that’s my home you know when I open up a new tab it’s ChatGPT right and I use cloud for a lot of stuff too and I think that’s becoming the new interface of the internet right is agents and AI and so a lot of smart people saw this including this analyst who stumped the Google execs on the earnings call why didn’t they just build this themselves

Jochen Madler (00:08:30): So I think there is a lot of companies now in the space, of course, and like it’s a it’s a growing space, especially in the monitoring piece. But I always felt like and it’s true, like this is the first step. So letting people basically peek into the box, right, like and opening this up. having a lot of synthetic searches going on to tell people, hey, what are LLMs now seeing? Do they see your brand or don’t they see your brand? And I think this part is one thing, but this is really close to SEO. And so I think the most obvious thing, which we still hear today is like, hey, yeah, but this, I mean, sure, this is a new thing, but it’s actually just SEO 2.0. And so I think this is something we don’t believe in. I think it’s way bigger. Yeah. And so this is why I think many miss this or just dismiss this.

Brian Bell (00:09:13): Yeah. So you’re a special kind of founder who started in academia and switched to startups. That’s a very unique path. You know, others have done it. But like, what did you believe about markets and maybe startups back when you were in academia that you now think is completely wrong?

Jochen Madler (00:09:28): So I think like I was actually in energy markets with an energy study for the German government assimilating these kind of things. And I think like market theory and like search is a zero sum game, right? It’s attention. And so you have these 10 blue links, at least in the old world in Google. And so whenever like it’s it’s you on there it’s not someone else on there so it’s really like this game of getting to the top and in the serious game I think like the interesting piece about markets in my opinion is what kind of like assumptions you take of people for like humans essentially we have this like homo economicus where like people are rational and like take all of these choices and I think the interesting thing with which the whole SEO industry in my opinion is built around is that people are actually humans right and they don’t they’re not rational they actually like go for these fancy CTA optimized titles and so it’s really not like yeah it’s really not an efficient market in my opinion and with LLMs and the whole thing gets, in my opinion, more rational now. I’m happy to dive in how, but I think the search space is essentially getting bigger than just template links. And also it’s getting more rational because it’s actually not humans now reading these pages, getting maybe blown off by cool parallax animations or colors or images. It’s really now more about product-like specs and information and data.

Brian Bell (00:10:41): Right yeah because the AI you got this 150 160 postdoc PhD level assistant and almost every human domain that can digest books worth whole textbooks the whole academic literature on something and then say hey I’ve crunched it and I’ve actually I’ve looked at I have this whole context window on you as a person because of all the all the data you’ve given me and you know I’ve crunched this all into this nice little paragraph of stuff you need to know right and so it is creating this like hyper rationalization around information dissemination and discovery and search. That’s a really fascinating take.

Jochen Madler (00:11:19): Yeah and I think like so for example if we like insurance is in my opinion a great example because usually like insurance is something like a semi-complex product but you You usually have like some broker or something helping you out and they would like know stuff and now a lot of people are going to CHPT and like starting this information journey pretty high up in a like not in a buy intent level but more like a learning experience like hey what is an insurance what kind of what different insurances are there and then for going from there and I think now it’s like the search in Google search you could bet on these keywords and just focus on the very transactional ones where you actually already buy an insurance but you lose so much context and with the LLMs you now have this whole like learning infrastructure the whole chat window essentially which carries so much more context and it gives also actually brands and a better way upstream to jump in there and provide also not only like a product but also value in explanation

Brian Bell (00:12:12): Yeah, so you’re essentially just applying this optimization to human attention. What part of marketing is actually just a math problem?

Jochen Madler (00:12:19): So I think, I mean, as I told you with the zero sum game, I think it’s really like marketing is essentially like trying to map what product, like what brand is selling to people who want to buy it, right? And that’s really the keyword idea of Google which in my opinion is actually genius we have people are looking for something and they encode to something in keywords they type into a search bar and brands are selling something and they also encode it by essentially putting these keywords on their site and so it’s really nice like that’s that’s working and it’s a huge like it’s this optimization problem of actually which keywords do I want to target and then we have of course other competitors also targeting these keywords and you get these bidding dynamics and this SEO and I think this still is this is marketing in a nutshell and the interesting thing now with AI search is with GEO this space this like encoding is way more like there’s way the information bandwidth is way higher now I don’t need to in court like encoders previously Google searches were pretty much like three maybe four maybe five keywords that’s it and I need to encode all of my knowledge all of my desires all of my background into those four or five keywords and now without elements we have all of this context we have way more longer prompts and so we can basically match this way better and that’s really exciting

Brian Bell (00:13:33): Most people think the AI search is just like another channel, right? Just like Google search or social media. And I think you think different. How do you think about it?

Jochen Madler (00:13:42): So in my opinion, like websites, controversial statement maybe, but I think websites are going away. So like over the short But I think this will become, as the systems get better over time, this will become less and less Actually, the web that’s being browsed will be browsed by agents and not humans. And therefore, the information and how it’s presented will also change. It will change for agents essentially. And the most exciting part about this is that it’s not only about visiting a web page and structuring it with more text. but of course we see this now with OpenClaw these agents can now not only read stuff but actually also do stuff on a webpage and once we enter this this gets really exciting because then the whole online funnel with like marketing and checkouts and everything can be done by agents actually it’s already done by agents and so this is really like this this blows the whole thing open because now agents can actually not only do the research but also the transactions

Brian Bell (00:14:50): yeah I can go talk to my agent like I would talk to my EA and say hey I need to you know buy some basketball tickets tonight and I want some good seats I like to sit you know kind of center court you know find me the best price on SeatGeek or whatever and I need two tickets I know the game’s at seven o’clock tonight and go and that’s it and just goes and does it and what’s your budget oh did you want to spend five is it okay to spend 500 or whatever the number is yeah that’s fine cool and it’s just done and you think about you know how that used to work in you know the standard web you kind of have to download the app or go to the site got to create an account and then you gotta like run the search and then you gotta like get your credit card info in there and then and you know increasingly we’re all gonna have these super intelligent super helpful agents assistants that can execute tasks do research and and buy stuff for us yeah so it’s a it’s a completely different thing how do you think about uh and what are you seeing in the platform today between kind of human traffic versus Asian traffic

Jochen Madler (00:15:49): like in the web per se or for our customers or yeah or just in the zeitgeist your customers on the web yeah I mean it’s it’s kind of early we live in the future right you and I like in San Francisco NYC You know it’s definitely not like the future is here it’s just not equally distributed yet is the phrase and I’m curious what you’re seeing in your platform of like how equally distributed is the future so yeah that’s a great tag actually and also it’s in my opinion really like important to state so you take like all the humans and you put them in this Bay Area it’s already like a set but then all of this also in YC and so it’s really like everyone for example like in this ecosystem everyone thinks everyone is using cloud code which is not the case right like it’s just up one percent and so yeah it’s always important to notice then when doing the reality check and talking to our customers which are really like much go to go to brands they usually see AI traffic in like the single digit percents of all their web traffic and so it’s really like sub 10% for sure and for most it’s also sub 5% so it’s it’s small but it’s not negligible anymore and actually it’s and the whole point is it’s really growing faster and so that’s that’s the the first derivative essentially of like how fast is it growing that’s what matters.

Brian Bell (00:17:00): yeah and technology has this way of sticking around we think of You know AOL is dead but I think there’s like still like tens of millions of people that log into AOL every day still and you know I think the same can be said for like Yahoo like we think of like the Yahoo directory kind of web experience is dead but people still log into Yahoo and kind of use the directory and you know I still use Google for some stuff you know sometimes I just want a quick answer or I need I need to actually find a resource online so technology has this way of Kevin Kelly wrote this book I’d love to Kevin if you’re listening I’d love to have you on the pod someday but he wrote a book called what technology wants and in there he kind of goes through all the evolution the evolution of technologies right and how technology evolves and it proliferates and there’s definitely like kind of the you know the crossing the chasm early majority kind of framework but he makes this point that the technology never really goes away it just sort of maybe gets more pushed to the side It’s still kind of there, but a new technology comes and usurps the whole experience over time. And I definitely feel like agents will do that. We’re living in the future. We’re seeing what could happen, how the world could be. But I think the rest of the world is probably still five or ten years away from really going, oh, this is just my default experience, talking to my agent every day, and it just does stuff for me.

Jochen Madler (00:18:19): Yeah, I agree. Like, especially the web, in my opinion, there will be like a super, like, for sure, there will be both types of webs for humans and for ages coexisting for a long time. for sure but I also like this proliferation of technologies it’s getting faster every year I feel and with ChatGPT for example like sure it started already now I think three years ago maybe even three and a half years ago in November 2022

Brian Bell (00:18:43): I think yeah right a little over yeah three and a half years ago maybe yeah even like close to four but still like now like I think is it a billion or eight eight nine hundred million users now so it’s almost like almost a billion users yeah yeah an eighth of the human population right in like four years

Jochen Madler (00:18:58): An agent will come right behind that, right? I think a lot of us have had that OpenClaw experience where it’s like, oh yeah, it just does stuff or the cloud coworker or the cloud code or codex. But I don’t think a lot of people have had that ChatGPT moment yet with agents. I don’t think OpenClaw is going to be it, especially because OpenAI bought them. Maybe it’ll just be ChatGPT one day. You’ll just be talking to it and it’s like, oh yeah, by the way, you want me to like check out and buy that for you? Can I just go get that for you?

Jochen Madler (00:19:25): Actually I think like Chbt acquired like the Open Cloud Founder in a sense where like he joined OpenAI I think that’s the right phrase To build agents inside yeah To put this inside right so we can really expect these like systems Chbt in particular to have these kind of capabilities sooner rather than later and I think they already have this distribution right of like 900 million users and so like with a snip of a button if they just roll out one more feature all of a sudden this now is like agent ready and can do stuff on websites so yeah I think we’re not too far away

Brian Bell (00:19:55): yeah fascinating world but what is in this area of and what do you call in this category is it AEO is it like answer engine optimization like what is what do you call it

Jochen Madler (00:20:05): So yeah I think there’s this big debate right between GEO generative engine optimization and then answer AEO answer engine optimization in my opinion like I don’t care we started with the GEO keyword which we just use for now but I really don’t have strong opinions and in my opinion like this GEO it’s it really narrows people into this 0.0 thinking which for sure with this monitoring is a part but it’s not the main part in my opinion it’s the first one but more like the more interesting thing is marketing to agents so agent experience instead of customer experience I think this would be the new thing which really will accrue a lot of value and will get many people excited

Brian Bell (00:20:42): Yeah, love that. So what is winning in this AI answer space actually mean in practice?

Jochen Madler (00:20:48): So for brands, like why are brands working with Sidefire and us is because, of course, they want to be, first of all, present, like pretty much marketing 101, right? Like it’s you want to be present in front of the user. So you want to be mentioned by these, first of all. So when they type in something about an insurance, the one insurance company wants to mention them, like primarily in a positive light. And then the second thing is, so that’s brand awareness and it’s brand mentioned. And so the second objective for these companies is not only getting brand mentioned, but also having their site, their link in there as a citation, such that people click on it and do a checkout. And so it’s really just true optimization, but it actually goes like it’s it goes both ways. So when you do a search, there is these web searches in the background happening. They got all these pages, the LLMs go through these pages, understand what’s interesting enough to be cited and these citations and these sources, they also inform which brands are being mentioned in the answer. They ground the whole thing. And so it really goes hand in hand.

Brian Bell (00:21:54): So imagine Google and OpenAI disappeared tomorrow. Let’s take them one at a time. If Google disappeared tomorrow, what percentage of companies would lose their primary source of distribution?

Jochen Madler (00:22:04): Well, a lot. I think over 50%. So we see this now with companies. Google is really everywhere. And I think like even with this AI, I mean, they’re doing their best drive with Gemini and so on. But Google is huge and Google search. So I think at least we did this like estimation napkin thing. So in my opinion, it’s like 60, 70% of companies are having their main distribution channel or as organic traffic, either paid or organic via Google, which is crazy. So take away anything, of course, which is really like outbound driven sales or like farming, mining, stuff like this, defense tech, but anything else, actually, anything consumer related is really, really just.

Brian Bell (00:22:45): Yeah. And what about OpenAI?

Jochen Madler (00:22:47): I mean OpenAI is really directly eating into the cake of Google so it’s really the same number in my opinion they just share this one thing because ChatGPT is really it’s using web search in the background it’s actually using Bing and so it’s the

Brian Bell (00:23:00): same kind of system okay interesting so in your opinion is GEO just SEO with a different label or is it different

Jochen Madler (00:23:09): So we or I think of it as like GEO in the narrow sense is really like this SEO 2.0 so making sure that you have the right web pages such that people right now are being like the brands are being mentioned in these answers and their links are being cited so that’s GEO and I think the broader term then which is exciting future is this agent experience maybe then it will be subsumed by GEO maybe it’ll be a different thing I think we don’t know yet yeah

Brian Bell (00:23:34): I’ve already learned to use markdown files now more when I interact with cloud and OpenAI because it just reads it easier, right? So I’m just like, okay, like you prefer a markdown file here. I’ll just start using those and they’re kind of cleaner anyway.

Jochen Madler (00:23:47): It’s interesting. There’s this like, yeah, just the first version of my opinion of this parallel web we’re seeing that like a lot of customers of ours now actually Cloudflare launched this as well just to serve a markdown clone, a markdown mirroring essentially of every web page for the agents. Strip it of any like pictures anything just raw information but who would have thought like marketing is such an old technology and now we are back there it is so

Brian Bell (00:24:10): your product doesn’t just analyze it executes maybe you could tell us what that means

Jochen Madler (00:24:13): yeah so I think that’s the interesting part which we really like built out during YC now this year this is so we started with this monitoring so you have all of these synthetic prompts you run through and you can tell people what web pages are being shown in the answers and once you know this you can then start like improving right and so Improving means there’s really two kind of broad categories you can improve as a brand and it’s always through the web so you put out different content on the web about yourself about others and this influences these background web searches of these models because right now we don’t I don’t have a scalable way of infiltrating the training data of JTBD5 or something. This is over time, of course, with Common Crawl, but then it’s still stuff you put on the web. So in my opinion, what we do is you have on page and you have off page. And for on page as a brand, it’s really your own website. you can either create new content or you can create or you can improve your existing content to better tailor it to elements and for off page that’s also something with the whole trust issue or like the trust thing so we have these editorial coverage sites so it could be PRs outlets that write about you and mention you as a brand and then UGC, User Generated Content. This is really like YouTube, Reddit, Quora, all of this. And so as a brand, you have really these like the on-page stuff. It’s really scalable. This is what we automate with Sitefire. We can create blog posts and we use this information, essentially what we see. and then reverse engineer what’s working right now to tailor and to tell you what to do and then actually do it for you and so I think there’s one big difference which we’ve seen with like many companies and we also tried this in the beginning is gathering a lot of information and then running like a lot of like tests to find out elements of websites that are working and then applying them as a general set. This, in my opinion, doesn’t work because the world is too complex and LLM is a true statistic in their nature. So what we found with a lot of models and also other companies out there, they are running these monitorings and then they are trying to extract out certain snippets but in order to fit everyone it becomes so generic as of yeah you should add FAQs to your site or something like this right which is so generic it doesn’t help and so what we do with our system we for each topic you want to appear in we gather the top-sided pages so we can understand And what is it? It could be anything. It could be Reddit. It could be your own page. It could be a competitor’s blog post. And then depending on what works, we have agents that essentially do the whole SEO experts work for you in digging into those pages, understanding what’s the reason they’re being cited there, and then applying this to your own web page to tell you either, hey, you don’t have something on this topic yet. You should write about it. Or, hey, it’s actually these three Reddit threads that are really driving this answer. You should post something in there. Stuff like this.

Brian Bell (00:26:52): Yeah, I wonder how they’re doing that at the OpenAI cloud and Google and Grok for that matter. Are they using some version of PageRank to try to figure out what’s authoritative? You know, because back when I was in demand gen marketing, it was all about like backlinks and Google every month or two or three would publish a new version of their algorithm. And you had to like go back and like redo all your pages to try to rank higher for different keywords and stuff. Do you know how I was kind of working under the hood there?

Jochen Madler (00:27:21): Yes, so I think the interesting thing here to understand is, so you have a prompt, you input it into jhpt, and then jhpt takes this one prompt of you and expands it into multiple search queries. And these usually are longer than what humans would type in. They look weirdly specific for human Google search queries, but they carry all of this like context you bring in and so they run maybe like three to five to even ten web web searches at a time in parallel gather all of these pages and for each of these web queries it’s really Google’s ranking at work right it’s really still the same Google ranking which has as you said rightly page rank baked in and so essentially you get the top ten Google searches for each topic which already includes this page rank and authority and so on and then they gather these pages and then there’s one so it’s proprietary right we don’t know for sure but there’s something like a rank fusion algorithm going on where you basically compare across these 10 web pages was there any web page that appeared for example twice and was it an in one search it was number one and the other one it was like number eight and then maybe we have some other ones and so essentially like a waiting across the search grace to find out the most important pages and then once you have them they go into these pages and read them top to bottom with like these snippets and that’s where then the markdown mirroring comes in and so you can optimize your page on a technical level to once it’s crossed the border of it’s being found in the background it’s also passed this rank fusion algorithm now it’s being considered by the LLM and then it reads through all of these snippets and then you can have a structure and like content, for example, facts or like quotations, stuff like this, which then gets extracted.

Brian Bell (00:28:53): What was the first version of the product that actually worked for you guys?

Jochen Madler (00:28:58): I think, yeah, so an official or like what sold and what like what was used was already like back in 2025 as a monitoring product. But we always felt like this didn’t really work because it’s really just an analytics solution. And so it was really until, in my opinion, in the beginning of March in NYC when we started with this actions, which by the way, we also tried first distilling a lot of features didn’t work really. But then the first thing which our agentic approach was in beginning of March when we had one customer who was really like going ham with our situation. I was really one of the first ones to really try it they’re like a programming website like Coursera but for like only for programming courses and they really started applying our suggestions and like creating the content the tired side fire outputs launching I think one block was a day and we increased their AI traffic by 170% in just 10 days and we also connected the network log so we could see every bot visit in Cloudflare and this was the first time and now it’s sustaining so it’s really nice I think then there was this aha moment of holy this actually works yeah

Brian Bell (00:30:01): I’m wondering if you have a blog and you’re a brand, a startup, brand, enterprise, whatever, should you have that blog? A lot of people put them on Substack these days and maybe they put a subdomain on their main website. Will the AI read it just as well if it’s on Substack or are you not getting that accrual of all that crawl?

Jochen Madler (00:30:22): It’s better to have it on your own domain. It’s better to not have it on a subdomain, but really on your top domain. And that’s what we see works best. And then again, it really depends on the topic you’re looking to appear in, especially in these programming courses we had for this one customer. Some of them were really driven by Reddit, especially in JGPT. So there’s also a difference in the models. Many people know that JGPT really likes Reddit. But the thing is, with our system, there isn’t such a thing as like there’s one size fits all and then you’ll appear it’s really that for different topics sometimes you need to be in Reddit but sometimes you need to you can have a blog post and this works really well and sometimes we actually see that medium posts and substacks are being like referred more it’s also some sort of UGC content yeah so it really depends

Brian Bell (00:31:11): that’s interesting yeah I just wonder you know I have a blog for team ignite chose to put on Substack it was easy to do and I could have put on my website I guess I could use WordPress or whatever and but I just kind of decided that Substack has better distribution

Jochen Madler (00:31:21): and I also like Substack I mean that’s yeah I think for humans and the app and the ecosystem I think if you if this you have a lot of like yeah I think then it’s stronger it’s better I think purely from an LLM perspective and if that yeah it’s probably the website

Brian Bell (00:31:35): so how much of Sitefire today is product versus services

Jochen Madler (00:31:40): I think like the actions that we have it’s really fully automated right now it’s an agentic pipeline that runs but we are really like in like many for many customers we have like two tiers the ones who would want to do it self-serve and then the ones who have like some sort of human in the loop and of course we are like we want to see the systems we want to like know all the edge cases and so we are checking these outputs right and refining every day and so I think there’s the whole system is automated but we are really on the sidelines looking at these outputs and then thinking about is this really something only for this one customer this one topic or is this a general issue that we should improve our system so

Brian Bell (00:32:18): that’s what we do nice and then so far what has been the hardest technical problem you guys have had to solve I think

Jochen Madler (00:32:25): the yeah really just like the actions it really took like quite some time to get something that actually works and it’s not too generic so it’s probably the same thing we experienced with cloud code which I think a week ago the whole source code got leaked and so there isn’t really that much magic

Brian Bell (00:32:42): come out and kind of say like oh there’s we didn’t it was April Fool’s joke there was nothing actually like do you think they’re just kind of trying to cover cover up and yeah I think so yeah

Jochen Madler (00:32:51): But, I mean, like we had this, Boris, the creator of Cloud Code, he was at YC, gave a talk. And back then in the talk, you already said, like, they don’t do a big secret about Cloud Code. They had, like, competitors or customers, like enterprise customers, checking the source code, like, for any, like, things. They would happily give it out to them because the thing is, it’s not magic. but it’s a lot of tiny pieces put on top and I think we experience the same thing with our harnesses that we put our agents into and there’s not one single thing but you need to get a lot of things right on top and then suddenly it’s amazing and so for us that meant having the right harness and the right system problems but then also giving these agents the right tools at the right time so they have for example they have firecrawl they have certain APIs that they can use which are used in the SEO world and so they can really decide for certain things which makes them so flexible to employ different tools and this really is what took a lot of time yeah yeah I mean we’re the same and we have a lot of little AI thing processes we run you know you can just easily replicate them because they’ve been iterated on dozens if not hundreds of times right you have to kind of work with the AI over months and years to get it to do just the thing you want to want it to do what’s a widely held belief about AI marketing that you think is completely wrong so yeah I think like the obvious thing is this SEO thing that people yeah

Jochen Madler (00:34:14): I just think you want to get on a webpage and then yeah my opinion like so we have a strong belief that like the whole point of marketing agencies is something that will probably be automated most of it I think not all of it there’s always room for like experts but I think this is something that many people haven’t really gotten their head around because it’s just so ingrained. But I think why is this the case? I think there’s a deeper reason. With SEO, there was always a new hack in town. Every six months or something, there was a new SEO agency that could claim, hey, now things change, a new update. We figured it out. Let’s do a project. and then they applied some system until it didn’t work anymore and you needed these people to constantly be on the ball and follow these updates but now with AI and our systems it’s reactive anyways we monitor each day we see shifts in the content that’s being cited and now we don’t have people anymore needing to look at updates we can have all of these updates in markdown files and feed it to the agents they know in 10 minutes they know more about how the algorithm changed than any human could And so, yeah, the whole system, the whole machinery about SEO agencies, in my opinion, is something that’s going to go away.

Brian Bell (00:35:25): Do you think that content will get commoditized to zero as well?

Jochen Madler (00:35:28): Yeah, I think that’s an interesting I don’t know for sure how it will happen but for sure you see that a lot of people I mean with cloud code right like you can put out web pages left right and center like as many as you want and I think really the content creation is commoditized it’s going to zero everyone can create lots of content the question now is creating the right content that works and that’s a big difference so yeah I think the in my opinion what will happen is that Yeah, this backlink thing and like this reputation will become more important again. And because the cost of creating new content is just so low and keeps decreasing.

Brian Bell (00:36:09): So what’s to stop like SEMrush or HubSpot launching this tomorrow? Why wouldn’t they win in this space?

Jochen Madler (00:36:14): I mean, they’re trying, right? Like they rebranded, they’re now launching their suite. I think, yeah, the interesting thing is they had this SEO thing going on for 20 years and they didn’t launch any actions. I think they have a lot of people, but I think you could have said the same thing for like many companies. So far, I don’t have any problems actually.

Brian Bell (00:36:32): It’s the hardest thing about being an investor, by the way. is like when you look at a company and you’re like, will this company get big before the incumbents decide to eat their lunch, right? And just build it themselves. And so I’m constantly like kind of watching the OpenAI, Google, Claude, Grok roadmap, you know, what they’re building. to try to inform myself of like okay because if they move in and they build in your space you’re done unless they choose to buy you right but you know if Microsoft just decides like hey agents are cool like we’re just going to build like this like super powerful agent that does everything I’m sure and they are they have the co-pilot agents and stuff and it’s not very good but like how do you think about that like if you were like put your VC hat on

Jochen Madler (00:37:14): you know and you’re looking at a company there was like a brutal scene in my opinion actually with the cloud code shock that was at YC there was like 100 people of YC founders invited and then one guy was building memory management for these coding engines and he was pitching his thing like in the Q&A and Boris was in the end like of his yeah you know what you should pivot because we’re launching this next week

Brian Bell (00:37:35): yeah that’s a feature I mean and that’s exactly the thing I’m talking about it’s like okay cool feature and you know they’ll just put a few engineers on that and just bundle it right in you know you have to be really careful as an investor that’s why I do a lot of vertical or new category like you guys like hey this is new it’s so new that it’s going to take Google a long time to figure it out and maybe they’ll never figure it out and maybe they’re not really incentivized to help you figure it out either you know yeah that’s actually a big thing

Jochen Madler (00:38:03): right so yeah so in my opinion like actually we are safe in a sense at least from the foundation labs because we are optimizing sort of against their system or against their objective function to serve users and so Google cannot offer I mean it’s the same with SEO services they are doing their best to monetize with ads right like and they’re you know they’ll give you the AdWords tool they’ll give you the keyword kind of tool and stuff but they’re not going to tell you exactly how to like game their system right because their customer is the consumer right they want to make sure the consumer has a good experience not the business has a good like the business is like a means to an end of like paying for the consumer experience but the consumer is the customer well consumer is kind of the product too in this equation but yeah yeah I mean they cannot if you have a platform and you’re selling ads anyways you cannot have this organic algorithm and put one company above the other on your platform right like that so I think it’s not happening and so that’s not a problem.

Brian Bell (00:38:55): Well let’s talk about the future walk us through a world where AI agents fully replace search what does that customer journey look like?

Jochen Madler (00:39:02): So actually we had this happening to one of our customers it’s another YC company two weeks ago and so they are selling are using some sort of API software, like a wrapper for certain go-to-market APIs. And they told us they got their first user from OpenCloud without any human in the loop. And so what happened was that right now OpenCloud, if you configure it correctly, it can do this already. They had an API. It found their service via web search. It found that it had docs, so it visited their docs page. And then there was a free authentication on it you needed to install. And so it did. It authenticated. It started using the API and draw the credits until it’s empty. And so they got paid usage from without any human loop. And I think this is crazy, right? And then they came to Sidefire, of course, like, hey, we want to know more. And now we’re ramping this up.

Brian Bell (00:39:51): Yeah, we got to figure this out. I’d love to get an LP to invest in our fund without talking to them. That would be amazing. Not that I don’t want to talk to LPs. If LPs are listening, I want to talk to you, but you know, that’d be great, right? Like, oh, I just got like a 250K check into my fund and I didn’t even talk to this LP. The agent just kind of did that.

Jochen Madler (00:40:09): And from a perspective of the end user who is using this, you don’t really care for many things. You don’t really care what service exactly you’re using, especially if it’s APIs. You want to have it robust, you want to have it cheap, and that’s it. I think for humans we have this human shopping experience where we have these review like the checkout wasn’t great or like the product wasn’t great actually it was cheap but it didn’t last long and so the same thing in my opinion will be for agents because you can have a great onboarding experience but if it’s like super cumbersome to have these tools and they are down every like third time you call them actually you don’t have a great agent experience so in my opinion a really cool thing would be if there is like some sort of you know this claw hub right where open claw people or like open claw agents would talk to each other where you have some sort of review forum for like APIs where agents basically report how their experience was using this API

Brian Bell (00:41:12): Yeah, kind of like a moltbook Yelp review G2 crowd for agents I got this massive markdown file of all the products with all the SKUs and all the customer experiences and they’re like oh yeah I searched the G2 crowd like product hunt whatever review site for for this product and it turns out yeah it’s like has good reviews on Amazon but you know six months later the thing falls apart and we know that because all these agents were reporting that into the moltbook like markdown file that’s really interesting what do you think value accrues in the world going forward you got the model layer you got the interface a sas you have infrastructure then you have all the like the raw materials you have the chips

Jochen Madler (00:41:56): so for sure like the chips and like selling the shovels for a long time do you think Sure. And I think like it’s interesting for a lot of software companies to think about if their service, which is true, and we see this with many customers now, especially the ones and the more modern facing ones are no longer inclined to log into a new dashboard. They rather would have an MCP or something. and just use their super apps or codecs or cloud coworker or something to deal with software. And the thing is, for the software providers, you really reduce your whole value proposition to an API that’s being run or searched in the background. And so this is a problem, of course, for so API like software as a service is being eaten by AI and the service you provide will really be become infra essentially because it’s an API and so yeah that’s that’s one thing that’s going to be true and then I think the also if you are then I think scaling and if you’re doing this marketing right like using Sapphire or some other tool to basically be the one the API found this distribution is almost via API it’s almost instant this is why also this marketing to agents is so important because just connecting your API if it’s your API that’s being found and it’s being used there’s no incentive to switch and so yeah really you’re locked in and you have all of this memory which these agents create so I think I think like But the most interesting shift in value will be around SAS. So at these dashboards and these software solutions, I think for traditional infra, I mean, they just stay infra. I think that’s fine. There, I don’t have any strong things or takes or something about this.

Brian Bell (00:43:23): Yeah. Yeah. And it’s weird. I think if you look at the history of tech technological revolutions and as they kind of proliferate they kind of whipsaw up the stack I think I don’t think SAS goes away necessarily I think SAS without AI goes away right I think some people are just going to want to log in just like some people just want to use Google I just want to get like a quick answer I don’t want to wait for chat GBT to like search stuff and like just like give me the answer really fast so I still think there’ll be room for SAS yeah I think it’s going to be a rising tide across all layers of the stack, honestly. That’s what I personally think. If you guys succeed massively looking out five or ten years, what do you actually own and what does that look like? What’s your vision for the future?

Jochen Madler (00:44:00): So right now we are helping companies to be discovered by agents. So basically The next step is, as we do it with this one company, that their API is being found and then integrated by OpenClaw or in the future of other agents. And so this is really just marketing to agents then that their API is used more in or like cells essentially and is being drawn from and so the content that we put out in the web now it’s really blog posts and so it’s still in a human readable form you could go in and like read these blog posts as well they’re not that pretty they’re optimized for agents but still like you could in the future But like these content and webpages we put out, it’s the first thing that agents see when they interact with their brand. And so from there on, once these agents get more capable and cannot only read content but also do something with it, authenticate or something, we can also help brands to not only be the first front door to the site, but then we can also serve these tools on our website to develop this agent funnel. So we can have certain tools on our website that get a quote for our service or authenticate or even do the checkout and we can then and also help these companies basically revert all their business logic which was for humans to business logic for agents which is probably something around WebMCP and I thought in my opinion like the implementation isn’t complex it’s just some code on the website but the monitoring layer is interesting to see because just embedding some tool on your website won’t cut it. It’s really like this is marketing now too, essentially. And so you need to understand where you need to embed it, what works, what doesn’t, and A-B testing, all of this. And so mapping out the agent journey on a webpage, that’s something we can also help with. And I think this is really exciting.

Brian Bell (00:45:41): Okay, so let’s wrap up with a rapid fire. What did you believe about startups two years ago that you no longer believe?

Jochen Madler (00:45:48): that you need a lot of people and need to scale headcount I think now it’s not true you just need a few people with great tools yeah so basically just kind of set up all the AI agentic workflows like oh we need to do this thing we need a SDR to go qualify demos like cool agentic workflow we need another agent to like reach out to people agentic workflow we need somebody to follow up with people when they’re not using the product anymore and check in on them agentic workflow Yeah exactly Is that basically it now?

Jochen Madler (00:46:17): Yeah and I think of course like it’s not dystopia that like no humans will be used anymore but I think like the first inkling of if there’s something to be done it’s now can I do this with AI and if not of course Yeah I mean I’ve had this experience I’ve set up a lot of AI here and I was like a very early user of OpenClaw so it just it was burning through like I was burning like a hundred dollars of tokens a day I was like this is a lot kind of a lot you know I find so far that if I have a contractor overseas that’s pretty well educated and I give them a process with AI to follow I get a lot better outcomes right now than just a pure AI driven I mean that’s not the case if you really really tune the AI it’s like you’re really really good at tuning the AI you get a lot of time to like work with it and it’s very domain specific vertical specific you can get it to do exactly what you want it to do but yeah I mean I can see the promise of that

Brian Bell (00:47:09): What’s the decision you made early at SiteFire that you would reverse today?

Jochen Madler (00:47:12): I think this first layer of following what everyone was trying to do with these actions of basically running a huge regression and trying to distill some pieces that work everywhere, like a grand model of how a geo works. I think this was doomed to fail. But yeah, we tried it anyways.

Brian Bell (00:47:29): I don’t understand the context. What is the direct actions?

Jochen Madler (00:47:31): and so like the idea is if you do a lot of monitoring and then you’re basically trying to see why are these pages ranking and you can basically do a lot of statistics right on this like what are the features and there’s so many studies on this what are the features that are being like used in these pages and this has some value but it’s really like the world is more complex than that and so basically going all in and trusting these agents basically putting in little SEO researchers instead of applying some model this really is something that I would do differently now

Brian Bell (00:48:00): What’s the most dangerous assumption in your current strategy?

Jochen Madler (00:48:03): So I think we are we are banking on the people that AI will win essentially like that people will use ChatGPT, Cloud Coworker and all of this as super apps and so if this is not true and actually that people are still using Google for a lot of things or if they are using platforms such as for example Airbnb it’s quite interesting I talked to the founder and he said yeah we’re not doing this JTBG integration we want our app that people should be on our app right not on JTBG and so if there’s enough platforms that can manage to have the power like this I think within Airbnb there is no optimization to be had right it’s a platform game and so I think yeah this is a risky assumption

Brian Bell (00:48:41): Yeah I think you’ll have both I think you’ll have like kind of the Google and Bing of AI which is starting to coalesce and then you’ll have lots of local experiences right that are kind of locked down right what’s something your customers say they want but you’re just ignoring them you think it’s like a bad idea so I mean there’s like right now since we are working with them on this new thing but of course like once they see there’s some working and there’s like tech developing they also want to have it for SEO and like for the traditional web pages and like nice human interactive elements stuff like this and yeah this is something we need to say no because we need to focus but yeah it’s a short-term thing that would work right now but I’m really sure it’s not working anymore in the future

Brian Bell (00:49:22): What’s a metric that you care about that most people wouldn’t understand?

Jochen Madler (00:49:26): So I think for me, this bot requests, so like connecting with the network logs of the customers and AI referrals from Google Analytics and looking at those metrics. First of all, what’s your traffic to the site from agents? This is really these background searches hitting your site. And then, of course, what’s the click through? They are not directly the same thing. But of course, you can see essentially which pages are working in AI and which ones are driving the traffic. and which ones are driving the traffic. And I think, yeah, this is something I care about because ultimately what spins up this whole agent funnel.

Brian Bell (00:49:57): What’s a recent opinion that you changed your mind on?

Jochen Madler (00:50:00): Yeah this SEO agency thing I think at the beginning we worked with them I think and of course we wanted to learn about the space and I mean there’s a lot of knowledgeable people in there and so we were really trying hard to sell our tool to them like as partners or something and I think now yeah I don’t believe like this is too strongly the case anymore Yeah, because essentially they also had a hard time justifying the price because they obviously want to have their own knowledge and power and work and consulting on top of it. And so I think this is really a channel which we stopped doing.

Brian Bell (00:50:36): If you had to bet against Sitefire, your own company, what would the argument be?

Jochen Madler (00:50:41): I mean something that we get a lot of of course competition you can always that’s something you can bet on and I think the other bad you can do is that like there will be this so I think a risk is or like something that that is maybe true is like that these AI these foundation labs they don’t like their system to be game too much right and so if there’s a company that’s too successful and systematically outperforms too largely on their algorithm they of course like they could do something about it I think they have the technical capabilities I think at that point I think we are we are in a good spot in my opinion if this is true and also the main thing is of course yes making people discoverable but also making this whole agent journey possible so I think back then it’s a thing but I think this is something if you have these companies against you I think it’s really a bad spot to be in that this is a risk that we run but actually we now manage to get them on board on our company some of them so actually we’re in a good spot

Brian Bell (00:51:41): Awesome I really enjoyed the conversation where can folks find you online?

Jochen Madler (00:51:42): So we are like our website is sitefire.ai yeah and I think anything from there we have the same thing on X but yeah that’s really it

Brian Bell (00:51:51): That’s awesome. I’m surprised you were able to get sitefire.ai. That feels like you had to buy it from somebody or what?

Jochen Madler (00:51:58): No, actually not. We spent a lot of time like researching and we were previously were called something else, which was, we really liked the name, but we needed to switch it during YC before we all launched. And we had two intense days of searching domains and then we finally found this.

Brian Bell (00:52:12): That’s awesome. Yeah. I was so, so lucky when team ignite was not, that did not really exist. I was like, cool. because that’s what I wanted I didn’t have to actually do the two days of intense searching but that’s always like a thing where you got the perfect name you’re like oh man somebody owns it they want 50 for it you know and also there’s like this

Jochen Madler (00:52:30): immediate thing of like you think of this great name and then you’re disappointed because someone else thought of the same name right yeah but I think you guys got a good name so yeah thanks for coming on really enjoyed it yes me too thank you so much

That lens has shaped his path from teaching middle school students in the Bronx to running Generous Ventures, a pre-seed fund built around behavioral science and distribution-first thinking. His edge is not access or capital. It is how he filters signal from noise.

Here is what matters from the conversation.

Subscribe now

The Core Idea: “Of Course” Businesses

Mike looks for companies where the value proposition is so clear that customers do not need convincing. Not “interesting.” Not “worth a pilot.” Immediate adoption.

A simple example from his portfolio: a company that reduces radiology costs by up to 80% for self-insured employers. No heavy sales motion needed. The ROI is obvious.

This is the standard he uses. If a founder needs long explanations, heavy demos, or market education, friction is already too high.

As a founder, you can pressure test this directly:

• Can you explain your product in one sentence with a clear economic impact?

• Would a buyer forward it to peers without being asked?

• Does adoption spread naturally within a network?

If the answer is no, you are likely building a “maybe” product instead of an “of course” one.

Distribution Is the Real Moat

Most early-stage investors say they care about product. Mike prioritizes distribution.

He looks for:

• High-trust networks where one customer unlocks many others

• Industries with low competitive secrecy, where buyers openly share tools

• Built-in referral loops instead of founder-led sales

A strong signal: one customer brings in five more.

He avoids businesses that rely on constant outbound sales or long enterprise cycles without natural expansion. That model can work, but it is slower and more expensive.

For founders, this shifts the focus:

• Stop asking “How big is the market?”

• Start asking “How fast can this spread?”

Why Most Founders Misprice Their Rounds

One of the clearest mistakes Mike sees is valuation anchoring.

Founders compare themselves to a handful of visible companies and assume similar pricing. That leads to raising too high, too early, with too little margin for error.

He shared a simple contrast:

• A company raising modest early rounds with strong fundamentals can scale into a $250M+ valuation cleanly

• A company raising aggressively too early must grow perfectly just to justify the next round

The key point: valuation is not a trophy. It is a constraint.

A lower valuation with faster execution often leads to a better outcome than a high valuation with pressure and limited flexibility.

What Actually Matters in Founder Quality

Mike does not optimize for pedigree or storytelling. He looks for patterns in behavior:

1. Speed of learning
Great founders update their thinking weekly. They test, adapt, and move.

2. Clear prioritization
They know what matters and what does not. They can say no without hesitation.

3. Bottoms-up insight
They have either lived the problem or spoken to enough users to understand it deeply.

4. Resilience with direction
Not just persistence, but persistence combined with learning. Grinding without improvement is not enough.

One signal he values: frequent, high-quality updates. Founders who communicate clearly and consistently tend to execute better.

Community Is Not a Buzzword (If Done Right)

Mike has built hundreds of curated groups across founders, investors, and operators. But he is careful with how he frames it.

He does not call himself a “community investor.” He runs a fund that uses communities as leverage.

The difference is execution.

His communities work because they have:

• Clear context (what the group is for)

• Tight curation (who belongs)

• Direct value (what members get)

For example, a group might be:

• Only pre-seed investors sharing deals

• Only CPG founders discussing operations

• Only GPs, no LPs, to avoid pitching behavior

Anything off-topic gets removed immediately. No spam, no self-promotion.

The result: signal stays high.

A Different Model for Venture Capital

Mike also structured his fund differently.

He shares up to 70% of GP carry with LPs who:

• Source deals

• Help portfolio companies

• Support other LPs

This turns LPs into active contributors instead of passive capital.

It also creates a network effect:

• More sourcing

• Better diligence

• More support for founders

For early-stage funds without massive capital, this kind of leverage matters.

The Contrarian Take on Portfolio Strategy

There is an ongoing debate in venture:

• High-volume portfolios (100+ companies)

• Concentrated portfolios (20–30 companies)

Mike sits in the second camp.

His view:

• You do not need massive diversification if you are disciplined

• A focused portfolio allows deeper support and stronger conviction

• A well-chosen company can return the entire fund

The tradeoff is clear. It is harder to execute, but it allows for higher engagement and clearer decision-making.

The Biggest Mistake Investors Make

One belief he rejects completely: that other investors know better.

He has seen strong companies passed on early, then oversubscribed later. The signal was always there. The market just missed it.

This leads to herd behavior:

• Investors follow brand names

• They chase momentum instead of conviction

• They optimize for safety instead of upside

His approach is simpler:

• Understand the fundamentals

• Make your own decision

• Accept that you will be wrong often

But when you are right, it matters.

Final Thought

Mike’s long-term goal is ambitious: become the first person founders and investors think of when it comes to pre-seed deal flow.

Not by controlling access, but by creating so much value that going through him becomes the obvious choice.

That idea ties back to everything in this conversation.

The best companies, the best founders, and the best investors all share one trait:

They make the right decision feel obvious.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:

00:01 Introduction and Guest Background
00:32 Mike’s Origin Story and Early Career
01:29 Transition from Teaching to Venture Capital
02:37 Building Communities and Entering VC
03:50 Fund Strategy and Investment Thesis
05:02 Evaluating Distribution and Market Dynamics
06:05 Customer Concentration and Market Size
07:15 Behavioral Science in Venture Investing
09:06 Lessons from Early VC Experience
10:47 Founder Traits and Decision-Making
11:42 Follow-On Strategy and Portfolio Management
12:40 Community as a Competitive Advantage
14:32 Fund Structure and LP Incentives
17:02 Building and Managing High-Value Communities
19:53 Experiments in Founder-Investor Connections
22:21 Personal Story and Resilience
25:14 Evaluating Startup Potential and Risk
27:51 Portfolio Strategy and Diversification Debate
31:56 Investment Philosophy and Decision Frameworks
34:26 Founder Mistakes in Fundraising
36:42 Secondary Markets and Liquidity Strategy
39:28 Founder Priorities and Non-Negotiables
41:38 Learning, Reflection, and Continuous Improvement
44:16 Network Effects and Deal Flow Growth
47:00 Identifying High-Potential Founders
49:31 Changes in Venture Capital Landscape
51:23 Staying Relevant as an Emerging VC
53:25 Tools, Systems, and Personal Workflow
55:16 Vision for the Future
56:54 Rapid Fire Questions and Closing

Transcript

Brian Bell (00:00:56): hey everyone welcome back to the ignite podcast today we’re thrilled to have Mike McCombie on the mic he is an investor community builder and behavioral scientist enthusiast based in New York City Mike has spent years helping founders navigate the early stages of company building from entering at Techstars to investing with next-gen venture partners and building venture communities across the startup ecosystem, which is how I got connected to Mike in the first place. So we’re happy to have him on. Thanks for coming on, Mike. My pleasure. Thanks for having me on here. So I’d love to start with your origin story. What’s your background?

Mike MacCombie (00:01:25): I did not expect to be in venture catalog. I can say that much I am the son of two child developmental psychologists who became a teacher to middle school students with special needs in the Bronx for four years then became a behavior science based consultant and then I joined a VC fund built out their ecosystem and decided to start my own fund after that so I didn’t catch it you were teaching in the Bronx you said teaching in the Bronx students with special needs in middle school so that makes I work about seven days a week on this role this is nothing in terms of the emotional labor compared to what I had back in the day so this is relatively straightforward

Brian Bell (00:01:56): Yeah, I taught in teaching fellows program when I washed out of Wall Street 20 years ago. So I was taking the, I forget which train it was, it was probably the sixth train. I was in the Upper East Side. I taught a semester of math in high school and I was like, ah, this isn’t for me.

Mike MacCombie (00:02:08): So I could just imagine. It is a particular personality and mindset to go into teaching. I could say that for sure.

Brian Bell (00:02:14): Yeah, that’s amazing. And then how did you land a gig in venture? I mean, that’s pretty impressive. It’s hard to do.

Mike MacCombie (00:02:20): Took some time. So you can imagine being a teacher trying to get into tech, you get a lot of eyes glazing over and saying, oh, maybe you could get a customer success job at an ed tech startup or something. And I said, well, that’s not what we’re doing. So I ended up hosting a large number of events for people where it was actually set up in a way that it wasn’t about networking. So it was curated by personality for positive, pleasant, trusting, generous, and open-minded people. Couldn’t bring business cards, couldn’t talk about work for the first question, and you couldn’t move on from a conversation just because they didn’t have anything professionally in line with you. So that event for 55, 60 events across from 20, 16 or so until 2019, I guess we’re going to 2020. And I built a bunch of other events. And so I got really good at ecosystem building. And I met a firm that was looking to hire somebody as an ecosystem builder. So I sort of went in on a functional expertise that built me into getting exposure to the investing side. and then that got me into the investing way in the other direction where now I’m running my own fund.

Brian Bell (00:03:15): That’s amazing. And is this your, what fund number are you on? What’s the thesis and stuff like that?

Mike MacCombie (00:03:20): Yeah, so I’m on fund two. Fund one was an extremely stealthy $3 million fund investing in priests and seed. The thesis for me is I spent two careers finding out what we get people to say yes, of course, without much friction, without much pain, and getting them to remember, to share, and to act. And so taking the same thesis, I said, okay, where are the companies where we actually can get an of course decision from customers? So when I say my fund is focused on is finding the truffles, places and ecosystems where there’s dense value, clear customer demand, distribution-led businesses that are able to ramp up their go-to-market strategy with very lean expenditure, basically own their ecosystem. I focus a lot on vertical-specific AI, and I’m not talking legal and health, but saying radiology, food safety, simulated medical education. The two companies in my second fund so far is an AI data layer for animal health and customs compliance for the supply chain. So I go for very specific ecosystems where I believe somebody can own it and build up the data, the orchestration layer, and or the embedded fintech pipelines that allows them to avoid anthropification over time.

rian Bell (00:03:24): I love that I guess the word I would use is category defining I was a category manager at AWS and Microsoft I ran all the categories and so I borrowed that kind of term is that kind of how you think about it as like an ecosystem I’d say owning

Mike MacCombie (00:04:35): the ecosystem you might not be defined in the category you might be the second player in but if you can completely own the ecosystem by getting to everybody faster there was a company I was looking at in 911 dispatch calls there was a company way ahead of them but they actually found a go-to-market strategy that was faster to get in to own all the customers before the other ones could get there There’s a company that I’m looking at investing in a space where there’s a competitor that’s raised at a half billion dollar valuation, but that’s on 700 customers and this one has distribution leverage to 36,000.

Brian Bell (00:05:01): Revenue multiple on the 500, do you know? About a 10x.

Mike MacCombie (00:05:05): Okay, that’s not too bad. Not unreasonable. But again, that’s going off of a 700 base. So it’s sort of where somebody can own the distribution and own the ecosystem. They don’t have to be the first one, they have to be the one that has the most.

Brian Bell (00:05:16): How do you assess this? This is something I kind of struggle with. And it’s a question I always ask founders as I interview them. It’s like, hey, what’s your distribution wedge? How do you assess that as an investor?

Mike MacCombie (00:05:24): So I look at a few things. I think about why people will say yes across the board. And it’s usually thinking about the competitive dynamics within an industry or thinking about the sort of social sharing dynamics. So I usually look at high trust, high velocity or high social proof customer networks, meaning if one is in of any category, the rest will come in. Or if one big name is in, all the other names will come in. or they have an abundant mindset when it comes to sharing who they use because there’s not a hyper-competitiveness between each of them. So if we’re looking at, for example, food safety, nobody’s really going to try to say, hey, we’re not going to share with you how we keep our food safe. They say, yeah, this is what we use. This is great. We recommend it. So it’s a lot of an abundant mindset when it comes to the ecosystem. I also look for high referral coefficients rather than founder-led sales and hand-to-hand combat. So I’m looking at a combination of ACV sales cycles and sort of touch requirements in order for somebody to get that customer through the pipeline. Ideally, the perfect case is

Brian Bell (00:06:21): How do you think about that in terms of buyer concentration? I was just looking at a railroad startup this morning where there’s like six big railroads and then there’s like maybe like a hundred more small ones and so it’s a pretty concentrated industry. Would that kind of turn you off as an investor?

Mike MacCombie (00:06:34): I don’t mind customer concentration if the contract sizes are big enough. I think the one that we’re both talking about, the revenue didn’t have a way of scaling exponentially after the first customers. In fact, when you look at the distance from the rail stations, it actually decreased. So it was a very geoproximate way of looking at how they got their customers. Customer concentration, I don’t mind as long as it’s big enough. So I’ve had companies that have sold to the government. I have companies that sell to states. As long as there’s a play where either it’s a large enough contract that their cash flow is abundant or there’s a data play where somebody else wants to acquire them for their building. So I prefer not to have it be somewhere where there’s five total customers. But if there’s a long enough tail or a big enough contract size, it’s not a stress point for me.

Brian Bell (00:07:18): So what is the behavioral science background and how do you apply that today in venture capital?

Mike MacCombie (00:07:23): Well, for me, it’s very basic first principles thinking. Instead of saying, oh, this founder’s great, they had a background in this, I’m saying, okay, what is the actual reason somebody will buy? Why would they use it? Why is it an obvious business case for them as approximate to revenue or approximate to savings. And so it’s really, it’s a way of discerning away all of the fluff. My companies don’t need hype to succeed because their value prop is clear enough and apparent enough to the customers. So my best performing company, my first fund was in the radiology space. They had no press until their series A. And I can’t say the numbers, but they are increasingly growing even on an absolute basis without really needing any hype. because their value prompt was, we can cut your cost of imaging spend by 80% if you’re a self-insured employer. You don’t really need to have a lot of argument for it. So it’s getting to the fundamentals. I think the right companies are not trying to fight friction. They’re not trying to use PLG to get in and sort of prove their case out.

Brian Bell (00:08:55): That’s amazing what did you pick up I mean you’ve picked up a lot of how to’s it sounds like on investing from your early experience because you kind of if I understand it correctly you were I would say a lot of my learning was self-taught, but I think I’m grateful that I had a lot of reps and looks at where we saw successes and where we saw failures. And the companies where there was the most clear value proposition from the beginning

Mike MacCombie (00:09:35): saying we do this fundamental thing that you absolutely need to do at a cost-efficient basis. That’s where they go. I think a lot of the things that I’ve learned lessons from across the board from all of my experiences play in. I do not follow on to save a company. I follow on if I think that company has a higher potential than anything else. I do not chase good money after bad I do not try to stick to a pure rule evaluation if I think a company could still achieve massive returns then I will move to that company and put money in it’s sort of seeing enough reps other things I’ve seen is a founder having a A deep passion and inability to run through walls is difficult if they don’t have high resilience or high focus. So I’d look for the companies that can cockroach it out if they need to, that have high iterative speed and that are learning faster than I could teach them anything or learn myself as they go through. The companies where it just feels like every update they’ve learned something new and they are completely self-sufficient, but they also know how to take leverage of the things going from their investors and from their supporters.

Brian Bell (00:10:25): Yeah, that’s when you know you’ve made the right choice, right? When you’re getting the monthly or quarterly updates, you’re like, these guys don’t need any help. I don’t even know.

Mike MacCombie (00:10:32): There’s one company in my first fund that was amazing at their customer updates. They knew how to ask for help, but they made it so ridiculously That’s so easy and so particularly pointed that you say, well, this is a one-click solution to help them. They would shadow the people that did the work. So I don’t think it’s the ones that don’t need the help. I think it’s the ones that know how to properly leverage the help that they can get or the help that is available. Somebody might be completely independent, but they could still find customer serendipity. They could still find talent serendipity as they go through. Yeah, I love the ones that are completely independent, but if I can still help, I want to make sure that I’m adding to what they can get in enterprise value.

Brian Bell (00:11:03): Yeah, that makes sense. I want to drill in on the follow-on. So on a $3 million fund, I’m guessing you weren’t reserving much capital. How much did you decide to reserve, if any at all?

Mike MacCombie (00:11:13): I ended up investing in only five companies in follow-on. One of them I tripled down on. I actually basically quintupled my total dollars invested into the company. I’m very grateful that that’s my best performing company that theoretically could return my fund about 8x if it goes as I think it will. So I I was very conservative I say I follow on when a dollar’s into that company is a better bet than a fresh new company either based on the information asymmetry that I have or simply at the price it is it’s being underappreciated or at the trajectory that I think it can take at some point you know if you’re looking at a hundred million valuation on eight million revenue that’s good but how much they ramp up is simply dependent on what you think they can get to so I’m conservative of follow-ons frankly because I’m grateful to see a lot of good opportunities as first checks and unless I’m deeply convicted and nobody else is buying in I allow that one to continue But it does make a huge difference when you’ve got a big investment in a company that’s doing well.

Brian Bell (00:12:02): Yeah, that makes sense. So you seem to be following a network-driven model for investing. You know, a lot of investors focus on capital, but you focus on community. Why do you think community is such an underappreciated asset in VC?

Mike MacCombie (00:12:14): Well, I will call one thing out. I think I’ve seen a lot of community-driven and it sounds fluffy. I do not ever call myself a community driven fund. I say I am a fund manager who happens to run a lot of communities that are deeply high leverage for what I do. Because it’s fluffy, I don’t think people see it. If they’re a lead investor, community doesn’t matter as much. You have a few GPs that you share deals with, you take a stab at. For me, because I’m a second check into any company that I do, I take a highly federated approach. So actually the first two companies I’ve done, one there and one there, both came from investor referrals from other companies through communities that I run. And so for me, I don’t mind being the first thought as the second check. I have been the first company, the first check into some companies in the past, but that’s not a necessity. It’s my strategy. Like my fund size is concentrated at a small size, but even as a $10 million fund too, that I’m working on 250K check, it’s usually not going to box somebody out who’s coming in. And so it’s very easy. And I incentivize all my LPs in my second fund to support the companies. I give 70% of my GP carry is up for grabs between me and my LPs. if somebody sources they can get 30% of the carry on that company once the whole fund doesn’t carry if they help they can get 30% if I do no work at all and nobody else helps and if they help other LPs they can get up to another 10% so effectively they are my federated approach of saying look if I have a pricing strategist if I have somebody who is an expert when it comes to alternative lending sources somebody who’s we all get the leverage there and so being on the small side allows me to play well to get into the allocations and to frankly be somebody that they want to come in because I’m providing a lot more leverage for the dollars than most people would

Brian Bell (00:13:44): yeah that’s really cool something you know team ignites in our name we were also network driven part of the reason I want to come have you come on the podcast just to learn from you but it about also just compare notes and I like this idea of sharing carry how do you kind of execute that like what are what’s your who’s your fund admin how do you actually do that at the GP level well there’s a few things

Mike MacCombie (00:14:03): one it’s the incentives are for them to report so the actual data gathering if they don’t report I get to keep it so the incentives are aligned to and there’s a way of Diverting carry into different entities As a portion and at the end of the life of the fund You can basically share out shares of those LLCs That allows you sort of to effectively share the carry Without having to do it all directly AngelList is my fund at the moment They’re fantastic in many ways And they can do deal by deal carry I think They can do deal by deal carry on the sourcing side But it’s hard to do post facto of down the line I want to actually add in some more carry here and there That’s where you sort of think about the shares Diverting into different entities For parts of the carry

Brian Bell (00:14:43): Yeah, I famously broke away from AngelList three years ago because they were cross-marketing to my LPs, competing funds, which I did not like. But I did like their administration. I thought they had really, really good fund up in. I ran my first fund with them, a little rolling fund. Yeah, you can do the deal by deal carry and they just make it really easy, right? It’s just all kind of done for you.

Mike MacCombie (00:15:03): Yeah, it’s very straightforward, especially when I have 229 LPs in Fund 1 and I’m going for 240 in Fund 2. I’m not doing those K1s on my own. I’m not doing the distribution. No, yeah, it’s just too much. It’s the perfect size for me. I don’t need to go larger.

Brian Bell (00:15:15): Yeah, that’s amazing. So you’ve built and hosted hundreds of these events and communities. What separates a community that thrives from one that quietly dies?

Mike MacCombie (00:15:25): Well, I think there’s a few things. The biggest thing, and as fundamental as it is, is setting the proper expectations. I will have communities where I tell people, And I sort of set those I think the other reason that communities, even when expectations are set higher, they don’t function in that way, is because they fail on one of the fundamental three things. You need to either have a very clear context, a very clear curation, or a very clear overt value created from the community. this is where we share pre-seed deals this is where we share best practices for CPG founders it is only these kinds of things that are allowed as soon as people start posting and plugging their own events their own newsletters and saying hey I wrote this article please like it on LinkedIn I delete with an iron fist in terms of like if it doesn’t fit the group and it’s clearly not valuable for everyone to see it gets deleted so at most I see maybe 30 groups a day that are posting I’ve run 322 groups as of right now and so that works in terms of curation it avoids temptation so for example I have a GP focused group no LPs allowed in fact I had one person who was both and they sort of said hey can you let that person go and I have an LP group no GPs allowed because otherwise you’re going to want to go in and directly pitch every single person etc and so having a pure curation provides that everybody’s there for the same reason even having a this is all pre-seed founders no they don’t have enough curation overlap in terms of what’s meaningful CPG founders they talk all the time because they’re all dealing with the exact same issues right And then the last thing is overt value. If I have a group where people are just sharing poker-based events for poker players, that’s overt value. It doesn’t really matter what it is. If I have a group where I’m sharing referral codes for something where everybody gets a discount, it’s overt value. If you have something that’s worthwhile enough If I miss it for more than a few days and it takes forever to catch up on them, I don’t need that either.

Brian Bell (00:17:35): Yeah, I got a lot of these WhatsApp groups as well, including a couple that you’ve set up. And typically I’ll just turn off notifications and every once in a while I’ll dip in there and see what’s going on. But yeah, it’s hard to keep up with the fire hose of everybody’s comments and questions.

Mike MacCombie (00:17:50): Nobody needs to say, I’ll DM you or I’ll be there. No, just DM them or put an emoji on the message and that’s fine

Brian Bell (00:17:56): Yeah, totally. So you’ve run lots of experiments helping founders raise by connecting them with lots of investors like myself, right? I’ve done a couple deals with you over the years now. What inspired those experiments?

Mike MacCombie (00:18:08): I am a big person when it comes to limiting beliefs about what’s possible for So long story short, mom died when I was young, I felt very helpless and I figured out what are all the ways that I can have agency in the world and one of the best things is by people telling me what’s not true and proving that actually people can behave differently. So anytime somebody tells me, oh, people won’t be vulnerable with strangers, disprove that. and mystery trips to other countries without knowing where they’re going. Disprove that. People won’t go to an event if it’s not focused on the industry that they want. Disprove that. So every single one of those is a challenge. And actually I worked at a fund where they pass it with other people. It’s bad signal. I said, well, I don’t claim to be the arbiter of perfection. Everybody has this belief that if you’re a third of the time wrong and a third of the time sort of zombie companies and a third of the time right, you’re good. I don’t assume anybody is the perfect knowledge so I started a group of people shared deals even the ones they’ve passed on I’ve done deals that people have passed on people have done deals I’ve passed on so to say we all underwrite different things I’ve passed on unicorns people have passed on companies of mine that are on the way to be unicorns like we are not that perfect the best thing is people who are true to their game that they are under able to underwrite a portfolio of enough companies that have shots on goal so the long story short my view is tell me something that people can’t do and I’ll say there’s probably a system where we can get them to enjoy doing it

Brian Bell (00:19:19): My mom died when I was young too. I mean, not too young. I was 16, but how old were you?

Mike MacCombie (00:19:23): Eight.

Brian Bell (00:19:24): That’s a little younger. Yeah, my sister was about that age when our mom died. And yeah, it’s hard. I think a lot about like how much that drives and motivates me. And it drove and motivated me so much, you know, to get into the wrong career early in my career. And so I kind of put the cart before the horse and I ended up on Wall Street, like really hitting my life, right? And I was like, you know, I studied for the CFA and... I got there you know I made it I’m making good money and I’ve really hated my life how do you kind of prevent kind of the negative effects of a traumatic event like that and how do you think it motivates you today well I think it’s probably about your narrative of what you’re going to do about the world I think you look at the way that I see superheroes and villains defined as they both go through trauma the villains say I went through it so everybody else should and superheroes say I went through it nobody else should and so it’s a way of saying that I think it’s also how one finds their way through what’s the narrative like I’ve I will admit I’ve probably always been a pretty optimistic person and I will say I was deeply affected by going through what I went through I personally was very grateful that I had a very strong community in my school that came around me when we needed it so I didn’t feel like I was going to be completely alone but I think it’s somebody who’s able to reflect take in appreciate and be grateful for what they have and who are able to say that was terrible What do I do with it? And how do I respond? Rather than saying it was terrible, my life is ruined, nothing else can go. I have navigated more things that are high stress and high cortisol inducing over the life. And every time I realize I’ve still gone through every single time. It may not be perfect, but provided people are alive and that I’m not affecting my health.

Brian Bell (00:20:50): I’ll make it through.

Mike MacCombie (00:20:51): And so I would say psychologically resilience is probably the highest factor of founders in terms of any trait that leads to success. You can have a resilient founder who’s beating their head against the wall and not learning at all, but one who gives it two tries and one who gives it 10, you have an increased likelihood purely by the number of at-bats that somebody’s taking.

Brian Bell (00:21:06): Yeah, it seems to be, it’s not the only defining feature of very resilient founders, but it seems to be very common amongst high performance founders, right? There’s some traumatic experience that they’re overcoming. Something put a chip on their shoulder. And I think, you know, people like us that have been through trauma early in our lives, I think it gives us an edge as investors because I think we can spot it in people a little bit better.

Mike MacCombie (00:21:28): Yeah, I think we can spot it. I think, well, one, we can see it in somebody. I think you can only see other people as deeply as you see yourself. If you haven’t gone through traumatic events, you don’t understand the impact of it. You don’t understand the layers that it takes. And once you’ve reflected enough on yourself, you also are able to see those reflections in somebody else. So if you’ve gone 10 layers DPC 10 if you’ve gone 100 you can see 100 and so that’s one I think two is also there’s a hunger and there’s sort of like a non-negotiability when it comes through it somebody who hasn’t got through much might be they face less and tribulation where they’re they don’t know that they can get through and they don’t know what it looks like to push through and I think the founders and the investors who have gone through that kind of trial know one that they can persist they know two what it means to persist and they also know three like what is worth persisting for once you’ve gone through some part of my language like you’ve gone through some shit you’re very selective about the things that are worth your time and so it’s a it’s a high discernment that comes through too and frankly prioritizing what matters I would say goes into it as well

Brian Bell (00:22:24): Yeah, I love that. One thing that Sam Altman said recently that’s always stuck with me as an investor is when you look at a company, you’re not trying to think about what can go wrong. You’re trying to think about what can go right. That always just stuck with me. What do you think? What do you think of that phrase? Do you agree when you look at a company? Yeah, you’re kind of thinking about the things that go wrong. But if it goes, are you trying to think about, okay, the optimal Rosie scenario, if everything goes to plan and the vision comes to reality, is it like a multi billion dollar company or not? Like, what do you think of that phrase?

Mike MacCombie (00:22:51): I think there’s a way of doing both. I think there’s about eight different ways that founders are underwritten by investors for the profiles of what investors look for. Some go pure traction, some go for pure hustle, some go for pure P&L analysis, some go for pure product. some go for is this a history making idea some go for pedigree some go for distribution some go for first principles every single one of those can be underwritten differently in terms of like East Coast mentalities West Coast mentalities and so for me when I think about it I care I don’t care about the fact that the TAM’s not too small, that they can get to something that’s massive and acquisitively interesting. But I also don’t want them to say like, this is going to be a pure binary bet. I’m not going to do asteroid mining on the moon. I’m not. I looked at a company. I love the founder who was building an underground pipeline for delivery networks, kind of like what Elon was doing, but purely for packages. It was great. That’s a binary bet. It’s not a whole lot of middle ground where you’re going. And as a pre-seed investor who’s not trying to have an insane amount of dilution for a binary bet, I like the companies that are a little bit more capital efficient as they get there. But, like, I have companies in my first fund where I was running concentrated enough checks that I could underwrite to a $200, $300 million valuation and have a fund returner. I’m very grateful that one of those companies is looking at a $2 billion valuation in probably a year or two. Like, okay, we’re We’re happy. So I think if I underwrite you a base case where I need it to survive, but I also am not going to try to go for something that’s just like a good triple or a good double. I say it’s a solid triple with a concentrated enough value. With a concentrated enough check, a triple becomes a home run for you in terms of fund returning. And so that’s why I don’t do Spray and Pray. There are some funds that have 500 companies per portfolio. There’s some that have 70. they’re great funds my approach is I like to focus on the fundamentals get to the first principles and say if I can find 30 companies that are first principles based very interesting targets not only will one I have ones I can hit it but two I don’t feel like I get a billion dollar outcome and say oh man I just really need them to get to 10 billion in return the fund because I’m so spread that’s interesting.

Brian Bell (00:24:45): I think I approached it from a first principles like CFA brain I kind of said okay if I was making a portfolio of early stage bets, like what would be the expected return based on portfolio sizing. Assuming I have average picking ability, average ability to win, average ability to add value, because everybody thinks they’re above average, right? And so the number I came up with is, you know, you actually don’t you don’t asymptote your expected return until you have hundreds of early stage positions. Now, that doesn’t mean you can’t have a concentrated portfolio and win. I just think it’s harder. It’s much, much harder. Especially if you’re someone like you with all these groups and this ecosystem and you’re getting a massive amount of deal flow and you’re having to say no X more than me, right? Because you’re only making 30 bets. That’s 10 per year, call it, right? About one per month.

Mike MacCombie (00:25:36): Well the nice thing is one I’m getting into ones of the early enough valuations where I have that opportunity set and two you’re right like I would love I know funds that I’ve done in-desk approach and they can guarantee like a two to three X but also I know it’s way higher than that never I’ve never like the data I’ve looked at and I’ve pulled a lot of data on this and I’ve written like 10 articles on this I’ve done a lot of research Anybody with over 100 positions in their portfolio has never had under a 10x TVPI fund. Ever. Never happened. Nobody I’ve ever met. And I’ve met like two dozen at this point. They all have like 10x TVPI funds. Interesting. I can think of students. If you find a counterexample, let me know. But I’ve never met anybody and I’ve pulled a lot of research and data on this. If you think about it just mathematically, there’s an expected return of like any particular asset. And then the more assets you add to a portfolio, there’s an emerging quality return. of expected return, depending on how many assets. And there’s an asymptote, right? Like for like public equities, it’s like 40. Like 30 to 40, you’ve kind of asymptoked, you’ve kind of diversified away the market risk. For early stage, it’s like three to 500. That’s where you’ve actually like, there’s no like additional like diversification benefit.

Mike MacCombie (00:26:42): I think it’s also depending on how one’s approaches when it comes to sourcing. So if you have access to every YC company, you will have a good returning fund if you decide to get in every single one, especially if you can get the assets really out.

Brian Bell (00:26:51): Yeah, YC does 800 a year, right? They’re doing 2,400, 3,000 of fund now.

Mike MacCombie (00:26:56): Right? Yeah. But they’re also getting in early enough in terms of their ownership where they can get a meaningful return to them. Like there’s a fund that I know that partners with every accelerator in every studio and gets their companies pretty much at the same terms. They’ll return. I don’t think they’ve had a 10x TVPI fund yet. And I think there are a number of funds in. I’d love to meet those guys But I think you also see I think any angel investor will say you have to have at least 20 bets to have a shot at getting a fund returning portfolio or a portfolio returning portfolio at that point and my view is as a solo GP who is very grateful to have seen about 20,000 companies over time Yeah I could easily do sprinkle checks into a bunch of companies but in terms of where I’m best and I like to provide a lot of leverage for the companies I support for me the right size is that I think when you look at some funds that are massive AUM they have to be a lead check they do not have the time for the board meetings so at some point logistics say take your bets I think you could also point to companies like I think Equal Ventures for example I want to say they have like 15-20 companies per fund and they’re deeply thesis driven and I’m sure their returns are probably doing just fine I would also say I probably don’t think 2048 Ventures is a deeply diversified portfolio they have the particular realms that they go for the view I’ve come across over time is there’s going to be a thousand ways to approach it and I think you can be successful you just have to be really good at your approach and there will also be for every approach will be 100 people. There won’t be 50 that can’t even return the median returns of like we got all the money back. So every strategy can be done successfully. Every strategy can be done poorly. Some might be a little bit more successful, but also they’re harder to implement in terms of one, fundraising and two, sourcing and in terms of three, portfolio support. That doesn’t mean it’s not doable. Like I have a friend who’s a fund manager who also invested in my fund. We both look at each other’s portfolios and say, I would not invest in what you did, but I know why you did it. And he said, I would not invest in what you did, but I know why you did it. So I just learned to say, be true to oneself.

Brian Bell (00:28:41): So a topic you talk about a lot is the founder decision making. Where do behavioral biases show up most often in startup building?

Mike MacCombie (00:28:48): I think there’s a few things. One is external anchoring to context. So if somebody has too many data points of SF side and they say, oh yeah, I should be raising on a 50 mil evaluation pre-product pre-revenue because I know three people who did it. They’re anchoring to an external and local data point rather than saying across the entire network. And I think there’s a lot of decision making of going for the wrong things in the beginning saying what is the proper path of how I raise compared to my revenue profile profitability to reduce the likelihood of me being cut short on my at bats. So I will see companies, they raise too high with too little capital and they don’t figure out that the price you raise it does not matter. My best performing company raised 1.4 million on a 5 million post and then they raised 3 on 18 and then they raised 30 on 250. They’re getting back to the point. I don’t think they’ll need to raise again before the IPO. So the founder took, you know, we’re talking 30% dilution plus 17, so probably 60% dilution, but that’s all probably he’ll need to take. So I think one is the anchoring to that. I think the two is in terms of what they prioritize, being too attached to the solution rather than the problem. They’re stuck in their iteration loops. For example, I can think of a company building an AI pin that you might know about where the founder was too anchored to their own decision making. And I recall from team members saying you could not make a decision without the CEO’s approval, but then the CEO wouldn’t make themselves available for three weeks at a time. So you just were not iterating a product fast enough. And so I’ve seen inability to properly hire and source and delegate teamwork, anchoring to external biases. And I think also not being able to face reality when they say, hey, I’m raising this round slowly. I have a key outcome that requires me raising quickly. And they say, but I can’t raise at a lower valuation. I say, look, you’ll get this raise done in half the time with a 20% cut in your valuation. And making the smart layup of saying, what is the right choice for the health of the business in the long term and the growth that I can achieve?

Brian Bell (00:30:25): Yeah, I remember I was an investor in 2021. Back then I was just running a syndicate. I hadn’t quite launched my fund one yet. And I was seeing these like 50 caps pre-launch, pre-revenue, particularly out of YC, I remember. And I was like, this is crazy. And now I think we’re at a whole new level of crazy in 2026 as we record this.

Mike MacCombie (00:30:44): We are at a level of crazy, but it’s interesting to see because I’ve gotten some feedback from people that have invested my funds saying, like, you’ve got to make sure you’re not too disciplined about price. You’ve got to be willing to go for the ones when it can turn up. Like, yeah, Stripe at $70 million looked pretty aggressive, and now nobody’s regretting getting into Stripe at $70 million. And so I think keeping an opportunity... Right I think it’s fundamentally underwriting to what can this company become and does that justify the price but also some people just completely forget to think two rounds ahead and saying hey if I raise this now what’s going to happen in two raises if I’m expecting if I’m at 50k ARR right now and I expect to be I have to really hope for massive revenue inflection or deep buy-in from convicted investors together.

Brian Bell (00:31:33): I definitely see this in my portfolio because I invest the gamut. I’d invest in the five cap stuff that you’ve talked about, but I also invest in the 30, 40, 50 cap stuff out of YC because I have the two funds. I have a seed fund and a YC fund and I’m investing in both of those parallels. It’s pretty interesting because I’ll go from meeting all the YC founders all raising at 30 caps and they have 5,000 of monthly recurring revenue and then I’ll go meet a company with a million I just put in the offer rate as we’re starting to record this podcast they got a million of ARR at a 12 cap and growing way faster than a lot of YC companies but it’s interesting because I’ve seen this a lot and I have a lot of portfolio companies I am a spray and prayer but that’s why I see a lot right and I see a lot of these you know these 25-30 cap YC companies they raised $300 million Series A’s you know and then they raise a billion Series B or 2 billion Series B I mean if you’ve got a

Mike MacCombie (00:32:26): good secondary strategy Then go for it. If you’re saying, hey, I don’t mind holding it for two rounds and selling it at 10x and do that across every company I’ve got, go for it.

Brian Bell (00:32:34): I’ve been thinking about that. How are you thinking about that as you kind of, you know, your portfolio starts maturing, you start getting some markups? Because I’m sitting on some 20, 30, 40, 50xers, a lot of them, right? And I’m like, okay, it’s 2026. You know, things are getting frothy. How much do I sell when these startups are raising at 200x revenue? I’ll tell you a story. I got into a bunch with my own money back in 2020 and went 33x in like 13, 14 months. I’m like, I should have just sold that, right? Because they raised at like 100x from Tiger and things are great and then they kind of imploded and they’re still around but I don’t think it’s gonna be a 33x it’s kind of impaired at this point but if you know if I would have sold some percentage of that I could have locked in a nice little gain you know so I’m starting to think about that now as you know fun one was a 22 vintage you know fun two to 23 vintage and so on so I’m starting to kind of how are you thinking about that as you look at your portfolio

Mike MacCombie (00:33:29): I think there’s three things to look at when it comes to any company one is trajectory two is future possibility and three is position within a portfolio because if you have a any one of those factors can play a part so like right if the company is far ahead of their revenue valuations and it’s just overpriced that’s one thing and I have a company that was way overpriced but I knew that their revenue was likely going to 50x in the next two years and so I said hey we’ll sell a little bit we’re not going to go for everything because I think I’m going to deeply regret if I do that So I think one is where they’re at right now. I think two is saying, okay, if this is a company that may be a little bit ahead of their skis right now, but I believe it could still get to a Decacorn outcome, I would be a fool to sell too much. I think the third thing is looking at the position of the portfolio. If it is your highest flyer versus, oh, this is my third or fourth place company, it plays a part as well. And so get a look at saying, like, how do I get to a 5X plus? That’s sort of like the baseline of what gets me there. If my best company, I can sell everything to return the fund 3x, but then I don’t have a few followers that have given me the other 2x. And if I think it still has an inflection point ahead of it, I say, okay, we’ll do a little bit. We’re not going to go there. So it’s sort of saying, how much can this contribute to returns amongst the entire portfolio? How much can this contribute to returns given where it’s at right now? And how much can it go beyond? I think when a company is getting at a 40, 50x multiple, I start to look and consider that question. but it really depends I’ve got some companies that I know if they get marked up 10x I will sell all of it and I’ve got some that if they get marked up 100x I’m not selling any of it it really depends on the factor at play of trajectory now trajectory future and position within the portfolio respectively

Brian Bell (00:34:59): Yeah that’s pretty wise you’ve asked founders about their non-negotiables what does that mean I like to see how somebody prioritizes I

Mike MacCombie (00:35:03): I’ve seen way too many founders fail because of them saying, oh, we can do everything. They have too much optimism and trust me, I’ve been the optimistic type and saying, what are the things that you, if you’ve got to cut everything else, what matters most? And if you can’t force rank it, then it’s very hard to make some decisions if you don’t have an operating procedure in your own mind. So like if somebody is non-negotiable on dilution, then I suddenly say they’re going to be too rigid for the market. If they say I’m non-negotiable on bringing our customers deep, immense value. That’s probably more indicative of where I want to be going directionally. They say I’m non-negotiable on a unicorn exit. Then I say, okay, they might hold out for the opportunity of doing something.

Brian Bell (00:35:37): It’s basically saying, what are the priorities and are they straight? And if they,

Mike MacCombie (00:35:41): it’s like when you’re in a relationship, you’ll negotiate on who does this is we’re not going to negotiate on somebody who can rupture and repair in a relationship. And so putting the first things first is deeply important, especially when you are absolutely overwhelmed and you need to prioritize with an abandon.

Brian Bell (00:35:53): Very wise. How did you get so wise for such a young person, no offense? Out of curiosity, how old do you think I am? I don’t know, like 30s?

Mike MacCombie (00:36:02): 37. 37, okay. I’m more older than you. I’ve gotten carded way more times than I would have expected, but I take it as a compliment. Yeah, you look young for your age, yeah. I would say when you run your own fund and you are a high thinker in terms of just repetitiveness of how much like I’ve replayed every deal I’ve done in my head multiple times over to the point where I’m very grateful that most of my LPs I said that my last LP update was 42 pages long and it was an update on every single company roughly about a page each Oh wow Here’s where they’re going right Here’s where they’re going wrong Here’s my take on the company And so I’m able to triangulate across a lot of things I triangulate a lot I have a deep retroactive source of anxiety I’m perfectly fine in the moment But if I think I’m messing up I’m like I’m going to replay it Until I can figure out what do I learn from it So I don’t make the same mistake twice I’m also grateful Both of my parents are psychologists So I’ve always had a little bit of a good metacognition Based context around me I don’t think I’m not going to be single because of a lack of effort I don’t think I’m going to fail because of a lack of effort If I have a lessons learned Or there’s a macro environment factor That’s different But from the youngest age, I always thought that I found the most agency when I learned the most, and so I like to continue doing that. It’s funny, when I tell people the ceiling of my first fund is the floor of my second fund, I mean it genuinely. I expect every company I do, fund one, company one, fund two, company one, I expect company one to be way better than company one, two better than two. And so if I’m getting at this level of returns for fund one, I expect to be at that for fund two just by default.

Brian Bell (00:37:22): That seems to be how it goes in venture, actually. Over time, you do get better, I think, to a certain age. It’s like you look at pop musicians or really great musicians of history. They have that, you know, 10 to 20 year span where they’re just great. Call out the first, you know, four or 5 to 7 funds and then I think like VCs get a little too old and too out of touch you know or maybe they’ve been scarred too much and they kind of get become risk avoidant or something like that maybe the same thing happens with the musicians and they’re seeing It seems to be a kind of a, with venture capital especially, fund after fund, if you’re still in the game, you get better at it.

Mike MacCombie (00:37:59): And why do you think that is? I think there’s a couple things. I think one, there is a hunger. Like it’s very easy to get complacent. You have a couple big winners. I know people at large funds. They have a winner in one company say, great, I don’t care about any of my companies. All of you can shut down and burn out.

Brian Bell (00:38:10): Yeah, I got my taxer and I’m set for life.

Mike MacCombie (00:38:13): Oh, I’ve seen companies that got completely screwed by, frankly, like the Tiger Globals of the world, where they were told to burn, burn, burn, burn, burn, grow. And then they said, yeah, we’re not funding you now that you’ve burned all the money. so one is complacency I think two is beyond that there’s a network information flow effect people who are going through their cohort you know they were the founders at the companies they were the first VCs and everybody that they know for a while there’s a good density of information going there but as everybody ages out some people have kids some people go to operator roles some people decide they don’t want to work in venture anymore some people say hey I’ve got my good I’m going to become a passive investor and other things some people say I want to go to real estate and some people just say hey I’ve got a family now that network that cohort they come in with with that energy density around them loosens up more and more so the only way that happens is if they’re continually repeating the information flows they get in or they’re finding a way to expand outwards of what’s going on around them so for me yeah I have a lot of friends that have gotten married and gone through but I am adding 30 investors a month to my pre-seed deal flow community calls I’m adding probably 50 people to my community groups I’m getting the fresh inflows and I’m building more of an influence It’s sort of funny when I hear a lot of LPs ask, you know, do you maintain? And I say that’s a really fundamental base case. I say, I don’t worry about maintaining because I’m continuously growing.

Brian Bell (00:39:50): Yeah, and I think everything you’re saying is correct, which is, and I want to build on it, which is, you know, over time, I think you get, your network grows as a VC, right? You meet more people, you meet more startups, you back more startups, you add value to the startups, those startups refer you to their friends that are starting startups and I think that that’s why there’s kind of this escalation of ability in venture capital right because like it is just network right it’s just you’re getting more deal flow because you’ve been in the deal flow ecosystem for a long time and so people are just sending you stuff and then you build out your VC community like you’re doing and they send you deals and you’re sending them deals I need to work on the sending LPs deals I haven’t been doing that as well I mean we started as a syndicate so you know I’ve run like probably 200 250 syndicates over the years but I should probably give my LPs opportunities to invest directly I’d ask my founders I guess I’d ask my founders hey do you want me to blast out the deal to all of Team Ignite I can do that you know I haven’t been doing that that’s a good idea

Mike MacCombie (00:40:53): I think there’s a way of going about it that I actually really like is every time I can take a call with somebody, I say I’m wearing two hats. I’m wearing the hat for myself and my fund. I’m wearing a hat for my LPs. And I say, hey, if I think this is interesting enough, I’ll schedule a group call. We’ll get some LPs on and we’ll see if they want to invest. I’m doing six of those calls this week. One we did today, eight investors showed and took questions. And I say to the founders, one, you get to bundle a yes or a no in one call. If it’s 10 no’s, at least you got it done with one call. If it’s 10 yeses, you got that done with one call. And so I value prop that to my LPs of saying, hey, you’re not just investing in me, you’re investing in all the sourcing that I do for your particular interest. And I think people underestimate that. They say, hey, I just want to see the one that they’re doing. I don’t claim perfection. I have deals that people have invested in that I did not do. And so it’s how I work with the LPs that I have. Usually fund-to-funds that are doing it for deal flow, they don’t like that because they say, you’re only going to be good at what you’re actually investing in and what you say yes to and I say then we have a fundamental disagreement because I have companies that my LPs have invested in that I haven’t that have done great and I have companies that I have seen that they’ve done that weren’t even in my wheelhouse that they’ve invested in so yeah I highly recommend it if that’s one’s value prop some people just don’t want it they want the passive returns

Brian Bell (00:42:02): I think it’s I mean I’ve already this podcast has already paid for itself for me anyway is I need to add more of my LPs to my deal flow list so thank you for that idea I will be doing that

Mike MacCombie (00:42:09): Or give them the option to Yeah, like frankly, I love bringing in LPs Like I have somebody who was an early investor in like Galileo and One Medical Any health tech companies like, hey, if you want to join, give some questions They ask way better questions I was looking at a company that was doing a formulation additive to make biomolecules and pharmaceutical products Shelf Stable Without Ever Needing Cold Chain Storage I Would Not Have Had The Deep Questions To Ask He Came In He Was Asking All The Questions I Could Hear His Diligence So It’s Advantageous For Red Teaming It’s Advantageous For Helping And It’s Advantageous For Them Getting Direct Access Which Is Why My LPs I’m Grateful They Like 5x From Their Fund 1 Commitment To The Fund 2 Commitment It’s More Value Than They Would Get From Hiring Their Own Analyst Or Associate Or Principal Or Partner

Brian Bell (00:42:50): That’s Really Smart Really smart. So when you meet a founder for the first time, what signals tell you they might build something meaningful?

Mike MacCombie (00:42:56): I think a few things. One, if I ask them what is making this worth the 10 to 15 years of cortisol inducing emotional labor, they have a reason that’s either deeply curiosity driven or deeply probable. I look for bottoms up customer insight like they’ve either talked with hundreds of customers or they say I’ve been this person I know exactly what I would want and I’m looking to make sure that I have it and I look for the iterative curiosity they say hey we’re going to launch we’re going to give you an update every week of what we’re doing with the pace that it feels like they’re impatient not to get to results but they’re impatient to take action on what they’re learning and so that’s meaningful for me I think also when they have a view as to not just right now but the growth strategy of saying at this stage we would expect to do this and we’re going to roll out to that and they’re making the right decisions of prioritization they say hey I don’t expect this or that I love it when a founder is able to tell me like we’re not doing this thing because of xyz they take a point of view and they take a stand rather than just saying we’re going to own everything I had a founder who said, hey, we’re not doing direct to consumer for our distribution strategy because it’s ridiculously expensive. The only way this makes sense is if we go this way. And I said, yeah, I completely agree. So they see around corners faster. They learn faster from the mistakes. They have an insight from the bottoms up rather than the top down. And they try to learn as quickly as they can for everything they’ve done because they know that that’s their best resource.

Brian Bell (00:44:07): Yeah. Amazing. So you’ve been in early stage game for a while. What’s changed since you got started and kind of what shifts are you seeing today?

Mike MacCombie (00:44:14): I want to say normalcy in pricing. I mean, I remember looking back at deals and it was surprising when we started seeing $2 million rounds. It was always a one and a half on eight. It was a one million on seven.

Brian Bell (00:44:23): I feel like two on 10 has become kind of the norm for pre-seed right now. Is that what you’re saying? It’s like, oh, we’re raising two on 10. That used to be like a normal seed five years ago. It used to be one on five or one on six or seven. And now it feels like two on 10 is like everything I’m seeing is two on 10.

Mike MacCombie (00:44:37): But I think what we’re also seeing is that the outcomes can be bigger. So across the board, when you say, I can see Decacorns, then people are able to underwrite to that kind of valuation. Now, in this market where we’re seeing Sass Hall tools being compressed, you need to have a lower entry point unless you’re building something that can actually outgrow the comparisons there. So I’ve seen that there’s this response to the general market there. I think there’s still shiny object syndrome. There’s still social proof following. I’ve had LPs where I said, hey, here’s a company. I’m not going to do it, but I know Andreessen’s investing in the next round. $500K was put into that company within two weeks. It’s like, cool. They’re doing well, but it was just like, There’s still the following of that in terms of social proofing. And I think people are trying to go to the safest bet rather than the highest upside bet. So yeah, the amount of secondary demand and deal sharing that I’ve seen in terms of Anthropic, SpaceX and OpenAI and Databricks and people either want really early stage or they want really exciting they want really late stage really early or just the ridiculously exciting like I saw a company that was doing a robotics platform for fishermen that found a way to just make better quality fish people did that at like a 140 million valuation so it’s like either really unique in the middle really early or really late stage depending on what their first cultures are

Brian Bell (00:45:43): So they’re more investors than ever competing for these deals. I mean, what do emerging VCs like us need to do differently to stay relevant?

Mike MacCombie (00:45:49): Well, I think one is finding your own game that you go for. Like I could be going for every agentic payments platform for e-commerce. It doesn’t really affect me when there’s a thousand truffle pigs out there because I still know my forest. I know the truffles that I’m going for. I think finding your game, continuing to build leverage within your game. You are great on a broader indexing approach. So continue to be valuable to the people that you want to go across. And for me, it’s finding the highest alignment with my first principles to the point where I’m grateful to say I’ve got six companies that I could say yes to right now but I’m going to pick two because they are the ones that are even in that top six

Brian Bell (00:46:18): or passed on a deal and came back and invested later or do you kind of stick to your guns nope I passed once I’m not looking at this again

Mike MacCombie (00:46:25): Oh no, I would not dare to be saying I am the arbiter of all you can do in the future. The company, there’s a fund that I support deal flow sharing with and I pass on the pre-seed for the company and they’re raising a Series A right now and I shared it with that company and they’re doing it. Usually if I pass, it’s more just like, oh, the valuation’s too late now, even though I think it’s a really valuable company. No, I mean, I have a very abundant anti-portfolio. It’s actually this blue set over here in terms of all the companies that I track. of what’s up there and I track them very clearly and there’s actually these four companies above so these are the two that I’ve done these are the four that if there ever was an SPV opportunity I would probably do them and then the six that I’m looking at for myself so yeah I track them

Brian Bell (00:47:03): What are all the colors for all your post-its?

Mike MacCombie (00:47:05): I mean, some of them are intentional, some of them are not. So I mean, for example, blue over here is all the deal flow sourcing channels that I have. This is the most serendipitous people that I know. Passes and diligence on fund two, fund one, diligence fund one, admin, LPs fund one, LPs fund two, small check, big check, and then sort of close.

Brian Bell (00:47:23): That’s amazing.

Mike MacCombie (00:47:24): Content, long-term projects, and revenue channels.

Brian Bell (00:47:26): That’s amazing. So just post-its all over your house.

Mike MacCombie (00:47:30): I eventually get any affiliate revenue from 3M, but I’m not holding out for it.

Brian Bell (00:47:34): So cool though. Have you looked at digital like Kanban boards that can kind of do this for you?

Mike MacCombie (00:47:39): Oh, I mean, I have digital versions of all of these. So it’s not, there’s no fire, fire, DNO kind of insurance concerns here. But these are more the things where I think if you have a wide aperture on things, it’s very easy to scan and go across. And so yes, I have a database of all my deal flow sourcing channels, but it’s also just nice to say, oh, I haven’t talked with Will, I haven’t talked with John or I haven’t talked with Maggie in a while. Yeah. It makes it easier that way.

Brian Bell (00:48:01): It’s something I struggle with. What CRM are you using?

Mike MacCombie (00:48:04): I have three answers to that. One is operational CRM. There’s a platform called Ace Workflow that actually is using me as a case study. They built out everything automated for me on Airtable. So every time I meet somebody new, there’s a form they fill out. I get everything automated and I can track them and get them into every loop that I run.

Brian Bell (00:48:20): that’s right I recall that from your deal flow thing it’s very organized on Airtable I use Airtable as well but I didn’t use it for my CRM it’s too much heavy lifting to use right now but maybe someday you know well that’s for the master one so I don’t lose track of things I think there’s the don’t lose track and there’s the how to action the don’t lose track is there

Mike MacCombie (00:48:37): I have all these WhatsApp groups where if I took a meeting with you, I’d write down the 15 communities that are a fit for you. That’s my actioning. I don’t want to have to reach out to 50 PropTech investors when I see a PropTech deal. I just say, hey, here’s my PropTech community of 150 investors. You all can take a look at the deal. If you miss it because you didn’t look, that’s on you. It’s so much faster to think of distribution on demand with one click rather than saying, how do I do 50 more emails because then I’m going to get replies via email it’s so much easier just to say what is the linear leverage so I have shotgun approach to visualization and I have a laser approach to what’s the single action I can take to get leverage for the thing I need to get done

I told him both, probably.

That same morning, OpenAI published Symphony, an open spec that turns Linear into a control plane for continuously running coding agents. The company said some internal teams had seen a 500 percent jump in landed pull requests in the first three weeks. By Wednesday, OpenAI and Microsoft had quietly rewritten their seven-year-old partnership so OpenAI can now ship products on any cloud. By Thursday, AWS announced OpenAI models were coming to Bedrock. By Friday, the Fed held rates with four dissenters voting for cuts, the most disagreement on the committee since October 1992, and the Q1 inflation print came in at 4.5 percent annualized.

If you take only one thing from the last seven days, take this. The frontier model layer, which a year ago looked like the moat, is becoming distribution. It is becoming infrastructure. It is becoming, in some cases, a commodity. The venture economy is reorganizing around that fact while we watch.

The capex print

Big technology companies reported earnings together this week, and the numbers told a single story.

Microsoft’s AI business now runs at a 37 billion dollar annual revenue rate, up 123 percent year over year. Azure grew 40 percent. Google’s cloud backlog reached 462 billion dollars, and the company raised its 2026 capital spending guide to 180 to 190 billion, with 2027 set to climb further. Meta lifted its capex range to 125 to 145 billion, citing component pricing and the cost of building future data centers. Amazon’s AWS hit a 15-quarter growth high at 28 percent. Free cash flow at Amazon dropped sharply because property and equipment purchases rose by 59.3 billion dollars, almost all of it AI-related.

Add it up across the four biggest spenders and you land somewhere between 650 and 700 billion dollars in 2026 spending on data centers, chips, power, and cooling. That is roughly the size of the entire United States defense budget.

The market reaction was telling. Microsoft and Meta both fell three to six percent after hours despite beating revenue. Public investors, for the first time in this cycle, are starting to ask whether the spending will pay back. They are not yet selling. They are starting to underwrite. That is a different posture, and it propagates downward into private markets.

The data center pipeline is also running into walls. On Wednesday, Compass Datacenters, owned by Brookfield, abandoned its 2,100-acre Prince William County project in Virginia after a March court ruling and a county vote against the rezoning. Sightline Climate estimates roughly half of the 16 gigawatts of US 2026 pipeline now faces delay or cancellation. Two-thirds of the new builds are heading to rural farm country, where transmission lines do not exist yet and local opposition often arrives before the steel does.

The bottleneck moved this week. It is no longer chips. It is power, land, and time.

What got funded, and why it matters

Six rounds last week tell a clearer story than any benchmark.

Hightouch raised 150 million dollars on April 29 to build an AI platform for marketers. Netomi raised 110 million dollars the next day for enterprise customer-service agents. JuliaHub raised 65 million dollars and shipped Dyad 3.0, an agentic platform for industrial digital twins. Scout AI raised 100 million dollars for a foundation model aimed at unmanned warfare. Aidoc raised 150 million dollars to expand its clinical AI deployment platform. And Ineffable Intelligence, the new lab from former DeepMind reinforcement learning lead David Silver, was reported to have closed a 1.1 billion dollar seed round at a 5.1 billion dollar valuation.

Five of those six are doing the same thing in different domains. They are putting AI inside a workflow that already had a budget owner, a system of record, and a cost of failure. Marketing operations. Customer service queues. Industrial design loops. Drone autonomy. Radiology departments.

The sixth, Ineffable, is a different category of bet. It is the third European AI lab in two months to raise a billion-dollar seed before shipping a product, alongside AMI Labs (Yann LeCun) and Recursive Superintelligence (which raised 500 million from GV and NVIDIA in mid-April). The thesis is that pure language model scaling has a ceiling and that whatever comes next is worth funding now, even at unicorn-plus prices, even without a roadmap. Sequoia, Lightspeed, NVIDIA, Google, DST, Index, and the UK Sovereign AI Fund all believe this enough to write checks.

The implication for early-stage investors is uncomfortable. The four or five labs taking pedigree-priced rounds are now sitting on cap tables that smaller firms cannot enter. The price of admission moved out of reach in a single quarter. Experienced early-stage investors are responding by funding the application and infrastructure layers that any of those bets will need to ride on. That market is wide open. The pedigree market is closed.

The True Anomaly round is in its own category. The space-defense company raised 650 million dollars on April 28 at a 2.2 billion valuation, four days after the Space Force named it among twelve primes for the Golden Dome interceptor procurement. The Golden Dome program is now estimated at 185 billion dollars, with 17.5 billion in the FY2027 request alone. True Anomaly’s previous round, less than twelve months earlier, valued the company at 260 million. Procurement visibility carried that eight-times step-up, with no help from the usual ARR-multiple math.

Defense procurement has become a real venture asset class this week, in the same way that healthcare AI did two years ago.

The partnership that broke

Three years ago, Microsoft’s deal with OpenAI looked permanent. Exclusive cloud rights. Exclusive IP license. A revenue share running both directions. An AGI trigger clause that gave Microsoft access to anything OpenAI built, until OpenAI declared itself something more than a normal company.

On Monday, all of that changed.

Under the new agreement, OpenAI can ship products on any cloud. Microsoft’s IP license remains in place through 2032 but is no longer exclusive. The Microsoft-to-OpenAI revenue share is gone. The OpenAI-to-Microsoft revenue share continues through 2030 but is capped. The AGI trigger clause is dead. By Tuesday, Andy Jassy at AWS had announced that OpenAI models were coming to Bedrock in the next few weeks. By the end of the week, AWS’s Matt Garman was publicly pitching that AWS could be a better partner to OpenAI than Microsoft had been.

The startup implication is the part that matters. For the last two years, a meaningful share of AI venture pitches arrived with some version of “we are built on OpenAI.” That sentence used to imply a cloud choice and a stack of integrations. Now it implies neither. Buyers will increasingly pick the cloud where their compliance, billing, and security already live, and pull the model from a menu. Selling access to a particular model has flattened as a wedge. What still works is multi-model routing, evaluation, observability, vertical data, and workflow ownership.

DeepSeek V4, released the week before, made the price-performance case more aggressive. The Pro version posts Codeforces and SWE-bench scores within shouting distance of frontier closed models, costs about a third as much per token, and ships under an MIT license. The US Commerce Department’s CAISI evaluation pegs the gap to closed frontier at roughly eight months. For any application where absolute frontier capability is not required, a 60 to 80 percent inference cost reduction is now available immediately.

The thin-wrapper thesis was already weakening. This week buried it.

The liquidity question

A different kind of news landed in private markets this week, quieter but worth tracking.

On Monday, Vinted closed an 880 million euro secondary at an 8 billion euro valuation, with EQT, Schroders Capital, and Teachers’ Venture Growth leading. Vinted is profitable, category-leading, and disciplined on unit economics. Secondary buyers love that profile. On Thursday, Lazard agreed to acquire Campbell Lutyens, a major secondary and GP advisory firm, for 575 million dollars plus up to 85 million in earn-outs. Avalyn Pharma priced an upsized IPO at the top of its range, raising 300 million dollars.

These three transactions together are a market-structure signal. The infrastructure of private liquidity, including secondaries, continuation funds, and tender offers, is being repriced upward. The SEC made this easier on April 16 by shortening the minimum issuer-led tender window from 20 business days to 10. This week, that change started showing up in client alerts and term sheets. Issuer-led tenders are about to become the dominant liquidity mechanism for top-tier private AI names, and the firms that intermediate them are positioning accordingly.

The takeaway for late-stage investors is that liquidity quality is opening, but not liquidity quantity. Vinted gets cleaner exits. Anthropic, reported by Bloomberg this week to be in talks with Google for as much as 40 billion dollars in additional cash and compute, gets cleaner pricing. The middle tier of “story stocks without exits” remains stranded.

Anthropic is worth a separate note. The company moved from 9 billion dollars in annualized revenue at the end of 2025 to roughly 30 billion by early April. The recent employee tender at a 350 billion dollar mark reportedly saw less selling than expected, while secondary buyers were quoted at 800 billion plus. That gap is unusual. Employees who sell early in a hot AI company are often the smartest indicator of where the price is going. When they refuse to sell at 350 billion, the market for the next tender will likely be higher.

What the macro is doing to the model

The Federal Reserve held rates at 3.50 to 3.75 percent on April 29, with eight votes for and four against. That is the largest dissent on the committee since October 1992. Two members wanted to cut. Two wanted to keep rates higher. Powell confirmed his last press conference as chair. He will remain on the Board.

The first-quarter GDP advance came in at 2.0 percent, up from 0.5 percent in the prior quarter. The PCE price index rose 4.5 percent annualized, with core PCE at 4.3 percent. Most of the inflation acceleration traces back to the Iran war fuel shock, which has held Brent crude near 105 dollars a barrel for two months and pushed US gasoline above four dollars a gallon. Spirit Airlines ceased operations on May 2, citing fuel costs. Forty countries are now reportedly considering nuclear power.

For venture, the macro implication is straightforward. The case for cheap capital re-emerging in 2026 is weaker than it looked in January. Private market valuation models that assumed multiple Fed cuts this year deserve a haircut. Founders pitching models that need rate relief to clear should expect skepticism. Companies whose customers are exposed to consumer discretionary spend should expect harder Q3 and Q4 pipelines.

What insulates a startup from this environment is pricing power, budget-owner buyers, and a workflow that costs more than the software does when it breaks. Marketing automation, customer service, clinical workflow, defense procurement, industrial simulation. The pattern in the funded rounds is not a coincidence.

What got weird

A few things this week worth treating as radar rather than thesis.

OpenAI’s Symphony released alongside reporting that its internal coding agents had cleared 5,000 production pull requests in three weeks of dogfooding. Symphony is open, which means by next quarter Linear, Jira, and GitHub-native versions will exist. The implication is that engineering organizations will spend 2026 redesigning themselves around stateful work pipelines that hand tasks to synthetic labor. The investable surface is in the seams between issue tracker and IDE, between agent and reviewer, between policy engine and compliance audit.

Anthropic’s Mythos Preview, gated under Project Glasswing to about 50 partners, continued generating cybersecurity ripples. The UK AI Safety Institute reported the model solving 73 percent of expert-level capture-the-flag challenges and chaining 32-step network attacks in three of ten attempts. Wordfence reported that AI-assisted vulnerability submissions grew from 16 percent to 66 percent of the total between November 2025 and April 2026. Berkshire Hathaway, Chubb, and Travelers all received approval this month to drop AI-related damages from corporate insurance policies. The next twelve to eighteen months are going to involve a sharp re-pricing of cybersecurity insurance, and a new procurement category for AI-vulnerability indemnity. SaaS contracts will start carrying language they have never carried before.

Google DeepMind launched an AI co-clinician research initiative on April 30. In blind evaluations on 98 realistic primary-care queries, the system recorded zero critical errors in 97 of them. In telemedical simulations across 140 assessment areas, it matched or exceeded primary-care physicians in 68. Expert physicians still performed better overall, especially on red flags and emergency triage. The framing is supervised teammate. That framing is what makes the procurement and reimbursement pathways begin to function. Healthcare AI is becoming a deployment category for growth-stage capital, with real workflow integration and real money behind it.

Figure announced its BotQ humanoid factory has hit 24-times its prior throughput, producing 55 humanoids per week, one per hour, with 80 percent first-pass yield. The company has shipped over 350 Figure 03 units. On the same day, China’s State Grid deployed 500 humanoid robots for high-voltage operations. Japan Airlines began piloting humanoid baggage handling at Haneda. The Beijing E-Town half marathon was won by a fully autonomous humanoid in 50 minutes 26 seconds. Production volume is no longer the constraint for humanoid robots. Demand validation is.

The cluster across these four signals is the same cluster. AI is moving from chat interface into workflow control. Whatever you build, build it with that fact in mind.

What experienced early-stage investors are doing

The pattern in this week’s rounds suggests a few questions worth asking founders right now.

• What workflow state do you own, and what is the cost when that state goes wrong?

• If AWS, Azure, or Google Cloud bundles a model, an orchestrator, and a compliance layer into the base offering, why does your buyer still need you?

• How does your roadmap survive Symphony or Anthropic Skills becoming the default orchestration layer?

• If DeepSeek dropped your inference cost by 60 percent next quarter, what part of your defensibility would still be intact?

• For agent companies, can you describe your verification and audit story with the same depth as your generation story?

• For security and trust startups, do you have a path into the consortium of companies with Mythos-class access, or is your TAM the long tail outside it?

The companies that have a clean answer to most of these are the ones being funded right now. The companies that do not are the ones whose seed rounds are taking three months instead of three weeks, and whose Series A pitches are bouncing.

For experienced late-stage investors, the watch list is concentrated. Anthropic and SpaceX are the two pre-IPO names where the price discovery is most active and the next ninety days will likely contain the cleanest entry windows. True Anomaly, Anduril, and Figure are the defense-and-physical-AI names where revenue visibility is starting to underwrite the valuation rather than narrative carrying it. Fervo Energy and X-energy are the first nuclear and geothermal names where public market issuance has cracked open. Vinted and similar profitable consumer marketplace assets are the secondaries to take seriously.

What to watch in May

A handful of dates over the next four weeks will tell us whether the patterns that landed this week harden into a regime, or get muddled by something new.

May 13. The EU Digital Omnibus trilogue meets again on the AI Act enforcement timeline. If August 2 holds as the binding date for high-risk and watermarking obligations, every European-exposed startup needs a compliance plan in hand by July.

The week of May 11. Senate confirmation vote for Kevin Warsh. The composition of the Federal Reserve through 2028 starts being decided here.

Mid-May. Recursive Superintelligence’s public launch. The first product release from the post-LLM lab cohort is the moment we find out whether the seed prices are real or theatrical.

Late May. SpaceX public S-1, with the roadshow targeted at the week of June 8. The cleanest test of whether the AI-frontier-tech IPO window opens or stays closed.

Through May. The first publicly accessible Mythos-class capability lands somewhere. Insurance coverage and SaaS contract language start changing the next week.

The infrastructure layer is more crowded and more expensive than ever. The model layer is becoming distribution. The application and workflow layers are wide open, but only for companies that own a state somewhere a budget owner cares about. Power, land, regulation, and procurement have moved from background conditions into active investment variables.

A reader who started Monday thinking AI was about chatbots and benchmarks should finish this week thinking it is about workflows, control planes, contracts, and kilowatts.

That is the regime. Build accordingly.

“The rich need to pay their fair share.”

It sounds clean. It sounds moral. It gives the crowd a villain and the politician a microphone.

But once you look at the federal income tax system, the line starts to wobble.

The top 10% of earners already pay about 71% of all federal individual income taxes. The top 1% alone pays roughly 38%. The bottom half of taxpayers pays about 3%.

That does not mean every wealthy person is noble. It does not mean the tax code is perfect. It does not mean loopholes, bailouts, crony subsidies, or special carveouts should survive.

It means the story being told at these rallies is too simple.

The May Day protests had a familiar script: workers against billionaires, ordinary people against the powerful, government as the only force big enough to make life fair again. The anger is easy to understand. Housing is crushing families. Groceries feel expensive in a way people can feel every week. Healthcare bills can still wreck a household. A lot of Americans are working hard and feeling like they are falling behind anyway.

That pain deserves to be taken seriously.

But blaming the rich is not a plan. It is a pressure valve.

America’s problem is not that successful people pay too little federal income tax. America’s problem is that government has become too large, too expensive, too indebted, and too comfortable promising things it cannot sustainably deliver.

A country cannot deficit-spend its way into dignity.

When government grows without discipline, the bill always arrives. Sometimes it arrives as taxes. Sometimes as inflation. Sometimes as higher interest rates, weaker growth, regulatory drag, or fewer businesses getting started in the first place. The cost rarely shows up with a neat label. It just makes life heavier.

And who feels that first?

Workers. Renters. Young families. Small business owners. People without assets. People trying to climb.

The same people politicians claim to be protecting.

The better answer is not a larger state with a bigger appetite. The better answer is a freer society with more paths upward.

That means making it easier to build housing so rent does not devour a paycheck.

It means cutting licensing rules that protect incumbents and keep people from earning a living.

It means a tax code ordinary people can understand without hiring a professional guide.

It means schools that give families real options, including vocational paths and apprenticeships, rather than trapping children in systems that fail them because of where they live.

It means sound money, because inflation is one of the cruelest taxes in America. It hits the grocery cart before it hits the yacht.

It means treating entrepreneurship as a public good, not a loophole for the ambitious.

The American Dream was never a promise that everyone would end up in the same place. It was the promise that your starting point did not have to become your ceiling.

That promise needs oxygen. It needs room to move. It needs a government strong enough to protect basic rights and restrained enough to leave people space to build their own lives.

The rallies are right about one thing: too many Americans feel stuck.

But they are aiming their anger at the wrong target.

The enemy is not the founder who built a company, the doctor who worked brutal hours, the investor who took risk, or the executive whose income already funds a large share of the federal income tax base.

The enemy is a system that makes housing scarce, schools uneven, healthcare opaque, businesses hard to start, money less stable, and government programs impossible to pay for without borrowing from the future.

If we want more people to rise, we should rebuild the ladder. We should stop sawing at the top rung.

A freer economy is not a gift to the rich. It is how more people become less dependent on politicians, bureaucracies, and slogans.

That is the part of the American Dream worth defending.

Not envy. Not dependency. Not permanent resentment dressed up as justice.

The real dream is still older, tougher, and better than that: work hard, take responsibility, build something, own something, improve your life, and give your children a better shot than you had.

That dream does not need another tax hike.

It needs a government that gets out of the way.

By the time vulnerabilities are found, the code is already live. By the time reports are written, attackers may have already moved. And by the time fixes are prioritized, the damage is often done.

That’s the gap Derek Foster is focused on closing.

Derek is the Co-Founder and CTO of Best Defense, an AI-driven cybersecurity platform built for a world where software ships fast and breaks faster. His background spans SRE, penetration testing, and large-scale infrastructure security across fintech and enterprise systems. Across those roles, he kept seeing the same pattern: security was always behind.

In this conversation, he breaks down why that keeps happening and what needs to change.

Subscribe now

The Core Problem: Security Is Still Too Slow

Modern development cycles have changed. Teams now deploy code daily, sometimes multiple times a day. AI tools are pushing that even further. Some teams report 2x to 10x increases in output.

Security hasn’t caught up.

Most companies still rely on periodic checks like annual or quarterly penetration tests. That model assumes systems stay stable between audits. They don’t.

Every new feature, dependency, or API creates a new attack surface. Attackers don’t wait for your next audit window. They move as soon as something is exposed.

Derek puts it simply: security can’t be a point-in-time activity in a system that changes every day.

Detection Isn’t the Bottleneck. Fixing Is.

Many companies already have tools that detect vulnerabilities. The problem is what happens next.

Security teams get flooded with alerts. Some are real. Some aren’t. Engineers have to triage, validate, reproduce, and then fix the issue. That process can take days or weeks.

In the meantime, the risk is still live.

Best Defense focuses on a different approach. Instead of stopping at detection, the system:

• Proves whether a vulnerability is actually exploitable

• Explains why it matters in context

• Generates a targeted fix

• Pushes that fix directly into the developer workflow

That last step matters most. If a fix shows up where developers already work, adoption goes up. If it requires another tool, another dashboard, or another process, it gets ignored.

Why Continuous Security Is Becoming the Default

The shift happening in cybersecurity mirrors what already happened in DevOps.

Years ago, testing happened at the end of the development cycle. Now it’s continuous. Every commit triggers tests automatically.

Security is moving the same way.

Derek calls this “shifting left,” meaning security happens earlier in the development process, closer to where code is written. The goal is simple: catch and fix issues before they ever reach production.

This becomes even more important as AI enters the stack.

AI is helping developers move faster. It’s also helping attackers move faster. The cost of launching attacks is dropping, while the speed of execution is increasing.

That compresses the window between exposure and exploitation.

If your defense doesn’t move at the same speed, you fall behind.

The Mistakes Founders Keep Making

Early-stage founders often treat security as something to handle later.

The logic is understandable. You need to ship fast, find product-market fit, and keep customers happy. Security feels like overhead.

But small gaps compound quickly.

Derek highlights a few common issues:

• Weak access control and unclear identity boundaries

• Unchecked third-party dependencies

• Sensitive data flowing through systems without visibility

• Lack of logging and monitoring

Attackers don’t need complex exploits. They look for what’s already exposed. Once inside, they move laterally and escalate access.

The advice is straightforward: build good habits early.

Know what data you have, where it lives, who can access it, and how it moves. Even basic hygiene reduces a large percentage of risk.

The Hard Tradeoff: Automation vs Control

One of the hardest problems in building AI security tools isn’t technical. It’s trust.

How much should the system do automatically? When should a human step in?

Full automation sounds appealing. But in security, mistakes can be costly. Teams need visibility, auditability, and control.

Best Defense approaches this as a spectrum:

• Some actions are fully automated

• Some are recommendations

• Some require human approval

Getting that balance right determines whether teams actually adopt the tool.

If it’s too manual, it slows them down. If it’s too autonomous, they don’t trust it.

The Bigger Shift: Security as a Built-In System

Looking ahead, Derek expects security to become almost invisible.

Not because it’s less important, but because it’s fully integrated.

Instead of separate tools and reports, security will:

• Run continuously in the background

• Validate changes as they’re made

• Fix issues before they surface

• Provide clear context when human decisions are needed

The end state is simple: developers don’t think about security as a separate task. It’s part of how software gets built.

Why This Matters Now

The stakes are rising.

Recent data shows the average cost of a breach is around $4.4 million. At the same time, supply chain attacks and third-party vulnerabilities are increasing.

AI is accelerating both sides. More code is being written. More vulnerabilities are being introduced. And attackers have better tools to exploit them.

This creates a new baseline.

Security can’t rely on slower processes anymore. It has to match the speed of development.

Final Takeaway

Derek’s view is clear: security isn’t a checklist. It’s a system that needs to operate in real time.

If your development cycle is continuous, your security needs to be continuous too.

Everything else is just catching up.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:

00:01 Introduction to Derek Foster & Best Defense
00:37 Derek’s Background and Early Curiosity in Systems
02:25 From Gaming to Cybersecurity Foundations
04:41 First Experiences in Security and Problem Solving
07:24 Origin Story of Best Defense
10:42 AI, Developer Velocity, and Rising Security Risks
13:15 Target Customers and User Focus
14:29 Why Traditional Security Models Are Broken
17:30 AI, Trust, and Security in Developer Workflows
21:19 Fixing Vulnerabilities vs Detecting Them
23:12 Building Automated Security Remediation
26:32 Market Trends and Investor Blind Spots
29:32 Common Security Mistakes Founders Make
31:52 Hardest Engineering Decisions at Best Defense
33:56 Open Source vs Closed Source Strategy
36:12 Long-Term Vision for Cybersecurity
39:30 Red Team vs Blue Team Explained
42:28 The Future of Cybersecurity and Automation
46:07 Seamless Security in Developer Workflows
48:30 Metrics That Signal Industry Change
49:37 Governance Debt and AI Risk
51:19 SOC 2 Compliance and Security Standards

Transcript

Brian Bell (00:01:01): Hey everyone Welcome back to the Ignite Podcast today we’re thrilled to have Derek Foster on the mic he’s the co-founder and CTO of best defense an AI driven cyber security platform building what they call continuous security and resilience testing kind of a mouthful Derek has spent more than a decade working in SRE pen testing large-scale infra security across fintech and enterprise platforms and today’s helping build technology that automatically detects vulnerabilities and even ships fixes directly into developer workflows we’re excited to talk to him about his journey the future of cybersecurity and what he’s building thanks for coming on Derek

Derek Foster (00:01:34): Hey, no problem. I’m happy to be here just to talk shop.

Brian Bell (00:01:38): Yeah. Fair disclosure, Best Defense is a portfolio company, so we’re going to be nice. I’d love to get your origin story. What’s your background?

Derek Foster (00:01:44): You know, it really just started off with a lot of curiosity in the beginning more than anything else early on. So when I was young, I was kind of always drawn to systems, like not just using them, but understanding why they behave the way that they do. and funny enough it actually originated from the gaming space you know I play a lot of games with friends and I was kind of the one that was more particularly interested in the how that they were building it than playing the game itself so I tinkered a lot with it one of my favorite things that I picked up and I guess that would be related to cyber security in a way was something called a game shark and I may be dating myself a little bit here but it was this nifty little in-between that you could put in on this device and It would basically overwrite things during execution time You can put a little code

Brian Bell (00:02:32): in there and it’d give you infinite life on a game or something Yeah, exactly that Regular Super Nintendo kind of attachment on the cartridge Yes,

Derek Foster (00:02:41): absolutely And that’s just kind of got me fascinated as to how is that actually working behind the scenes I just plug in this thing in between and it’s doing something magical And I kind of picked that part and learned a little bit more about the memory management and how things happen on the on the board and the transfer of data, right? So early on that garnered a pretty good interest to computers in general and then it became networks and then software and then eventually security and infrared scale. Thankfully I’ve been blessed to have some good experience like that. So over time I realized that I was really attracted to that intersection of complexity and consequence from the gaming days. So I just liked the situations where the tech details really mattered more than anything.

Brian Bell (00:03:27): Back when video games were actually hard too like you needed a game shark in some cases like I don’t know I remember like Ninja Gaiden like it was almost like impossible like to beat Ninja Gaiden which I did a Christmas or two ago I sat there on the couch with the Nintendo Switch and beat Ninja Gaiden but it had the reverse thing where you could like press the two button the rockers and go back you know if you messed up and it still took me like four hours to beat the game and that was like with like redos yeah it was kind of you didn’t have a redo like you just know you just had to start

Derek Foster (00:03:59): over yeah they had it really good these days with that we built a certain amount of grit and resiliency with the games that came out early on especially Super Mario and stuff like that I mean yeah go back to entirely they still keep some of those philosophies in the Dark Souls games these days right yeah where you have to start over quite considerably if you fail but nowhere near

Brian Bell (00:04:20): as rough as it was before my main game is Escape from Tarkov which I still play a few times a week now and that game is very unforgiving on dying like you lose all your gear and it like it takes a while to kind of earn it back it reminds me of the

Derek Foster (00:04:34): RuneScape days yeah playing that you know you go into the wild and if you get lured out by higher levels on the promise of good good gear and then they just jump in the woods and you lose all your stuff

Brian Bell (00:04:46): Do you recall your favorite GameShark hack? My favorite GameShark hack?

Derek Foster (00:04:50): This is going to sound silly, but I mean, I just loved having the infinite life’s period in anything that I was doing because it gave me the chance to really just try to break the games over and over again in many different ways without, you know, consequence. So I guess I would say unashamedly, I chose God mode most of the time.

Brian Bell (00:05:10): So gaming kind of one thing led to another and you ended up in cybersecurity. Do you remember what the first technical challenge was in cybersecurity that really grabbed you?

Derek Foster (00:05:19): You know, I don’t think that it was like any specific challenge. Like it was mostly puzzles. I just like to solve puzzles. So really like it wasn’t just solving a technical problem per se. It was just seeing how certain weaknesses could become an operational or business problem. weakness very quickly. And when I got out of college to start doing this type of work, that’s when it became much more compelling to me because of that reason. So it wasn’t just a specialty in IT, right? It’s a place where it was one of the clearest places where engineering kind of meets business reality, right? So you get to both sides of the coin, right? So a lot of people picture the cyber sex space as this kind of abstract cat and mouse type game but what hooked me was just how tangible it actually is so like a bad assumption about permissions some overlooked depth or some weak validation paths some blind spots in observability like those just don’t stay small for super long so those turn into bad stuff outages incidents and really expensive cleanups so that’s really just what I loved about the field in general like you can get you get to combine technical depth with high-stakes reasoning at the end of the day. So you’re not just saying, can I do this? You’re asking, what does it mean if this fails and how to design the systems that are resilient under this type of stress, right? So I think the first problem I ever solved in security was literally just as simple as secure API connections through leveraging tokens and stuff where the token itself was exposed through one of the repos that we had. And we just didn’t know the proper way to obfuscate and transfer this type of information around without causing any proper issues. And I mean, this was way, way early on. It was probably my first job when I was writing just straight up raw PHP codes and files with really no framework or anything like that behind it.

Brian Bell (00:07:21): So when did the idea for Best Defense first come together?

Derek Foster (00:07:24): You know, so I spent a lot of time in a multitude of different domains. So fintech, e-commerce, Multifamily housing, marketing, and such. Not only just doing general app dev, but network administration and security because of compliances and controls that were necessary, right? We worked a lot as developers and with developers to push out code at light speed. I mean, the business needs were real. They wanted new features all the time. There was no excuse. If you said it could be done in two months, they wanted it in three weeks. So there was a lot of push on developers velocity and then just to be able to get stuff out. But the problem was, was that when you’re met with that reality, you tend to circumvent a lot of the normal practices that you might put into place to have the consideration for security. You don’t really think about what type of data you’re piping through this endpoint. Where is it ending up visually at the different result pages that you may have? You’re not really sure. You’re not really sure of what compliance control you may be going against by choosing certain tech. Perhaps you didn’t check a vendor that you’re using, a new package that you brought in for any issues that they might have. Recently, all in the name of just getting things done faster, right? So we just kept getting caught up in this situation where annually, traditionally the pin tests were annually, and we would go through this bum rush process where the audits were coming up and we would buy these temporary tools to kind of throw a bunch of alerts at us, which always ended up being thousands of alerts at the end of the day. And you’d really had to sift through and figure out what was priority, what mattered, what was taken care of by compensating controls and things of that nature, right? So that you can have some kind of actionable list of things that are actual issues that could have an actual impact. And then you had to problem rush to fix those things before the others had a chance to say fail. So what we really wanted to do, and before we got to the idea of best fence there, we had worked with AI already, my co-founder, I, Dan, and we had built systems that would leverage high amounts and volumes of data and just that information, make decisions based off of a multitude, a deep amount of rule sets to make decisions on documents to create resolutions that the agents needed at the time based on that information. And this was before AI was sexy by any means necessary. So it was really hard to do at the time. But with everything that’s happened more recently, AI becoming a little bit more accessible, much better, higher context windows and things like that

Brian Bell (00:10:08): It kind of got- Well, more software than ever is being generated, right? Absolutely Things like Cloud Code and Codex and Cursor like you’re shipping more and more features than ever right business like has a never ending hedonic treadmill of feature requests and that you know given you can ship you know I’ve heard estimates of at least double if not 10x faster I think you’re going to have a lot more security vulnerabilities now right than you had you know five or ten years ago

Derek Foster (00:10:33): Yeah because you know not only did you know just the needs of the business have these developers you know inadvertently circumventing those types of checks and balances. But now you just have folks who just have no clue, period, on how to approach security or how to understand it or what’s really necessary for what it is that they’re doing conductually. I mean, it’s just generating at light speed. And traditionally, the security tools and processes have just lagged behind in that same velocity, right? It’s becoming table stakes now to bring that security up to the velocity of development. And that’s what we’re trying to do here, right? It kind of came together when Dan and I’s repeated experience turned into this kind of shared thesis. We spent enough time around those situations, those systems, those real operations and real security issues to kind of see that gap from a couple of different angles. And since they were being asked to move faster, automate more, adopt AI, but also reduce exposure and maintain trust, the tooling around just still felt too, it was just too periodic and and just really noisy. So the more we looked at it, it wasn’t like a missing feature issue per se, it was just like a categorical issue. Like security testing, exploit validation,

Brian Bell (00:11:50): resilience validation and remediation workflows and all the stuff that we focus on at Best Defense,

Derek Foster (00:11:55): they were just too disconnected from each other and tools sprawl, so many tools involved compared to how engineering teams actually operate. So it started to feel a little bit of inevitable that we should create a platform that kind of married all those ideas together made it easy to use not only for people who just didn’t have huge tech backgrounds but also configurable and targetable enough for real red team operators to use like a sniper

Brian Bell (00:12:21): instead of a shotgun per se yeah so who’s typically your customer so we find that a lot of reasons that people come to us is really just for high compliance reasons so we end up speaking with

Derek Foster (00:12:38): development teams in general red team operators who are looking to kind of give themselves superpowers and amplify their capabilities but really it’s it’s really just about the people who are using the product at the end of the day not necessarily c-suite of course we we cater to them in a way where we create nice dashboards that give them the metrics and the trend analysis of what we find and what was fixed over time how it maps to their compliances and all that but being engineers ourselves we care mostly about making it very seamless and easy to use for for the operators at the end of the day so that’s who we tend to target the most

Brian Bell (00:13:16): yeah so what so I get the problem like what was your unique take when you looked at the market and you’re like okay this should exist and it doesn’t like what was it about what you were building where you just had to like see it through and build it well there was a time when companies would think about security and intervals so you’d run

Derek Foster (00:13:28): run a pen test produce a report fix a penetration test for anybody listening doesn’t know what security is so you’re basically trying to you’re hitting endpoints and you’re trying to like get into the system who do certain like API injections and just trying to like try to fool the system into giving you access well said Brian barely barely well said no it was good so you know you’re fixing these issues and trying to satisfy some compliance requirements and then move on but that model largely assumes that your environment stays sufficiently stable between those annual quarterly checkpoints and as we know now a lot of modern environments just absolutely do not

Brian Bell (00:14:17): Yeah, you’re doing continuous integration, CICD kind of pipelines where you’re deploying every day or two or a week. So technically you should be doing pen testing every time you have a pull request, more or less, which is a developer has a piece of code and I want to integrate it into the code base and ship it. You should be able to run it both pre-integration and post.

Derek Foster (00:14:38): Yeah, yeah, absolutely. I mean, code’s just shipping so fast. Throw a change is even faster. Tendencies on top of that. but you know in AI just around all of it accelerating it so it’s not the question just really becomes like how much confidence should you really have in a point in time assessment versus like a continuously changing system right so you hit the nail on the head that’s why continuous monitoring and validation matters so much now they had a report that came out that explicitly framed that monitoring and testing around continuous monitoring and periodic penetration testing and bone assessments depending on the monitoring maturity in place so this is the New York Department of Financial Services that’s the one okay yeah I can I always have to look this stuff up it drives me nuts yeah so that is a uh that’s a pretty good reflection of where like really things have gone when when we say the the model currently is broken really just mean I guess I would say I really just mean it’s too static for the dynamic environment

Brian Bell (00:15:44): right because it’s not just about detecting vulnerabilities what your platform does is it delivers the fixes directly into developer workflows maybe you could talk a little bit about what’s the technical breakthrough that made that possible yeah so

Derek Foster (00:15:55): I mean it was just a lot of different things that kind of coalesced right there was I think the biggest one being the trust factor when it comes to you know AI and how it how it’s trusted in your workflows and what you allow it to do We’re seeing a lot of that wall break down a bit more. It’s still a bit of a divide out there, but when we’re talking about adoption of AI into the development workflow, that seems to be where it has taken storm more than any place else as it is, right? So it’s kind of changing the narrative there that AI and programming can be a viable option. I think now with still more human in the loop involved, but it’s really helping amplify the skills of people who are out there, right? People are becoming more natural with that in their flow, you know, leveraging it for planning, checking against their work, looking for issues and refactors and stuff like that. So adding in something like that in the security space just became much more viable because people are now just understand how it can fit into their workflow on the day to day basis. And with them becoming much more accurate these days than before with the benchmarks that you’ve seen from Anthropic and stuff, it’s becoming a little bit easier of a bite to chew. I don’t think they’re fully ready for it, right? We’re definitely on our way and we’re seeing that with some of our customers. So I would say secondarily to that, it’s not just the AI tools in place that are causing this kind of shift in the market. But we’re seeing the narrative change from big names in the field like Chase and JP Morgan and even Google and stuff have been coming out with some keynotes and memos that they send out about how important it is for security to be more of a continuous function in your development environments than it is periodic because of the rates of adoption from not only developers who are building stuff, but also from the attackers who are trying to get into your systems. They’re not waiting for anybody. They’re already bought on. They’re on board. There’s been a few breaches even recently stemming from, you know, other countries that have leveraged AI in a way where it was more or less unmanaged. Just a few clicks for them to orchestrate attacks across dozens of different companies to exfil data at light speed. And, you know, unless you have the same kind of momentum and velocity in your defense work as the attackers do, you’re always going to be behind. So it just makes sense now to have this AI defense versus AI offense, right? But I think there’s still something about the operators that need to be in place so that we can leverage their creativity and their skills and their experience to leverage it more as a directed system than anything else. And so bringing this all, that’s the one side of the coin, right? The secondary side of the coin is the fixing it. so the fixing it’s always been an issue as I mentioned before you get inundated with all these alerts from these systems that find it but then you have to go through the triage of verification and then fixing the issue and doing making sure that it actually works so through regression testing and things like that right so leveraging on that side of the coin to be able to create a system that not only is is accurate enough to make a thin slice change to solve the problem but also gives a human in the loop for verification for review It makes it an easy addition to the team because we’ve seen with our benchmarks about 85% reduction and mean time to resolution just because after our system takes care of all the finding it part, it has all the context and information and awareness to tell you why it’s exploitable and the way that it was exploited and the reason why you should fix it in this way. And that makes it easy for the dev to sign off and say, okay, that looks like a great a great fix. Let’s click it through. And the red team operators, I’ll say from my recent ICP interviews, have been primarily excited about the fixing it part because there’s there’s just so much.

Brian Bell (00:19:48): they have to do they can detect a lot of issues right and like a bunch of exploitation surface area but now they got to go fix it right and I think that’s an important distinction is it’s not just detecting the exploit like points but also generating the fixes maybe maybe you talk about like how hard that was technically to build that kind of system

Derek Foster (00:20:05): so when you think about code fixing you know a lot of people will you know consider something like Anthropic and it it will fix your code and I mean it’s a powerful powerful AI and it it can fix issues when it finds them but there’s a few things around that that needed were necessary for this to be adoptable at a in a scalable sense and and for organizations that are high compliance rights it’s it’s really not about can the AI generate text for the security fix the unlock was just the bridge between that gap and between the detection and action right so this part of

Brian Bell (00:20:43): the system that was super difficult was and then verification so it’s not just the detection and the patching but then it’s like designing the regression tests to now okay now that we’ve merged this patch does it solve the problem that we detected and going forward every time we push more code into production, we can now run those same tests again.

Derek Foster (00:21:06): Yeah, you touched on a very key point, and that’s the exploit validation, right? A lot of the existing tools, they just tell you that something might be a problem, but to have it provably exploited is the key, you know, from that. And in that provable exploitation gives you the information and the context to create a highly targeted and effective fix for the issue.

Brian Bell (00:21:27): Yeah, it’s almost like a red team co-pilot tool. in a way. I mean, I know CodeBuy is kind of a dirty word now in software, but it is sort of like that. It’s a cloud code for red teaming.

Derek Foster (00:21:38): Yeah, I think that’s a pretty good way to put it. I’m going to steal that one. I’m going to write that down somewhere. Cloud code for red teaming. Write that down somewhere because it is, and it needs to be focused on those specific issues and to leverage the same tools that they use on a day-to-day basis and the way that they would use them because they just spend an enormous amount of time doing the reconnaissance on the targets that they’re given to do the enumeration to find out where those holes are and the scanning of the vulnerabilities to build that attack surface. There’s so much to that part of the job that takes so much time that by the time they get through that and run through all the tests and find things, the devs have already pushed out a ton of stuff. The attackers may have already found those things that needed to be patched before then. So that needed to be sped up. to meet today’s demand and velocity. And it was especially timely now because AI related risk is just a huge governance problem, not just a product problem per se. If you read up on IBM’s 2025 cost of data breach research, ungoverned AI systems are just way more likely to be breached and costly when they are. And that average right now is sitting at like 4.4 million per breach. So it’s not just creating opportunities for defenders, It’s also raising the cost of weak oversight, the late periodic checks that are there, right? So at the end of the day, it’s powerful to consider having this in your workflow, especially for the remediation. But yeah, that’s what I would say at the end of the day.

Brian Bell (00:23:12): Yeah, I mean, so this is a huge market. Obviously, I think the timing is excellent. AI is creating new gaps that can be exploited. The spending is huge. We’re seeing huge exits in this area with like Google acquiring Wiz for like $20 billion. You know, good for them. What shifts are you seeing on the ground that most investors and founders are underestimating? Just thinking.

Derek Foster (00:23:35): So dramatic pause, right? I think just the first underestimation here is just really it’s the shift of speed the attackers are just getting insanely fast like not just more sophisticated just super fast I mean AI itself is just lowering their costs fundamentally on research and contextual generation and like social engineering support that was a that was what we saw most most of what AI was being used by the threat actors first when that stuff came out was just making their Their emails sound, you know, better when they came through. So you might actually believe that it’s a prince that needs money. Not only that, but also the coding assistance lowers the barrier of entry for people who are interested in becoming hackers and exploiting businesses like that, right? So the time window between exposure and exploitation matters just way more than ever. And I would say secondarily, the supply chain. and third party risks that are involved, right? So everybody says they care about it, but a lot of organizations just operationalize it weekly. In Verizon’s report last year mentioned breaches through third parties doubled like 30%. And even OWASP in their top 10, they elevated the software supply chain failures to a top tier risk category.

Brian Bell (00:25:02): So maybe you can unpack what you just said. What is that acronym for people listening? OWASP?

Derek Foster (00:25:07): OWASP? yeah oh yeah yeah of course I apologize for that I know these acronyms can just get a little well especially especially in cybersecurity there’s lots of acronym shorthand yeah yeah I know and you know that that makes a lot of sense so that’s that’s basically the if I if I don’t mess this up open worldwide application security project is what that is so They’re kind of a non-profit. They dedicate mainly to just improving security through free and open source tools at this point.

Brian Bell (00:25:38): So we back a lot of founders that are building startups. What security mistakes do you see over and over again?

Derek Foster (00:25:46): If I were to boil it down to similar patterns that show up. So when it comes to founders, they have a lot of good intentions, right? They want to ship fast and get cool features out to their customers as they’re trying to find their product market fits and just cater to those who are on there currently so that they don’t experience crazy churn to start out right so but the problem with that is that the discipline of security becomes a bit more postponed so like a lot of founders will often assume that they can just clean up identity boundaries or access control stuff or clean up their dependencies, add logging in the right places, right? And do all their exposure management stuff later. And sometimes they can, but until they can’t, then it’s super dangerous because the thing people don’t understand is that attackers don’t really need your most exotic weakness to exploit per se. They just need whatever’s available to build the map. and to be able to move laterally within systems to do something with it, right? So the pattern data that we keep seeing reinforces that it’s mainly credential abuse, vulnerability exploitation, and operational weaknesses at that, that remain some major breach paths, right? So my advice is really just don’t act like, try not to act like a giant enterprise on day one per se, but just build your good habits early. Know what you have, know what data you’re managing, know what’s exposed, know who has the access to that, and make sure you take the time to validate it. Because if you just take those steps, just keep that proper hygiene, then you can prepare yourself for most things and be able to see the writing on the wall for where your weak points really are. And that’s where you can focus your limited bandwidth of security testing on or on those layers, those holes that you have.

Brian Bell (00:27:39): Yeah. So you’re CTO, right, for Best Defense?

Derek Foster (00:27:44): Yeah.

Brian Bell (00:27:44): What has been the hardest engineering decision you’ve had to make since founding the company?

Derek Foster (00:27:49): That’s a good question. The hardest engineering decision. This might be the one. So it’s not even specifically to the engineering. It’s really just the hardest recurring decisions is where to draw the line between the autonomy and the control, okay? Because it’s really easy to think Autonomy is great I mean because autonomy is exciting right people love the idea of AI just taking over the hard parts of everything but what we found is that really our customers really need scope control and and governance and evidence and auditability and the ability to just understand what the system did and why it did it so the hard the harder engineering decisions often come down to really like where should the system act automatically where should it recommend where should it automatically fix something where should it wait for some kind of a human and rights right so that’s not just a product design question per se that’s that’s like a trust design question is kind of a product philosophy question right Yeah, because in security, it just matters a lot to trust. Because they’re not just buying capability at the end of the day. They’re buying confidence in how capability is constrained, I guess I would say. Because with security, just so much can go wrong.

Brian Bell (00:29:18): And you guys are closed source, right?

Derek Foster (00:29:20): At this moment, yeah.

Brian Bell (00:29:21): What was that decision like? Because you probably looked at it and you probably considered open source. So there’s some security companies that kind of start with that kind of open source framework and then build a SaaS offering on top of an open source. Did you guys consider open source when you guys launched?

Derek Foster (00:29:37): Yeah, we certainly did. We certainly did. And I think that that’s not even out of the cards right now. We’re actually putting together one of our first open source contributions, which is going to kind of be a light version of the red teaming side of things and perhaps even with remediation coupled with it, right? A light version of that to help out the community. But I would say the core reason why we didn’t start off with that was simply because we were building these tools for our own internal use at the time for our own needs and really didn’t expand out into the space until we realized that other people might actually need it, right? And by that point, We’d already been working with other red teamers and stuff on design and figuring out how it can work at scale operationally. There’s pieces of this that will become open source and I probably already have, even from other people. But there’s a real need for the system that lives around that. And that’s when you start getting into the enterprise level, the public sector level with government and stuff. They can’t particularly just lean on on the open source side of things because they need a higher level of operationalization of governance and auditability that needs to surround these tools in order for them to be safe to use. So that’s one of the reasons why we chose to do that initially, because we want to figure it out as we’re going along here with our customers and as we grow. And then once we understand what would be the safest output for the general public so that they can do the same kind of things that we do to some degree, then that’s going to be when we do that.

That’s not how Arun Penmetsa sees it.

After more than a decade investing at Storm Ventures—and years building inside Oracle and Google—his view is simpler and stricter: if your product isn’t solving an urgent problem, nothing else matters.

This conversation breaks down how strong enterprise companies actually get built, why most teams stall after early traction, and what investors really pay attention to when deciding who to back.

Subscribe now

Product-Market Fit Starts With Urgency

Founders often point to metrics to prove they’ve hit product-market fit. Arun looks at something else first.

Do customers need this now?

If the answer is no, you’ll see the symptoms:

• Long sales cycles

• Endless pilots

• Budget delays

• Low conversion

You might still close deals, but growth won’t compound.

The goal is to find problems that sit at the top of a buyer’s priority list. The kind that forces action, not consideration.

A useful test: if your product disappeared tomorrow, would your customer scramble to replace it?

If not, you’re early.

The Hidden Gap: From Founder-Led Sales to Scalable Growth

Many early-stage startups get their first customers through the founder.

That works—until it doesn’t.

The real bottleneck shows up when you hire your first sales team. Suddenly:

• Deals slow down

• Messaging gets inconsistent

• Conversion drops

The issue isn’t talent. It’s missing structure.

What worked in the founder’s head never got translated into a repeatable system.

Storm Ventures focuses heavily on this transition. One key idea: build a customer journey alongside your sales pipeline.

Most teams track internal stages like:

• MQL → SQL → Close

But they ignore what the buyer is going through:

• What pain are they solving?

• Why should they trust you?

• What convinces finance and procurement?

If you can’t answer those clearly, scaling sales will break.

Enterprise Buying Hasn’t Changed as Much as You Think

Tools have changed. Buyers haven’t.

You can automate outreach, use AI for targeting, and close deals faster. But the core questions inside every enterprise are still the same:

• Does this save money or time?

• Does it reduce risk?

• Will this make the buyer look good internally?

Winning in enterprise means aligning with those incentives.

The best founders understand how decisions get made across teams—not just by a single champion.

Why Go-To-Market Mismatch Kills Startups

One of the most common mistakes Arun sees is misalignment between product and go-to-market.

Example:

• A company sells a $15K product using a heavy sales team

• There’s no clear path to larger contracts or expansion

The math doesn’t work.

You either need:

• A low-cost, high-volume motion (product-led)

• Or a path to higher contract value over time

Without one of those, growth stalls.

This is where many startups fail—not because the product is bad, but because the business model can’t scale.

AI Changes Speed, Not Fundamentals

There’s a lot of noise around AI replacing SaaS.

Arun’s view is more grounded.

AI will reshape how software is built and delivered. It will:

• Speed up product development

• Improve automation

• Increase efficiency

But it won’t eliminate the need for software companies.

Instead, it will change what good looks like.

Winners will:

• Own critical workflows

• Control valuable data

• Deliver clear outcomes

The deeper you sit in a customer’s operations, the harder you are to replace.

The Most Overlooked Risk: Deployment

Founders celebrate closing deals.

Customers feel stress.

Signing a contract is the moment of highest risk for the buyer. They’ve committed to a new system that might fail.

If deployment drags, risk increases.

That’s why speed to value matters more than closing speed.

The faster you:

• Integrate

• Show results

• Prove ROI

The more trust you build—and the more likely that customer expands.

Early Metrics Can Mislead You

ARR shows up in every pitch deck.

But the definition has gotten loose.

It might include:

• Usage-based projections

• Unactivated contracts

• Annualized short-term spikes

The number alone doesn’t tell you much.

What matters:

• How predictable the revenue is

• How quickly customers expand

• Whether usage translates into long-term retention

Strong companies don’t just grow revenue. They build reliable revenue.

What Great Founders Do Differently

As companies scale, founders naturally move away from day-to-day details.

That’s where many lose their edge.

The best ones stay close to customers:

• They join calls

• They hear objections firsthand

• They see where the product breaks

That direct exposure shapes better decisions—especially around product and strategy.

Distance creates blind spots.

Where the Next Opportunities Are

Arun sees several areas where startups can still win big:

• AI-driven enterprise applications
  Still early. Most workflows haven’t been rebuilt yet.

• Cybersecurity (especially AI-driven threats)
  Attack surfaces are expanding fast.

• Manual, overlooked workflows
  Many industries still rely on spreadsheets and fragmented tools.

• Retail and physical operations
  Large markets with slow tech adoption.

The pattern is consistent: big opportunities exist where complexity slows incumbents down.

Final Takeaway

Enterprise startups don’t fail because of lack of demand.

They fail because:

• The problem isn’t urgent

• The go-to-market doesn’t scale

• The product doesn’t embed deeply enough

The companies that win do three things well:

1. Solve a real, immediate problem

2. Turn founder knowledge into a repeatable system

3. Become critical to how customers operate

Everything else—funding, timing, even technology—follows from that.

If you get those right, growth stops being unpredictable. It becomes something you can actually build.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:

00:01 Intro & Arun Penmetsa Background
01:06 Early Career at Oracle & Google
03:00 Transition from Operator to VC
05:03 Is an MBA Necessary for Venture?
07:00 Joining Storm Ventures
08:26 Shifting to Investor Mindset
11:04 Storm’s Enterprise Focus
14:25 Storm’s Go-To-Market Playbook
18:55 Defining Product-Market Fit
20:52 Evolution of Go-To-Market
24:05 Investment Stage & Check Size
25:04 How Storm Evaluates Startups
27:58 Platform vs Feature Risk
30:07 Common Investment Mistakes
31:55 Cybersecurity Market Insights
33:29 Capital Efficiency in Startups
35:28 Technical Differentiation Debate
38:08 Where to Build in the Stack
42:20 Future of Work & AI Impact
43:48 SaaS vs AI Debate
46:00 Rapid Fire: Enterprise Insights
48:41 Future of Cybersecurity
50:05 Manual Workflows & Opportunities
53:21 Metrics VCs Don’t Trust

Transcript

Brian Bell (00:00:42): Hey everyone, welcome back to the Ignite podcast. Today we’re thrilled to have Arun Panmetza on the mic. He is a partner at Storm Ventures, an early stage VC firm focused on building enterprise leaders. Arun leads investments across SaaS, cybersecurity, and digital health. And before stepping into venture, he built enterprise software products at Oracle and was part of the early Google apps for enterprise team. Very interesting. He brings a rare mix of deep technical grounding and over a decade of early stage investing experience. Thanks for coming on, Arun. Yeah, great to be here. Thanks for having me. Yeah, so I’d love to start with your origin story. What is your background?

Arun Penmetsa (00:01:14): Yeah, so I was born and raised in India. I came to the US for undergrad many years ago now. I would say I have a somewhat traditional engineering background. undergrad and masters in computer engineering and then I spent as you briefly mentioned about five years briefly at Google but mostly at Oracle working on the in with enterprise teams there and then business school and then I joined storm about more than 11 years ago now so time has really flown by

Brian Bell (00:01:40): Yeah, I mean, it’s very, very standard background for Silicon Valley. You started your career building enterprise infrastructure. What did that teach you about, you know, deploying capital today? Like, what do you kind of still draw on from your Oracle and Google days?

Arun Penmetsa (00:01:54): Yeah, so it’s interesting if you think about just how technology trends evolve. Every time there’s a new wave, the infrastructure tends to get built first, you know, the picks and shovels, as people say. And then eventually you get to the application layer because a lot of the applications require that infrastructure to truly scale. Obviously you know being at Google for a little bit and then Oracle it was a completely different time and this was like 15 plus years ago. The cloud was just starting out. Companies were moving their products to the cloud and it was still very, very early days. But it was interesting to see that trend from the perspective of, let’s say, someone like Oracle, who even today and obviously at that time too, had a massive customer base. And at least someone like me who was just out of college, it was great to get the perspective of being able to watch these large companies support their customers through a technology transition one of the benefits of being at these companies was that you didn’t have to go find customers right like they would sort of they already had these customers they would come to your HQ for briefings and it was wonderful experience to interact with them understand how they thought about buying software what their priorities were how would sales marketing customer success engineering everybody work together So I think that was helpful to as I think today and obviously it’s a different world right now given everything with of how these big businesses operate and how they think about their incentives when they’re adopting new technology.

Brian Bell (00:03:17): So at what point did you realize that you wanted to move from building products to backing the people building the products?

Arun Penmetsa (00:03:23): Yeah, I mean, there wasn’t any master plan or anything like that. So after Oracle, you know, I had applied to business school and I was fortunate to go here at Stanford. And so I happened to meet a lot of investors on campus, either teaching classes or there were guests in class. and I just got more curious about venture so it’s an interesting story I literally cold emailed a number of GSB alumni who were in venture at that time and I was just curious like you know what their journey was what it takes to become an investor and I think maybe this speaks to one of the strengths of going to business school there but I would say like 80-90% of them responded and they were open to having a short conversation and I yeah eventually i ended up at storm over the summer and that you’re almost

Brian Bell (00:04:07): obligated right to take those calls yeah i mean those alumni reach outs i take i take them i you know you said 80 90 i probably take most of them as well most of

Arun Penmetsa (00:04:16): them yeah i think that’s one of the benefits especially because being in the valley there’s such a good concentration of alumni here as well so it kind of works out

Brian Bell (00:04:24): well i have a question for you actually and i want to get back to your story of gsb because i have an intern i won’t say his name that’s been working on team ignite for a while I have actually two one that got into Berkeley and deferred I have another guy who’s applying now to get his MBA do you think it’s like necessary or helpful I guess that’s a two-sided question we’ll take both parts of that to get an MBA to work in venture is it helpful and necessary.

Arun Penmetsa (00:04:48): So I don’t think it’s necessary at all. There are many paths into venture, as you know, you know, I think having an operating background, especially in today’s era where you’re building a lot yourself, I think is probably a really good way. I think what’s helpful is if you don’t have an established network to source from or to help you with diligence or to kind of break in. It can be helpful to have a network from a top tier MBA school to kind of get started. And let’s say in my case, I was coming from a pure tech background. I didn’t know really anything about investing or finance or all those areas. So having that network definitely helped a little bit. I was living here, so I did have an established network through my colleagues, you know, the school I went to and so forth. But I think it can be helpful if you don’t have that. So yeah, that would be my view.

Brian Bell (00:05:38): Yeah, I think especially if you come from an engineering background and you’ve never operated, you’ve never shipped a product other than you’ve coded products, but you’ve never like ran an org that shipped a product or you never went zero to one as a founder or anything like that. You don’t have the business side of things. You’re very strong technically.

Arun Penmetsa (00:05:54): I think it could be I don’t think it’s necessary, as you said, at all, right? I mean, I think the best learning is just by doing, right? So if you go work at a startup, you’ll get that experience. Or start a startup. Yeah.

Brian Bell (00:06:05): Or start a startup company. That’s probably the best, yeah. That’s what I tell young people. Don’t try to work in VC. I mean, yeah, get an intern at a VC, whatever, get some experience or be an associate, but go start a startup as soon as you can. Exactly. especially today there’s just so much greenfield opportunity out there and it’s

Arun Penmetsa (00:06:23): never been easier to start a company given all the infrastructure and resources and

Brian Bell (00:06:27): tooling you have today so you’re reaching out to your network you’re getting some informational interviews as they say and then at some point you landed as an

Arun Penmetsa (00:06:34): associate at Storm tell us about that yeah so I spent the summer here at Storm it was a great experience I mean Storm and I can talk more about the firm but It’s a relatively small team It’s been around now for a little over 25 years You know, a great track record And I have to really credit the culture at Storm Because I got great mentorship over the summer Everybody engaged with me really well And again, because it’s a small firm It’s just partners Everybody kind of does their own work

Brian Bell (00:07:02): Kind of like a benchmark model Everybody’s a partner Everybody does their work Yeah, that’s right

Arun Penmetsa (00:07:06): so everybody did their own work and I was involved we didn’t actually end up doing

Brian Bell (00:07:10): a deal when I was there over the summer so it’s a very concentrated firm so you guys are sitting on boards writing bigger checks and

Arun Penmetsa (00:07:18): yeah so i would say we’re relatively concentrated and the number of deals per year fluctuates it depends some years we may do like six to eight some years we may do three to four but we obviously spent time looking at a lot when i was there and i really enjoyed it and i really enjoyed working with the team so when i graduated you know i was fortunate to have you know i met the team after i graduated and i was fortunate to get an offer and uh yeah now it’s 11 years later that’s awesome

Brian Bell (00:07:40): what was the biggest adjustment for you you know going through GSB and shifting into investor mode.

Arun Penmetsa (00:07:47): Yeah, I think between operating and investing, I guess there’s a couple of things you really need to think about. One is you are looking at many different opportunities whether it’s new investments or companies you’ve already invested in so I think you need to be able to context switch quite often right but still be present and very thoughtful about each engagement so a simple example would be you know as you go through the day you may meet four five six companies and each of them may be doing slightly different things in different sectors at different stages I mean generally because we are early stage the stage doesn’t That’s the hardest thing actually about being a VC is the context switching, I think.

Brian Bell (00:08:29): Because every half hour is something different. Exactly.

Arun Penmetsa (00:08:32): But you still have to be present, you have to think through, you know, because this could be the next big deal, right? And I think that is something, that is a muscle I think you need to build. And obviously this happens with portfolio too, right? You can have a board meeting, then you can go to a different company and my founders call me all the time or text me. So I think that’s one aspect. I think as investors we have to be comfortable sort of realizing that we don’t have as much control over the business even though you obviously hear stories about investors being very involved and I think there’s a lot of value we can add but ultimately it’s the founders that are running the show. Right. And that’s the way you want it. I mean, ultimately, you back this team to go execute and you want to support them in whatever you can. And obviously not every, as you know, most investments probably don’t work out as well. But but that the idea that, you you have limited information ultimately because you’re not in the company day to day and being able to still be helpful and thoughtful in the kind of advice you give and the questions you ask is a delicate balance because you know when things are not going well if you come from an operating background or I suppose any background the instinct may be to like oh I want to roll up my sleeves and help but you have very little information in most cases and you have to be careful how you how you show up in that moment as an investor because it has a huge impact on the team

Brian Bell (00:09:51): Yeah, it’s a delicate dance. So you guys have a long-standing focus on enterprise. How do you kind of keep that focus and stay disciplined over all the different hype cycles that have occurred over the last 11 years that you’ve been there? Right.

Arun Penmetsa (00:10:07): So I think, first of all, I think enterprise, and it’s B2B more broadly, so it’s not like we’re only looking for companies that sell into large customers.

Brian Bell (00:10:15): Could be like SMB, B2B, things like that. Yeah, exactly.

Arun Penmetsa (00:10:20): I think one, obviously it’s a very big market. Huge amount of software spend goes into these companies as they procure products. I think two, the team realized over the years and the team has been investing together in some form since 99, that One of the key aspects of selling into these businesses is the way you go to market has a lot of similarities across sectors. So whether you’re selling into cybersecurity or sales tech or healthcare, obviously every sector has some nuance and the way they buy there’s potentially regulation, the sales cycles are different and so forth. But the way businesses make decisions The way incentives are set up, the way you go to your champion, then you have to convince the rest of the org, including finance and procurement. There’s similarities across sectors and businesses that you can learn and work on over time. One of the things that the team has really developed over the years is this really deep understanding of go-to-market, whether it’s sales-led, whether it’s product-led, whether it’s using channel or marketing-led. And we realized that using that context and playbooks that we’ve built can be beneficial no matter, like I said, whether you’re selling to hospitality or you’re selling to healthcare or some other sector. So the combination of one, this is a very big market and two, the expertise we’ve built on the go-to-market side is what keeps us really excited about this because we think we can find good opportunities but also help these companies, you know, as we say, get to unlock, as we say, they’re able to unlock growth, which is essentially go from that early product market fit stage to a stage where they can predictably scale growth.

Brian Bell (00:11:54): So I think that’s interesting. And the last thing I’ll say is, you know,

Arun Penmetsa (00:11:57): I think B2B also in some in some ways tends to come a little bit after the hype because these are larger companies making decisions. So it doesn’t it’s not like a trend hits it and then the next year everybody’s buying it. Right.

Brian Bell (00:12:10): It’s more like Web3 will do that. Yeah, that’s great. They’re going to evaluate it and say, how does this drive? You know, how does this save us money or save us time?

Arun Penmetsa (00:12:19): Yeah, yeah. Primarily, right? If you think about AI, right, as a trend, and we can definitely talk about that. I mean, now we’re seeing a lot of adoption, but you know, 18 months ago, a lot of people in CIOs, CTOs, CPOs, they were trialing products. They weren’t ready to like jump in and So there’s time and there’s a playbook you can run to get these companies successful.

Brian Bell (00:12:41): Love that. So what is that playbook for founders listening? Why would they want to take money from Storm? What have you guys developed and how do you guys help founders repeatedly get to that product market fit? Let’s say go zero to one, one to ten, things like that.

Arun Penmetsa (00:12:56): Yeah, yeah. So I think a lot of the learnings, one of my partners, Tehi, has actually written books on this. And there’s a website called Survival to Thrival where he’s published a lot of this. But the brief summary of it is we think a lot of go-to-market playbooks tend to focus, are more internal facing to the company selling. So if you think about the traditional pipeline, right? You get an MQL,

Brian Bell (00:13:17): you qualify it, then you go to the SAL, SQL, and there’s a whole sort of sales pipeline.

Arun Penmetsa (00:13:22): That in our view is great and obviously every company needs to run it, but it’s very internal facing and ultimately it doesn’t really address what is happening in the customer during the sales cycle. So one of the things we encourage our companies to really do is in parallel build Build a customer journey that mirrors the stages in your sales pipeline. So thinking through, you know, what is the urgent pain that you’re solving? And we spend a lot of time with our companies on urgency because we really try to understand, is the problem you’re addressing truly urgent? Like, do they need the solution today?

Brian Bell (00:13:53): Is it a hair on fire problem or is it just like a nice to solve problem someday? I mean, there’s many, many sort of ways of saying it, right?

Arun Penmetsa (00:13:59): Like, like pain killer versus vitamin and all this kind of stuff.

Arun Penmetsa (00:14:02): So we document that. Then we think about, okay, you really found a hair on fire problem, but why would they pick you? What do you have to say in the sales cycle? And we call them wow factors. And there's many different names for it as well. And as you go through the sales process, we encourage our companies to, and we work with them on this. I mean, literally I've sat in like three hour long sessions going through each prospect. We encourage them to come up with these wow factors for every meeting that you're coming up with because essentially we need something that will get you to the next meeting, right? And then we talk about how do you convince the finance team? How do you convince procurement? How do you convince the executive team? What are the things you need to build into your sales process to do that? And it's not like we have a right answer for each of these questions but I think what ends up happening is a lot of this information is in the founders head when they start because the founders are the ones who are truly understand the market they're very passionate they're able to go sell and convince those initial customers but when you bring in a sales team or even the first AE oftentimes downloading those insights from the founder into a playbook that you can give to a rep and say go execute this because we have figured all this out is a big sort of stopgap in the growth of a startup. and we are trying to bridge that so a lot of our efforts are like we sit with the team with the sales folks the marketing folks the product folks the founders and then try to build this out when we invest right so that's sort of the I don't know really high overview

Brian Bell (00:15:33): yeah and a lot of founders need that help too right I mean because a lot of founders are usually a couple technical people maybe a product guy and a technical coder CTO kind of guy or gal and they need help on the GTM that’s very common

Arun Penmetsa (00:15:48): that’s right and in many cases the information is in their head it’s not structured in a way where you can just give it to someone right but it’s in their head and we see a few times this plays out right great team they hire a really good AE or a head of sales and the instinct of the founders is well it can go in both ways but one is they continue to be hands-on which can cause some friction because if you don’t have a playbook then the founder is like still selling or like there’s a different methodology or they say I’ve hired these great people I’m going to step back which is also a gap because that knowledge is now not passing on to the team so anyway that’s what we think one of the which is why you know I mentioned this concept of unlock growth right like we believe when you go from product market fit to scaling there’s like a step in the middle where you have to like unlock the growth before you can scale it so a lot of our methodology is focused on doing that now again I’m using a lot of terms here but yeah yeah I love to unpack the three

Brian Bell (00:16:42): stages that you just talked like you just breezed right through them yeah like one I’d like to like you know from a storm perspective the storm ventures what is product market fit look like for you guys yeah so I think for us product market fit ultimately goes back to urgency

Arun Penmetsa (00:16:56): right there are different ways to look at it how many customers you have how much revenue you have when we meet conversion rates churn rates exactly right and I think a lot of those are like symptoms of what you are essentially trying to achieve and for us what we believe is if you can truly find something that’s urgent and the customer desperately needs it I think that’s the first step to finding product market fit right I mean obviously I mean I don’t think anybody any decent startup is building something that nobody wants right but I think it’s

Brian Bell (00:17:27): At least half, you know, because half go out of business, right?

Arun Penmetsa (00:17:32): I think there are other reasons they go out of business, right? I think if you get customers, I think you’re building, there is value you’re delivering. But I think the challenge is, especially when a market turns, you are not in those top one, two, three, four things that they need, right? And we see this all the time. I mean, there are customers, We’ll keep piloting your product, but they never pay. You know, you see customers who say we don’t have budget, which is a genuine problem. But anyway, so I think if we can truly get to urgency, and that’s what we’re trying to find out through the first 10 customers. Like, have you truly struck oil? Right. And it’s not easy to figure that out. And, you know, if we knew how to figure that out, all our companies would be successful. Right. But I think that’s the way I think about part marketplace.

Brian Bell (00:18:15): You’ve seen a lot in 11 years. What do you think has changed about go to market over the last 11 years? I mean, you have a lot of AI, SDR, AI outreach, AI ads, AI this and that. What’s changed? What remains the same?

Arun Penmetsa (00:18:27): Yeah, I think I think what’s changed is how automated things have gotten. Right. I mean, 11 years is not that long a time, but, you know, a couple of decades ago, you had to get on a plane. You know, cloud marketplaces have come about in the last decade or so. So you can go to AWS Marketplace and say, I want to buy this.

Brian Bell (00:19:06): Yeah, I came from AWS Marketplace and Azure Marketplace. I ran the category team at Azure, actually. How are you guys leveraging that in your portfolio? We’d love to hear the latest.

Arun Penmetsa (00:19:14): We are in the early stages, a handful of companies. Actually, you may be, this is maybe a side discussion and we can do it later, but we’re actually investors in a company called Flywheel that is kind of building out the buyer side platform for cloud marketplaces. And that team came from AWS Marketplace as well.

Brian Bell (00:19:29): Oh, wow. Okay.

Arun Penmetsa (00:19:30): so they’re trying to automate that entire journey from when a procurement person decides to buy to connecting with the vendor on the other side automating the transaction and making it much easier to find more of these deals so yeah there’s

Brian Bell (00:19:45): some unicorns on the seller side too like tackle and yeah

Arun Penmetsa (00:19:49): Right so anyway so that’s a new phenomenon probably in the last decade so I think there’s a lot more automation in how you find information about your buyers reach out to them move them through the journey build out the playbooks and all that I think what’s not changed is still kind of going back to the basics about like I mentioned right like why are they buying your product how do you price things How do you think about the incentives of your buyer in there? How do you think about making them look good? How do you think about showing ROI so they can go to their CFO and fight for the budget that you want them to fight for? How do you manage churn? Even though there’s some automation there too today. So I think the way and the reasons people buy is still the same. You can move them through the process faster. You have more data to leverage today than maybe 15-20 years ago. I mean, there’s companies in the portfolio you can talk about, you know, there’s this whole category of revenue intelligence today that didn’t exist 15-20 years ago on how do you really think about each customer, the revenue potential, risk, expansion, all of that. So that’s some ways I think things have changed, but in other ways not.

Brian Bell (00:21:00): So what stages are you guys investing again?

Arun Penmetsa (00:21:02): So we mostly, so there’s some variants, but I would say the traditional target would be sort of late seed and series A. I like to call it traditional series A because we write checks sort of in the $2 to $6 million range.

Brian Bell (00:21:14): Yeah, which is kind of seed now. Which is kind of seed now, $50 million A’s.

Arun Penmetsa (00:21:16): Yeah. Fund 7 which we’re investing out of right now is 230 million so that would be the ballpark in some cases we’ve gone early like Flybeal was one where we invested at inception because we had a thesis in the space it’s a great team and they’re doing very well so in some cases we go early but that would be sort of the one I mentioned previously would be the sweet spotter

Brian Bell (00:21:41): Got it. And then what’s the kind of the scorecard look like internally for you guys? How do you guys make decisions? What’s important to you? And what might be differentiated versus other VC firms that you know?

Arun Penmetsa (00:21:51): Yeah. So I think there’s a lot of commonalities to other firms. I mean, obviously the team matters a lot. I mean, given that we invest early and we tend to be very long term oriented.

Brian Bell (00:22:01): Yeah. Somebody on my podcast, I forget who was like, well, I care about five things and the first three are people. Right.

Arun Penmetsa (00:22:06): Yeah. And I think that’s true in many cases, right? It also goes back to what I said earlier. Ultimately, They are driving the ship and no matter how much influence or control you think you have, it’s actually very little, right?

Brian Bell (00:22:17): Yeah. You’re more like a coach. Yeah. You’re on the sideline and you can’t even call plays. It’s not even like being a coach. It’s more like being like, I don’t know, some sort of junior assistant coach that doesn’t call plays.

Arun Penmetsa (00:22:29): I mean, you’re absolutely right. I mean, that happens a lot. And it’s interesting. Ideally, listen, I want to be involved and I I do tend to get involved in in some aspects with my companies but ideally you want a company that’s doing well where you don’t have to do anything right like that’s

Brian Bell (00:22:44): that’s what I saw my LPs is like you know the best companies don’t need my help they don’t need team ignite exactly I never hear from them I just get I get their series A B and C announcements and I’m like great right so that’s that’s the best case scenario yeah

Arun Penmetsa (00:22:59): So the number one, as you said, is team, right? Then we try to have a good thesis on the market. We look at, you know, this storm through its history has done 200 plus investments. So we have a good sense, although markets change really quickly. Right. We have a good sense and I would say network in some of these markets where we try to have a view on what companies are doing. And the third, I think, is sort of the go-to-market side. Like we really want to understand is the product they’re building and the team really well tied to the go to market they want to pursue and it’s not like you can’t like like if it’s if it’s if you’re doing a sales led motion then the way the product is structured and the way pricing is set up has to support a sales that motion like the economics have to work out and that’s a very simple example but we think a lot about those things and can we truly help because if the economics are off there, it’s just very hard to build a business. You may need a tailwind or something in the market to really pull you along. So yeah, those are maybe three or four things. Like I would say team, you know, the market slash urgency thing is something we think a lot about. yeah and then severity yeah yeah and then and then we think about sort of the go-to market there’s probably other factors like we think about like obviously we think about traction you know how fast you’re growing and all these other things but if I had to boil it down it’s probably those three things

Brian Bell (00:24:19): yeah I love that how do you guys assess you know as I look at a lot of early stage companies a struggle I have as an investor is assessing the bundling risk of the incumbents yeah you know I’ll look at something really interesting and I go you know that that’s kind of a feature inside of somebody else’s platform like how do you yeah

Arun Penmetsa (00:24:35): So one of the things we do quite a bit and is we spend a lot of time with the teams on their product roadmap, whether it’s new companies that we’re investing in or even existing portfolio because the markets evolved, right? We invested four years ago on a certain thesis and the market has shifted. So we spend a lot of time with the founders on how do they think about building this product in the next two, three, four, five years. And even though you have an insertion point today that feels like it’s something that could get bundled. Do you have the appropriate or at least you’re thinking about it the right way where you’re either collecting a set of data or you’re building into a certain set of workflows? Is there a moat here that will form over time or now? Yeah. Yeah. And people have different views on what is truly a moat, but I think ultimately, from a buyer perspective, if something is working really well and it’s getting their job done and making them look good, it takes a lot to remove that and put something new in, right? and this obviously comes up with AI all the time today right like every company so anyway to answer your question we try to spend a lot of time in product roadmap and we try to be thoughtful about it we try to think through you know how does the founder think about the next three years like have they thought about okay the wedge product and I know When you’re starting the company and you’re early, you have 100 things going on as a founder. And you may not have thought 18 months ahead, 36 months ahead. But that is something we really try to understand.

Brian Bell (00:26:01): What has changed over time, over the last 10 years, 11 years that you’ve been there? What have you guys kind of incorporated where you’re like, you know what, we made a mistake and we’ve learned that mistake. And now when we see that again, we don’t invest. Is there anything like that where you’re kind of like red flag, yellow flag? yeah not gonna bet we’ve seen the story before we know it doesn’t work yeah I would

Arun Penmetsa (00:26:22): say like go-to-market mismatch is something we’re very careful about and again it’s it’s not like a black and white answer because there’s shades of gray here but we’ve seen as an example and you know we can advise them but it’s not like we don’t want to invest and say change what you’re doing right because we want ultimately the founders to build what they authentically think is the right thing to do But as an example, we’ve seen companies that hire a sales team and sell $10k price point, right? And I understand in the early days you have to do whatever it takes to get customers, but then we looked at the product roadmap, there’s really no viable path to driving expansion, right? You have to move up market quickly because the math just won’t work. And I’m giving maybe a simplistic answer here, but I would say that’s one. The other thing, which is maybe an obvious one, is we tend to spend, I would say, a fair bit of time now with founders, especially founders. founding teams together. And this is maybe more my learning. Like in the early days, I was trying to do my first deals and I was trying to get them done. But I think just understanding founder dynamics, can the two, three, four co-founders, how they work together? What are their strengths? Where’s sort of the gap in the team that we may need to help support? That’s something we think a lot about as well because that’s probably the most painful one when you have founder conflicts after you invested because it’s just hard to recover from that kind of stuff.

Brian Bell (00:27:41): Yeah, you guys have done a lot of cybersecurity. What is some common things you’ve seen that creates a platform versus a feature?

Arun Penmetsa (00:27:48): Yeah, I think there’s a few things. I think one is, especially for the larger platforms, you know, do you have an asset inside the customer’s environment? I think that matters a lot, especially these days. And if you think about the bigger platforms like CrowdStrike, Palo Alto, Zscaler, I think them having that footprint inside customer environments helps a lot. Whether you’re in-line and you’re looking at the data constantly, I think matters a lot, especially now with AI and how the things are changing. Cybersecurity is always interesting because it feels like there’s always budget, but there’s like a million companies all vying for that budget.

Brian Bell (00:28:23): Yeah, and it shifts so much too. So it’s a great place to invest because it’s, you know, every three to five years, there’s just a new... a new edge that you have to protect exactly yeah and there’s a new piece of

Arun Penmetsa (00:28:35): software with an acronym that you need to buy but those are some things I think companies do well you know we recently one of our companies was bought by CrowdStrike and they had a really interesting product they started out sort of in the browser security space and expanded from there to endpoint but that was I think one of the big reasons they were bought is they were giving a perspective on what was happening on the device that was hard to get for the platforms

Brian Bell (00:29:01): what’s changed recently is you know I feel like The essence of what it means to be capital efficient has shifted, especially over the last five years. How have you seen that in your portfolio play out?

Arun Penmetsa (00:29:09): Yeah. So for the companies that we’ve invested in recently, there’s a handful that are growing well and quite efficiently. I think it ultimately goes back to what kind of product you’re building. If you can get a go-to-market that’s more product-led, I think it would be very, very I think if you can automate parts of the sales process through the AI tools I think you can get very efficient a lot of our companies do sell into large enterprises and in those cases you still need people to go and communicate and manage deals and there it’s harder but just the whole go-to-market process is getting more efficient with these AI tools so we’ve seen some of that It’s nowhere close I think at least what I’ve seen to what we see with sort of the B2C use cases like the lovables and sort of those companies and how efficient they can get prosumer kind of approaches kind of things but we have seen in a handful of cases just sales efficiency go up when the difference is you know if you are I think I think the other thing we’ve seen is the pace of product launches is getting faster across the portfolio which also means if you do a good job your net retention is higher so it’s okay to be a little bit inefficient early on but it doesn’t take you 12 months or 18 months to launch your next product you can probably launch your next product in three months as an example right so you can start driving expansion up much quicker so over time hopefully these companies will need to raise less but I still think we’re kind of in the very early days on the B2B side and how that plays out but I can see a path

Brian Bell (00:30:36): on how these companies can get much more efficient a big refrain I hear online is you know there is no technical differentiation do you believe that and what have you seen over the last five years yeah so I think it depends on what you mean by technical differentiation right and where in the stack you’re building I think I think if

Arun Penmetsa (00:30:53): I’m for the solution. But in general, they tend to be more horizontal and not best of breed in any individual one. So I think you can do that. If you go more to the application layer, I think it’s harder to build truly something that’s technically different because you own so little of the stack. But I think where you can differentiate, like we talked about earlier, right, is you get access to a data set that is hard to get access to. you embed deeply in workflow so you’re just managing a lot of what the customers are doing and you’re just taking that over right so you know it’s interesting for some of our companies they have this data set they start doing one workflow now they’re doing two five ten fifteen and it’s At some point, because AI is running a lot of it in the backend, even the end customer doesn’t know exactly what in their workflow is done by this company’s product, right? Because the results show up or the output shows up and you don’t fully know which parts of that workflow are automated. So the risk of pulling them out can be quite large because you don’t even know, like before it would be like, if I have a SaaS product, I know like this is the dashboard, this is the API. If I cut off the API, I know exactly this is what will happen. but you can get so far embedded now with AI that you just don’t even know and then obviously you can personalize to a company’s workflow you can fine tune to their sort of use cases and data so I think at the application layer you have to do some version of that I also think some verticals are interesting you know like if you think about healthcare and I do a little bit there where there’s all this nuance and regulation that is I think it would be it’s not the focus for the broader models to really go fix right and you need to spend a lot of time in this So even those are opportunities today.

Brian Bell (00:32:48): So if you were starting a company today, which layer on the enterprise feels most ripe for reinvention?

Arun Penmetsa (00:32:53): I think multiple layers can be. I think we’re very early on the application side. And I think that’s true of most waves, right? As I mentioned earlier, The infrastructure gets built out first because that’s where the initial opportunity is. People build on top of that. Otherwise, you have to go build a full stack yourself.

Brian Bell (00:33:08): You know what’s interesting about AI is it kind of started at, you might call it the, I mean, it’s foundational model layer, but it’s really the platform layer. But then it kind of like I think it’s it’s it’s just changing so fast that people have had to reinvent

Arun Penmetsa (00:33:30): You can see more of the stack quickly to realize the potential, right? It’s just really interesting because if you think about the cloud wave, which obviously was a very big wave, a lot of the benefits initially were somewhat invisible to most people, right? Like you had to go convince a company, You’ll get savings over the next two, three, four, five years.

Brian Bell (00:33:48): Digital transformation is what we call it at Microsoft.

Arun Penmetsa (00:33:50): Exactly. We make this huge initial investment, but there’s value. And I fully believe in that. There’s a lot of value. But even now, after all these years, it’s not like everybody’s converted to the cloud. We’re still sort of, I would say.

Brian Bell (00:34:01): No, there’s still the laggards out there. And I think we’re probably just right in the middle if you kind of do the crossing the chasm.

Arun Penmetsa (00:34:06): That’s right. We’re probably right in the middle of digital transformation. Early to mid-innings, right? Yeah. But I think with AI, because the consumers can see it, like it’s like the phone moment, right? The mobile phone moment, the iPhone moment.

Brian Bell (00:34:17): Yeah.

Arun Penmetsa (00:34:17): Really, the demand picked up so quickly that, as you said, people were building I am a And then, you know, eventually you get to the application stages. And I think we’re still early in the application. Obviously, there’s some great examples out there of companies doing really well, but I would say a small percentage of the workflows in the enterprise are truly...

Brian Bell (00:34:58): It still takes 10, 15 years for it to play itself out, right?

Arun Penmetsa (00:35:03): Yeah, well, I mean, we’ll see. I mean, a lot of people are betting that it’s faster, but it’s probably in that range, right? So, and especially if you think about other verticals, like as I mentioned healthcare or construction or oil and gas where every decision you make is very consequential it’s not like a and I’m sort of oversimplifying but like

Brian Bell (00:35:23): yeah you can’t have like one nine of accuracy you need you need like email that suddenly like went out it’s it’s like you know the reading of somebody’s life on a rig or something

Arun Penmetsa (00:35:32): that’s right So you are going to be much more careful, which is why we’re seeing this entire way of, you know, simulation labs and guardrails and all these other ways of companies. This is the tooling that’s getting built out, right? And then I think we’ll get to true applications. I think we are still very early in seeing if you go back to sort of the iPhone era, like seeing the big, you know, like the Facebooks of the world, like those, the Twitters.

Brian Bell (00:35:59): I mean, you could argue that Facebook and social media was kind of a late internet thing. And then, you know, and then the iPhone created a new platform and, you know, the Ubers, you know, and the other, you know, unicorns. you’re right yeah you know that took 10 years really that took a while right and if

Arun Penmetsa (00:36:15): you look at the cloud wave and you think about the data dogs of the world and all they took a few years to emerge as well so I think we’ll see those companies coming up hopefully some of those in our portfolio but but I think that’ll come out in the next couple of years right

Brian Bell (00:36:26): so what do you think people are missed you know underestimating uh over the next decade in the enterprise shift

Arun Penmetsa (00:36:34): I think some things will change in the way products are built and delivered and used. And I think some things will still stay the same. I don’t think we’ll see as much job losses as people are fearing. We’ll see some and we’ll see some shifts. There’s certain kind of tasks that will fully be automated. I think when you are buying products for your business or you’re buying from a business, there is a level of port and reliability and assurance and planning and all this that’s required always require humans. Now, what those humans are doing will change. And in fact, some of the early studies that have come out, right, are showing that, right, in some cases, AI has improved productivity, but in many cases, that time has just moved to some other work, right? So it’s like, it’s not like people have more time. They’re just doing more work now, right? Right. I guess you classify that as productivity in the sense that more work is being done per person so we’ll see how it plays out but I think we’ll always there’ll always be something to do as we build products especially especially in complex industries so so yeah so I don’t think we’ll see the level of job losses that people are

Brian Bell (00:37:40): fearing I don’t think so either yeah speaking speaking of you know market overreactions you know recently we had the SaaS apocalypse right where a lot of SaaS companies lost 50-70% of their value seemingly overnight. Do you feel like that’s just a market overreaction or do you feel like the incumbents like, I don’t know, Salesforce and other SaaS players like that are really threatened in this new wave of like...

Arun Penmetsa (00:38:04): Yeah, so I think AI is definitely a real trend that will impact these businesses. I think the stock market reaction is overblown. I think the way that SaaS companies operate will evolve. to incorporate AI. So the way I think about it, it’s not SaaS, it’s like AI plus SaaS. So it’ll just look different. And you still have to do these workflows. I don’t think AI will do everything for you for a while, at least. And depending on where you are in that stack, you’re at different layers, different levels of risk, right? If you are a pure UI layer analytics product, then you’re probably at significant risk because the product can be replaced quickly. And there’s a lot around the business, like I said, around support and reliability and all that. So it’s not like you’ll suddenly drop off a cliff. But if you’re further down and you control You know, if you think about Salesforce and where they sit and the system of record, we may see layers on top that will pull data from them evolve quite, quite quickly. But I think just displacing Salesforce is just going to be so hard like you’ve built on that for like 15-20 years and you know they’re obviously building their own new AI products on top so I think they’ll evolve so I don’t think it’s like the SaaS apocalypse the SaaS companies have to evolve into something different they have to leverage more of the tooling they have to like like I said go deeper into workflows and all that and some will survive and some won’t but I think a broad market correction across everyone is is the overblown

Brian Bell (00:39:35): yeah I’d love to wrap up with some Fast questions, rapid fire. What’s one technical misconception founders consistently have about enterprise buyers? Technical misconception.

Arun Penmetsa (00:39:44): I don’t know if this is the answer to your question, but one of the things that I see a lot is people underestimate the integration work that’s needed to get a customer live. And this happens a lot where they see signing a customer as sort of a big win. And it is a win and you should celebrate. I’d say that moment when you sign the contract is the moment of highest stress for your customer because they just agreed to work with the startup and they don’t know if this thing is actually going to work so one of the things we advise a lot of our companies is get the deployment done as fast as possible and it’s not if it takes weeks like it’s it’s just adding a lot of risk so I don’t know if this is a technical misconception, but I think not enough focus is put on deployment, integration, timeline, and we push a lot to shorten that.

Brian Bell (00:40:35): Speaking of deployment, there’s been a recent rise of forward deployed engineers in the field. Is that going to be more important, less important, or just as important in the next few years?

Arun Penmetsa (00:40:46): I think it’s going to be more important. And I think it’s a model I think more companies should take because as I said with the models the AI models not everything today is completely like it’s not like it’s not like software 100% predictable if this is what the code says it’s what the code will do so being in that environment understanding workflows trying to get your hands into more of the things that you can build quickly matters a lot and in some ways it’s the old model of having customer support and success but but with SaaS it was more straightforward you know what your product did you deploy it online it’s not like you have to do a ton of work after that but I think with with with the capability of these models, you should have people, especially in your large customers, trying to figure out, trying to make sure they’re successful. Like this is a big thing, right? And go to market, like it’s not selling the product. Ultimately, the goal is how do I make my champion look like a hero? Because if you can do that, that customer and champion will be very sticky and there’ll be a great reference for you. And I think that’s even more important given how quickly companies are growing and how rapidly sales cycles are changing with AI. So I think we should do more of that.

Brian Bell (00:41:48): Yeah. In cybersecurity, where’s the next CrowdStrike sized opportunity hiding?

Arun Penmetsa (00:41:53): I mean, I think a lot of it has to do with, well, maybe two areas, right? One, I think broadly across application security and how much AI can solve for that, I think will be a big trend in the next few years because that’s such a large attack system. surface and if AI can meaningfully solve a lot of the application security or address a lot of the risks in the application security I think that’s one I think the second one is more consumer is what happens to consumer security like our passwords our logins everything with AI like I think it’s authentic security

Brian Bell (00:42:24): basically

Arun Penmetsa (00:42:25): Yeah, because the volume of attacks have increased so much that how can we get consumers to better protect themselves? Because if you think we had scams before, now with AI, it’s like 10x.

Brian Bell (00:42:37): I saw a meme, it was a really funny meme on X. It was something like, all we have to do is just ask their AI really nicely for their password.

Arun Penmetsa (00:42:44): Yeah, yeah, exactly, right? So how do you build the next generation of consumer security companies that protect their users from the new threats. I think there’s probably a big company out there somewhere.

Brian Bell (00:42:56): Is there an enterprise workflow that you see that is still painfully manual and overlooked?

Arun Penmetsa (00:43:03): I think an enterprise workflow. I want to say something in the on the service maintenance side. It depends on the vertical too in some cases. I’ll give you maybe a slightly different example and I am going to list one of my companies here. You can talk your book on this podcast. It’s totally fine. There’s a company called Compa in our portfolio. It’s a really interesting company and they do compensation management and intelligence. I really like this use case because if you think about it, what compensation teams do is incredible work and compensation tends to be 80% of our company’s right paying people but typically these teams don’t have a lot of people the compensation teams are small right and they do a lot of manual work to figure out how much should an offer be How do I plan for the year? How do I plan out my budget for hiring for this year? They often have to go work with sales, marketing, engineering. How many people are going to hire? What should we plan for this? Oh, the market for AI engineers changed. Now we have to offer, I don’t know, million dollar bonuses or something, right?

Brian Bell (00:44:05): I think there’s this interesting thing too where, you know, I think it was about in the 70s. I think Congress passed a law or the SEC passed a law where you had to disclose executive salaries. I think this was around like 1970 and you know from that point forward you saw this just astronomical increase in executive salaries it was like why is that well it’s because executives could look and see what everybody else was making you literally had to disclose it if you’re a public company any officer of the company exactly what their stock options are exactly what their comp is right And I think something similar has happened in the last 10 or 15 years with online tools like Levels and others where you can actually see, oh, a Level 7 engineer at that company makes that. I want that. So creating the transparency has actually increased the worker’s ability, the employee’s ability to negotiate. Right.

Arun Penmetsa (00:44:56): So anyway, I think compensation is one workflow which is heavily manual. It’s not a broad answer to your question, but I think it’s an interesting one because it’s often not thought of as much.

Brian Bell (00:45:06): Yeah, some of the best startup ideas are just kind of taking things that are just done manually and just AI-powered SaaS-ifying them. Exactly, exactly. Like, oh, they’re doing this in spreadsheets. Great.

Arun Penmetsa (00:45:19): You take healthcare or oil and gas like you said I mean there’s some back office that’s responsible for getting a wheelchair out to a patient right or getting a piece of equipment delivered at the right time and it is mission critical but it’s buried somewhere in the entire workflow that you never think about yeah

Brian Bell (00:45:37): What’s an early stage metric you trust the least? That I trust the least?

Arun Penmetsa (00:45:41): I would say, I think ARR has, the definition of ARR has changed so much over what it originally meant, right? Like true recurring revenue that’s signed and I think we see ARR put in every deck today but what it means varies a lot between being usage based or it’s contracted but not live this is if you take our monthly number and annualize it like this is how much it would be so there’s many variations of what the ARR number is so I think you have to dig in a little bit to understand what exactly yeah what are we talking about here

Brian Bell (00:46:18): about some of your highest usage day last month and you’d multiply by 365 that’d be probably the most egregious yeah so

Arun Penmetsa (00:46:23): so that I would say is is one that yeah that’s funny if you you could

Brian Bell (00:46:27): institutionalize one habit and every early stage founder what would it be

Arun Penmetsa (00:46:31): I would say I think most early stage founders do this already, but I would say as a company scales to continue staying very close to the customers, you know, and so this is maybe not very early stage, but if you get to a B, C rounds and your role changes and it should change, you have a full team. I think the best companies the founders are still very involved with customers they still do customer calls they take support calls they go meet because you never lose that pulse on the business I would say that you know one of the one of the things that does happen sometimes is you have a founder when they’re early they know every aspect of the business every number every customer and obviously you can do that as you scale but as you put layers between the founders and the customers that perspective changes and if you go back to something I said earlier about having the product vision one of the one of the best things that a founder brings to the table is their vision for how this market should evolve and what product needs to exist in it and how do you build towards that future and I think the further away you get from customers you lose a little bit of that vision because you’re not just getting the raw data anymore so A lot of founders do this already but if I had to say that’s something important that you should continue doing

Brian Bell (00:47:41): Is there a vertical in SaaS or enterprise that you feel is underappreciated or inversely overcapitalized right now?

Arun Penmetsa (00:47:53): I think I would say retail is probably underappreciated to a certain extent to think physical stores and you know this may be unfair I’m sure there’s a lot of a lot of retail startups now that that are like what do you mean but you know if you think about physical stores gas stations convenience store like like that’s a huge footprint across the country and they are slow to adopt there’s obvious reasons why it’s hard to break into that but that’s probably an underappreciated one I think go-to-market tech broadly is overcapitalized there’s so many startups building thought of it and there’s good reasons right it’s a very clear workflow it’s easy to test easy to get into but that’s what I would say my first instinct was to say healthcare but I think healthcare is not undercapitalized there’s a lot of healthcare startups out there it is a very difficult sector to get into and just manage everything that goes on in that sector but yeah that’s what I would say

Brian Bell (00:48:40): What’s something you’ve changed your mind on since you started investing? You were kind of sure it was true and then you’re like, no, it’s actually not true.

Arun Penmetsa (00:48:50): I think we tend to underestimate market sizes. Early in my career, I would be like, this can’t be a big market, right? I would do some diligence and be like, but I think this is not true of everything, but I think a lot of markets grow over time. So there’s value in in being a significant if you can capture a significant pocket of a small market as long as you as long as the market grows and there’s a tailwind I think you can build very big businesses and I didn’t appreciate that enough in the early days so so when we think about markets we try to be very careful about thinking bottoms up

Brian Bell (00:49:24): you kind of start from like first principles like how many people have this problem or how many good companies how much would they pay to solve it over what period of time kind of thing yeah

Arun Penmetsa (00:49:32): Exactly like that right but but I think I think because you’re especially in markets that are early that are late to tech adoption because such products have not existed they even the market doesn’t know that there’s a lot more you can do right like you you launch your first automation feature and they’re like this thing can actually be automated like like let’s take like and I’m you know let’s take oil and gas since you mentioned that earlier right and I don’t know much about oil and gas so I’m just sort of talking randomly here but like very manual very complex but I do think if you bring automation There’s probably things that will open up over time, whether it’s finding the location to drill in, whether it’s the right infrastructure to build, whether it’s around shipping, whether it’s about analytics, that you may be able to do things that are not possible now because of the technology that came in.

Brian Bell (00:50:21): You can bundle inside the vertical, basically. Yeah.

Arun Penmetsa (00:50:24): And I think your initial hypothesis on how much you could sell a company’s products for can grow. I mean, the price point can grow significantly. That may not have been obvious.

Brian Bell (00:50:35): What is a good rule of thumb from like a 10, $20 million company to a billion dollar company? How much do they expand their ACV over time over those five to 10 years?

Arun Penmetsa (00:50:44): Yeah. So I would say we’ve seen companies that have grown at least like next.

Brian Bell (00:50:50): Yeah, so it started with a 10,000 a year product and then went to a 100,000 a year product or 100,000 to a million.

Arun Penmetsa (00:50:55): Look at the cascade, right? You’ll get a few customers who will pay you a million because they’re at scale and all that. But I think I think there’s two ways to play this one is you find a very efficient go to market to land and if it’s truly efficient I think you can get away with like small price points but if you have a good product roadmap and you know you can get expansion then it’s okay to be inefficient upfront because your net retention will grow significantly right so your net potential will grow your revenue significantly so you can do either ones but I think it’s hard if you stay at a low price point because then you need tens of thousands of customers right and obviously there’s those companies that have sold and gotten at scale it’s just it’s just hard so yeah I think markets tend to be bigger than most people think they are at least in the early days because technology changes the buyer behavior and what they buy in those markets

Brian Bell (00:51:46): I’ll ask kind of a similar question in a different way what’s something that you believe to be true that’s a contrarian belief that most people believe to be false

Arun Penmetsa (00:51:54): and most people wouldn’t agree with you I’ll give you two answers I don’t know this is exactly what you’re looking for so one is like the job loss thing right I don’t know if it’s truly contrarian but like I agree with you there I don’t I think

Brian Bell (00:52:04): there’s actually going to be more jobs than ever created in the next 10 years yeah we’re going to invent so many new things to do yeah yeah

Brian Bell (00:52:12): and people just tell me it’s kind of like trying to explain to farmers what we’re going to do in factories right in the 1800s they’re like so wait we’re going to be in a building like you know on an assembly line like with a wrench in our hands like I don’t understand why that farms well won’t all the machines be building everything yeah no no there’ll be like things to do like we’ll be selling the machines to each other and servicing them and like maintaining them and it’ll be hard to explain to people what we’ll do in a post AI world like it’s just hard to explain

Brian Bell (00:52:41): It’s hard to have that imagination too.

Arun Penmetsa (00:52:43): Right. Because it changes so quickly, right? Like, I mean, imagine technology change over 20, like, I mean, in 1990s, can we, would you have imagined everything we have today?

Brian Bell (00:52:51): And the jobs are going to be so much more incredible because all the mundane, bureaucratic, lined up in cubicle work is going to disappear. And it’s going to be very creative and very collaborative and everybody’s going to have so much leverage now with AI. That’s right. Exactly.

Arun Penmetsa (00:53:06): So that would be my main thing. I think the second thing is, again, I don’t know this contrarian, but I think this is one of the best times to invest in venture. I mean, there’s obviously fear globally with prices being too high. We got a war now in the Middle East. Exactly right but I think the pace of innovation is so high that early stage companies somebody asked me the other day so many people are starting companies and it’s getting easier to start a company like should I start a company now like everybody else is starting a company and it’s so competitive and at least my view is I think it’ll get easier and easier to start a company like I think waiting doesn’t help if that’s your concern

Brian Bell (00:53:43): That’s what I tell young people they’re like hey should I get an MBA and I’m like nope start a company should I go work for Microsoft I’m like nope start a company it’s just never been easier to just start a company get somebody you like to work with and go talk to the customers and figure out what their problems are and go build it for them it’s as simple as that that’s right and now you can vibe code it you can just get cloud code to you know go figure it out and then you’ll just get more technical over time you’ll go deeper in the stack and you’ll expand your AECV all that stuff as you mentioned earlier

Arun Penmetsa (00:54:10): the learning around how do you sell how do you service a customer how do you manage your finances like all of that is so valuable no matter what you do in life right like being selling is a good skill no matter what you do in life right and VCs are just

Brian Bell (00:54:25): I tell people VCs are just salespeople yeah yeah We sell founders in taking our money and we sell LPs on giving us money. That’s basically a sales job. And we sell people to come on podcasts so we can learn for free and get a free education. But I really enjoyed the conversation. I learned a ton. I love doing this podcast because I just learn. I feel like every time I have a VC on like you that’s been doing this for a long time, longer than me, I take away nuggets and I get to improve my my craft my skills so thank you so much for coming on I learned a lot thank you thanks everyone all right take care bye

Christian Schroeder has seen all three—at scale.

Before founding 10x Value Partners, he was dropped into frontier markets with Rocket Internet in his early 20s, tasked with building e-commerce businesses from scratch. Not optimizing. Not iterating. Building from zero in places where customers didn’t even know online shopping existed.

That experience shaped how he thinks about startups today: speed matters, but clarity matters more.

Here’s what stands out from his approach.

Subscribe now

Talent beats experience—if you know what to look for

Christian doesn’t optimize for resumes. He looks for people who think clearly and move fast.

In interviews, he watches for two things:

• Can you explain something complex in a few sentences?

• Can you back up your claims with numbers?

Most candidates fail here. They talk too much, stay vague, and avoid specifics.

Strong operators do the opposite. They give precise answers and quantify impact. Instead of saying “I grew revenue,” they say “I increased revenue by 35% through X and added another 20% through Y.”

That level of clarity signals real ownership.

Another simple filter: if the conversation runs out of substance in 15 minutes, it’s probably not the right hire.

Founder-market fit matters more than the idea

Early in his career, Christian followed the classic venture studio model: find a working idea, replicate it, scale it.

It worked—but not always.

The failure pattern was clear. Founders weren’t deeply connected to the problem. They executed, but without conviction.

Now, he flips the process.

He keeps a shortlist of markets and ideas he believes in—longevity, AI automation, education—and waits for the right founder to match.

That shift sounds small, but it changes everything. Execution improves. Persistence improves. Outcomes improve.

Most venture studios do too much

A common mistake: trying to build too many companies at once.

Christian takes the opposite approach.

• 1 to 2 companies per year

• Deep involvement

• Clear thesis and advantage before starting

This comes from studying what actually works. The best venture builders don’t behave like accelerators. They act like focused operators placing concentrated bets.

If you’re building, the takeaway is simple: more shots don’t always increase your odds. Better shots do.

“Unsexy” markets create outsized returns

When most investors pile into the same trend, returns compress.

Christian looks elsewhere.

Historically, many of the best companies started in markets that were ignored:

• Government and defense (Palantir)

• Space hardware (SpaceX)

• Financial infrastructure before fintech became mainstream

These weren’t popular when they started. That’s exactly why they worked.

Less competition. Better talent concentration. More room to build something dominant.

If you’re a founder, this is worth thinking about. The obvious idea is rarely the best opportunity.

Longevity is real—but the business model is unclear

Christian has a personal reason for focusing on longevity.

At one point, he discovered his biological markers suggested he was aging far faster than expected. That forced him to study the space deeply and experiment aggressively.

He reversed it.

Now he’s looking at longevity as an investor—but with a critical lens.

The problem isn’t demand. People care about health and lifespan.

The problem is feedback loops.

If your product takes 10 years to prove results, adoption becomes difficult. That’s why he’s more interested in products that deliver:

• Short-term, measurable benefits (energy, sleep, performance)

• Long-term health improvements

There’s also a surprising angle: pet longevity.

Shorter lifespans mean faster validation cycles. Owners are willing to spend. Regulation is lighter in many regions.

That combination could create faster-moving businesses than human healthcare.

AI is accelerating everything—but margins may not last

There’s no question AI is changing how companies scale.

Startups are hitting $100M ARR faster than ever. Products distribute themselves. Teams are smaller.

But there’s a catch.

When production becomes easier, expectations increase.

If AI lets you create one logo in five minutes, customers will expect 50 options instead of one. Over time, that can compress margins back down.

So where does value accrue?

Christian sees two paths:

• Software layer: building tools that scale fast

• Operator layer: using AI to improve margins in existing businesses (roll-ups, services, operations)

He’s particularly interested in the second. If AI can significantly increase margins in traditional industries, owning the underlying business may be more valuable than selling the tool.

Small improvements compound into massive outcomes

One principle he repeats often: improve by 1% every day.

It sounds simple, but it works.

Compounding small gains over time creates outcomes that look like step changes from the outside.

Most founders look for breakthroughs. The better approach is consistency.

Founders misunderstand VC feedback

One of the more blunt insights: most feedback from investors isn’t the real reason they passed.

Sometimes the fund has no capital left. Sometimes the partner isn’t convinced but doesn’t want to say it directly. Sometimes it’s just not a fit.

If you adjust your entire strategy based on that feedback, you risk moving in the wrong direction.

Better approach: listen, but don’t overcorrect. Stay grounded in your own understanding of the business.

Don’t run out of money

It sounds obvious, but it’s still one of the most common mistakes.

Christian aims for at least two years of runway. That gives you:

• Time to iterate

• Leverage in fundraising

• Space to make better decisions

If you’re raising with only a few months left, you’ve already lost negotiating power.

Final thought

Christian’s career runs through very different environments—frontier markets, venture capital, company building—but the pattern is consistent.

Focus on people.
Stay precise.
Avoid crowded thinking.
Build with intention.

The founders who win aren’t chasing trends. They’re building with clarity in markets others overlook—and executing better than everyone else.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:
00:01 Introduction & Guest Overview
00:24 Christian Schroeder Origin Story
02:27 Rocket Internet & Frontier Market Experience
03:33 Key Lessons from Rocket Internet
05:32 Hiring Top Talent & Interview Signals
08:34 Applying Hiring Principles Today
09:48 10x Value Partners & Investment Approach
11:58 Longevity Thesis & Personal Health Journey
16:01 Telomeres, Biohacking & Reversal Strategy
20:50 Longevity Market Opportunities
23:32 Future of Longevity & Regenerative Science
27:49 Venture-Scale Business Models in Health
28:59 Evolution of 10x Value Partners
30:57 Venture Studio Strategy & Market Timing
32:28 Founder-Led Investing Philosophy
35:24 Venture Studio vs Accelerator Model
38:32 Macro Trends & AI Landscape
41:26 AI, Margins & Future Business Models

Transcript

Brian Bell (00:01:00): Hey, everyone, welcome back to the Ignite podcast. Today, we’re thrilled to have Christian Schroeder on the mic. He is a German serial entrepreneur, angel investor, company builder, founder and CEO of 10X Value Partners and Utopia Capital, and hands-on partner to founders building impactful early stage companies. Thanks for coming on, Christian. Yeah, thanks, Brian, for having me. But let’s just start with your origin story. What’s your background?

Christian Schroeder (00:01:20): Coming from a background of kind of an entrepreneurial family. So when I was growing up, we would always talk about like business and like the challenges of running entrepreneurial organization at the kitchen table, at the dinner table. Yeah, I mean, like that kind of gave me the mindset of like fingers. It’s kind of like looking at opportunities everywhere where you could find them. I studied business. I started my career as many people in like banking and private equities of the borrowing industries, but pretty quickly kind of like went back to like my roots of entrepreneurship. and had the opportunity to join an organization called Rocket Internet, which in the 2010s kind of like really drove venture building at an industrial scale globally. So we launched businesses like Rideshare, sharing e-commerce and food delivery in 50, 100 countries, predominantly Latin America, Southeast Asia, South Asia, the Middle East and Africa. I had the opportunity to be the pioneer for e-commerce the very fringe markets like the frontier markets that that usually you don’t associate with like a holiday destination so my colleagues were lucky to get kind of like Indonesia and the Philippines I build e-commerce in Bangladesh and Pakistan and well we also had Sri Lanka Sri Lanka was nice kind of but like what was interesting about this is that like we could really be the pioneers for The e-commerce in this region was like barely anything. It was like a green field and we basically had to even teach people like What kind of like it means to order something online that that’s even possible. So I did that for a couple of years. I was a bit tired, to be honest, Brian, to travel every week with a small carry on from country to country to make sure that operations were running. And I wanted to do something different. So I moved into the headquarter of Rocket Internet and joined the Venture Capital Fund. Global Founders Capital, and then spend another two years or so looking at thousands of business opportunities, point of view of investing, but also from a point of view of like, is this an interesting model that we could bring to a new market with our venture studio? And yeah, I did that for a little bit longer, but like the last eight years, I’ve been more focused on investing. So I’ve invested in around 40 companies in the private capacity as a private investor and also through my my venture studio, 10x value power. partners, created around 10 companies from scratch. And yeah, I mean, I’m very excited about solving problems that have a meaningful impact for society and the world. And that keeps me going, I guess.

Brian Bell (00:04:02): Yeah, it’s a fascinating background. I mean, Rocket Internet’s infamous, right? You know, the thesis there was, hey, let’s take Uber, which is taking off in San Francisco and kind of go build it. in europe or africa or asia what was your what was your favorite experience at rocket that you really learned the most from that you kind of still apply today i

Christian Schroeder (00:04:20): think like i think something that i still apply today is kind of like this bias towards hiring very talented people and over indexing on talent and less focusing on experience because the thesis It’s there that like young, hungry, driven people are like smart and capable enough to figure it out versus relying on the experience and then basically doing things that the way they were conventionally done. That is something that I’m still applying. today in many of my businesses. And I encourage founders specifically to bring on people like an entrepreneur in residence or founders associate to give them this like driven talent in their organization. I think like something that I’m really looking back to kind of is like the The experience of being parachuted into a company, it was like 10 people in Lahore in Pakistan. And like there were like some operations going on, but the processes that were really not scalable and to very rapidly assess the situation. and then come up with a plan of how to make something scalable where there’s a market demand, but you need to also execute efficiently. So that’s kind of like something as an experience. I’m very grateful for that because back then I was 22 years old, just out of undergrad, and getting this opportunity to basically take on huge responsibilities. It’s a great school.

Brian Bell (00:05:44): Yeah, I hear this a lot. And it’s almost like cliche, like hire A-level talent that will attract other A-level talent. What kind of tips and tricks have you learned along the way to spot? Because typically, you have the resume, you have the LinkedIn profile, you’re like, okay, the person person’s probably pretty smart. And then you meet them, right? And you’re kind of asking them questions and having a conversation in an interview kind of format. Where are you kind of like leaning in in those conversations versus leaning out? And you’ve done this a lot at Rocket Internet. You’ve done it as a VC. And now you’re doing it at 10X, both as an investor and like a venture studio. And so you’re Your whole job, like all your jobs have kind of really are predicated on that ability to spot, find and nurture talent. Where are you kind of like leaning in with all those hats and where do you kind of lean out?

Christian Schroeder (00:06:32): Yeah, I think like there’s a couple of things to look I think great people, they’re always very good at being extremely precise in their answering. So like people who don’t really know what they’re talking about, they give you a long answer. But I think this saying, there’s a saying that giving a short answer, it takes more effort. I think Einstein said this or like some like very famous

Brian Bell (00:06:54): Mark Twain said, I would have written you a short letter, but I didn’t have the time, you know, kind of thing.

Christian Schroeder (00:06:59): Exactly. That’s what I mean. Thank you, Brian. And I think this is something really when in the interview where you can pay attention to or when you meet someone, are they able to give concise answers? Secondly, are they able to back up or they are saying with numbers so like there are people who say like I grew revenue and then there are people saying I grew revenue by pulling level A and B and this contributed 35% in terms of like revenue expansion and then we added a new region which added another over 20%. So like people who are like very data driven and numbers oriented, usually they excel in kind of like scaling business. It’s not like the zero to one type of person, but kind of like when hiring at Rocket, we look more at like scaling something which is already proven that had product market fit because we took a business model that was already working. So we really like over index sort of on being on the data drivenness of people. And I think like Another factor where I said like I like people who give concise answers, I would say like if you don’t end, if there is no end to what you can talk about with people because they have so many interesting stories to tell, so many different experiences that are relevant to the role you’re hiring for, I would say that’s also a green flag. Usually if an interview is over after 10-15 minutes because you don’t have anything to talk about anymore, you just keep going another five, 10 minutes out of politeness, it’s like not a good sign.

Brian Bell (00:08:31): Yeah, these are all things that like, you know, I’ve hired probably 100 people in my career, 100 plus, I’m not even sure how many at this point. And I’m actually going through the interview process right now, hiring a portfolio manager for Team Ignite. All these lessons, I know all these things that you’re saying intuitively, but it’s great to hear them like explicitly in a way. And sit here and talk about it because these are all the same things I’m kind of like doing right now as I interview for a portfolio manager. Because we have 300 portfolio companies, just like so much to manage. And I got as far as I could go with like AI and automation and a contractor, but now I need to actually hire like a person. And so I’m going through the interview process right now. So it’s really, I’m almost like selfishly asking you how to hire the best people. But yeah, these are things intuitively that are great to hear, right? That you kind of get a sense of. You kind of get a sense for somebody.

Christian Schroeder (00:09:18): Yeah, makes total sense. Look, Brian, that’s why I like to read books, because it helps you realize lessons that you’ve already subconsciously learned, but that you can kind of like reiterate on through the reading of the book, right? It’s a great lesson.

Brian Bell (00:09:36): Yeah, I love that. So... You’ve been an operator, you’ve been a VC, and now you’re running a venture studio. Tell us about your thesis. What are you up to these days? You’re building out 10x value partners. Is it just a venture studio? Are you also investing? Are you doing both? Tell us about what you’re up to now.

Christian Schroeder (00:09:52): Yeah. So I mean, like it is not like really like clear cut. So I mean, like you can say like I’m doing my investing privately or I do through 10X value partners. But in essence, it’s just like my own capital. I had a few exits. I was fortunate enough and I’m running the venture studios like that. So we do both investments and also venture building. On the venture building, we’re taking very concentrated bets. So we’re doing a one or at most two companies a year. And what we’re looking for here is like large markets, like it’s also very founder driven. So we had around like 10,000 applications for founders over the years. And something that we learned from Rocket is that we really want to be very like focused on the founder product market fit and not just as a product market fit because we had like some issues in the past with Rocket where basically the founders were not really passionate about what they were doing. We have like this kind of like repository of ideas with like 10, 10, 15, 20 things that we really like and where we’re just waiting for the right person to come to us to work on something. So right now, very passionate about longevity, alternative health. I think that is like an industry waiting to be disrupted.

Brian Bell (00:11:10): Well, I mean, if you want to, let’s talk on longevity for a sec, which is a topic near and dear to my heart as well. I think I read some articles that you wrote about longevity. Was it you that did the telomeres lengthening?

Christian Schroeder (00:11:22): Yes, that’s right, Brian. So that was actually kind of my full-time job in 2013, 14, almost.

Brian Bell (00:11:29): Yeah.

Christian Schroeder (00:11:30): No, I mean, I got into this biohacking kind of step-by-step. So in 2016, 17, I worked super hard, 100-hour work, eating pizza every day. I gained a lot of weight, no surprise there. And I started like saying, okay, I really need to like kind of like look at my health. And first I just started going to the gym. But then if you have this entrepreneurial mind, you start thinking about how can I do this more effectively? How can I make progress faster? You start looking at your hormone levels and like optimizing your sleep and so on and so on and so on. yeah i mean like over the over the years it got more and more sophisticated than in 2022 23 i i did like a test to to see like the length of my telomere so for the viewer telomeres is kind of like a part of the dna the genome string which basically um keeps the helps the cell to keep its integrity and perform its functions as cells divide so with every cell division the telomere would would shorten. And after several shortenings, it’s called the Hale-Flect limit. Basically, the cell would lose its ability to continue replicating.

Brian Bell (00:12:43): Yeah, you start getting mutations in the DNA replication as the cells divide. It’s kind of like aging, basically, right?

Christian Schroeder (00:12:50): It is like one of the causes of aging. And people thought like 10, 15 years ago, that’s the main cause of aging. It’s not really what the science tells us today, but it’s still like a driver of aging. So telomere length is like something you would want to be concerned about. Fortunately, like Brian, I found out that my telomeres were extremely short. So I was in the one percentile for my age groups. It means like out of 100 people, 99 have longer telomeres. And if you would I would have had the telomere length on average of a 70-year-old at like 30 years of age. So I was like really concerned. And I started consuming all sorts of like literature on telomeres, what causes telomeres. I mean, probably like 50% of your listeners, Brian, have short telomeres because stress causes telomere shortening. All startups, CEOs are likely affected.

Brian Bell (00:13:47): Well, yeah, I mean, if you’re working 100-hour weeks, you’re eating junk food. you’re overweight and you’re stressed out, all of these are probably contributing to shortening them, right? Yes.

Christian Schroeder (00:13:57): And I mean, like, usually, I mean, what I was surprised about is like the severity of things. I think it can also be related to long COVID. So there is like some studies that suggest that COVID can shorten your telomeres. But like, I think like there is also like, yeah, I mean, like other aspects here that maybe I’m not aware of. So, but because like if you look at, for example, studies that show that severe alcoholics, they lose like three, four years of telomere aging. So to lose like 30 plus years, it’s really surprising. But anyways, so I ducked really into the literature. I learned also how we can increase the telomere length for like, for example, the enzyme activation of telomerase through... other things like the clearance of senescent cells.

Brian Bell (00:14:44): So basically like autophagy, like fasting, intermittent fasting, things like that.

Christian Schroeder (00:14:49): Yeah. And then taking cellulitics like phycetin, spermidin, resveratrol and high dosages that have to clear out the senescent cells, but then also to stimulate the Yeah, I mean like telomeric lengthening with like some supplements that you can take as well as also the red light laser therapy. I think that helped. I didn’t do any hyperbaric oxygen chamber, which treatments that can also facilitate telomere lengthening. But this is kind of like the cocktail of things you can throw at it. I’m quite lucky now that my telomeres are super healthy. I’m now in the 99th percentile for my age group. So it is really like a 60% length increase in the telomeres. It’s very significant.

Brian Bell (00:15:35): It’s really interesting because I always assumed, or maybe I thought an error, that the telomeres were just irreversible. Like once you lost them, they were lost forever. But it sounds like you can actually

Christian Schroeder (00:15:45): to do certain things to lengthen them yeah there are also some peptides such as i think epiphalilone i don’t know how to pronounce about 157 peptide have you heard of this one bp157 it’s more like for like accelerated healing and joint health so if you have like an injury from sports and you want to look into like there’s no medical advice here but it could yeah no medical advice here we’re just we’re just chatting do your own research Yes, it could potentially be used for like accelerated wound healing. But then there are also like so-called bioregulators where we have some studies to learn from telomeres. So there’s like a lot of things you can throw at the problem really. I’m in the position now where I don’t need to do that anymore. But yeah, I think it was an exciting journey. And I would like to really partner with entrepreneurs, even from an investing point of view, but also from a venture building point of view, who are doing something in the longevity, health, biohacking space. I think like if you look at US GDP and GDP, what it is spent for. It’s like healthcare is way too much and there’s very little if no efficiency gains and it’s just a system that is out of control. I’ve been thinking about this a lot.

Brian Bell (00:17:00): GDP is a really poor measure of economic viability because it’s like you cure cancer and it would lower GDP. Right. Or, you know, like if you if you solve aging, it would lower GDP in the short term anyway. I think in the long term, it would probably raise it.

Christian Schroeder (00:17:14): Yeah. If you have more people because you solve aging, so your population would like grow exponentially kind of.

Brian Bell (00:17:20): Yeah. And so like you think about like longevity escape velocity, Ray Kurzweil has been pretty consistent that we’ll we’ll reach that around 2032 is the early 2030s call it. And if you think about that, like if we solve aging, In the short run, it’s going to lower healthcare costs, right? So GDP will actually shrink. But in the long term, it should increase GDP because people will... create more value over a longer period of time. That’s kind of an interesting measure.

Christian Schroeder (00:17:47): Look, I think like if you would, I mean, I think GDP is like not a great measure for like economic prosperity. So like, I mean, like the most extreme version of this, if you would nuke all the major cities in the US, it would grow GDP kind of because you would need to rebuild everything. But I mean, of course, you evaporate the capital stock quite literally. So Like basically you should more think about like in terms of like capital stock, what’s the wealth that the nation has accumulated and kind of like what is the prosperity that they can derive from that capital stock versus like looking at expenditure and If you look at kind of what’s happening in the world, a lot of the GDP growth, especially in Europe and Germany and other countries, it comes from public sector spending. And that is just like money that’s just taken out of the pocket of the taxpayers. And we know, or at least it’s my thesis is that the private sector is just a more efficient allocator of capital, putting it towards usages which drive more economic growth. So I think this is kind of like really like a flawed way of looking at economic prosperity just in terms of GDP.

Brian Bell (00:18:55): Yeah. So you’re investing and growing companies in this space now, in the longevity space?

Christian Schroeder (00:19:00): Yeah, I haven’t really found the right investment yet. I’m quite excited about doing something potentially with regards to more like from a... The problem I have with longevity is really like, I think longevity for pets is a larger market than longevity for humans because like the feedback loop is just much shorter. I think a lot of people...

Brian Bell (00:19:22): Your dog only lives 10 or 15, maybe 20 years. And so you can... And you’re probably more willing... I don’t know what the laws are here in the States on this, but you could probably be a little bit more bleeding edge on longevity meds for your dogs than you can for humans. Are there FDA trials for pets? I’m not even sure.

Christian Schroeder (00:19:39): I think like it varies country by country, but like at least in the uk it’s also the case that you can prescribe like approved drugs off label so i took like a topical rapamycin off label to cure potentially cure my boldness didn’t work but like i’m i’m now like after that doing the telomeres i’m throwing stuff at my boldness and yeah i think like For the pets, definitely should be able to prescribe anything off-label as long as you give appropriate warnings that it might harm the dog or the cat. So you can do much more with that also.

Brian Bell (00:20:15): I think there’s like, I don’t know how many dogs there are in the States, like how many dog pets there are, but It’s tens of millions at least, right? And you think about like, you know, we have a couple dogs, right? Little Maltese, Shih Tzus, Malshies. And I think about how much my wife would probably pay to keep them alive. Like if they’re at the end of life, you know, and how much people do pay. I’ve heard stories of people spending 10 grand, 15 grand on a surgery to extend their dog’s life by even a year, which is crazy because you can just go get another dog, but people get really attached to their pets. So I think your thesis is absolutely spot on, which is people are going to spend thousands, if not tens of thousands of dollars to keep their pets young. Have you looked into Yamanaka factors and kind of what they’re doing out of some of the labs in San Diego and definitely the one out of Harvard with Dr. David Sinclair?

Christian Schroeder (00:21:04): Yeah, I mean, I’m in general familiar with the Yamanaka factors. And I also heard about, I met an entrepreneur a couple of years ago, was looking to kind of like reprogram cells within the body individually. So like in vivo, I mean, theoretically in vivo, using the Yamanaka factors to drive younging. I believe that also there are some herbal supplements that like kind of can influence. the same pathways the Yamanaka factors use. I think one of those was Kusumin, but please don’t quote me on this. It’s just like based off memory. So I think like that is definitely interesting. I think like the risk of the Yamanaka factors or the main problem they had like the basically you need to I mean like you need to kind of like do a partial reversion of the age to not go back to like the photos embryo status because then yeah and you basically like lose everything so I think that’s the challenge they wanted to solve is to use like the mechanism of the Yamanaka factors to do like a partial age reversal do you know like what what is the San Diego lab doing Brian

Brian Bell (00:22:17): I forget his name, but he’s the... Well, Yamanaka Factors is named after the doctor in Japan that discovered them. And he discovered them 20 years ago, I think. I think the paper is like 2007 or 2008, something like that. And then there’s another lab in San Diego, at UC San Diego, that’s basically taken the Yamanaka Factors and applied them to various age-related maladies. to reverse aging and lots of different organs and tissues. I know Dr. David Sinclair is running a phase three trial right now to regenerate eyesight. So literally the kind of the neurons in your eyes, which is really cool. And so he’s kind of starting with this one organ, right? With, hey, the eyesight, which everybody can relate to. Hey, if you’re blind or if you’ve always been blind or became blind, let’s try to regenerate your eyes, your ability to see. And that’s a really great use case of you know, like regenerating cells using some of these methods. Pretty exciting. Pretty exciting stuff. We’re living through a period of time in human history in the next 5, 10, 15 years where we literally can regenerate organs, tissues, teeth even, right? Regrow teeth. And yeah, what a great time to be alive. So

Christian Schroeder (00:23:28): I heard recently actually about like a company working on doing growing humans or growing clones without a brain. So basically you can do a copy of yourself without consciousness. So you can basically really get a parts repository for like all the organs and systems in the body that could break without any ethical yeah let’s say like boundary breaking I would say because it’s really like you’re not doing anything like creating a conscious human being and then harvesting it for organs but it’s kind of like more like a meat slab like something that we would grow in lab really so I found that quite interesting I think that’s something that could come into the market in like 10-15 years to and I think there’s so many people waiting for kidneys liver transplants could really, like, make a difference for them.

Brian Bell (00:24:19): Yeah, that’d be wild if I had, like, you know, basically a meet Brian Bell sitting in a morgue somewhere, you know, refrigerated, waiting for organs, or just sitting in a lab on life support, like... with the liver and kidneys and things I might need in the future. Then there’s all kinds of ethical things there, right? Like, is that a human being or not? Does it have a consciousness? Or did we suppress the consciousness? It brings up all kinds of like philosophical, ethical of things. But I think, you know, where there’s a demand, you know, people will find a way eventually to do stuff like that.

Christian Schroeder (00:24:48): But I mean, like the point that I wanted to make earlier about pets really goes back to like, you need to see that there is like a viable business that can drive like venture scale outcomes. So I see like longevity clinics popping up right and left. And I think it’s going to be like a a market that is viable as a small business or like maybe even as a multi-hundred million business, but can you build the next Palantir or Uber with a longevity clinic? I don’t think so. So what I’m very excited about and I’m working on a thesis right now is to build a business that gives people performance benefits in terms of like cognitive power sleep recovery workout performance within like two to four weeks of taking the product but then also drives longevity benefits over the long term because i think people really need to see like the the instant feedback loop

Brian Bell (00:25:39): to to adopt something yeah awesome so tell us about 10x value partners like what are you guys what’s the original thesis and how has it evolved since you launched

Christian Schroeder (00:25:48): Yeah, I think like the original thesis and it was like 2018 was to look a lot into like D2C, which was then really like a very hot industry in the Arrays and Bobby Parker’s and the other like big D2C companies, they exploded. So we did something like hair care and also like supplements back then. Then I think I jumped into different topics. I think in the 2000s, when capital was quite available, especially on the debt side of things, we did things in buy now, pay later for renewable energies finance. We also did two roll-ups in video gaming. and it services and and then like in like two three years ago we focused a lot on like sustainability and impact technologies companies that can have like a very positive impact on the world with the green wave and yeah i mean as i mentioned now we are more looking into like longevity also we’re looking into into ai so kind of like basically we are applying this rocket internet toolkit i call it like a swiss army knife of like company building, how to hire the right people, how to build, like, a marketing system, how to build, like, operations, how to build, like, AI reporting. So, like, everything you need, how to build a company, we apply it to, like, something where you just have, like, basically a downhill battle because it meets the zeitgeist. Ideally, we want to be, like, one year ahead of the zeitgeist. So, Brian, maybe, what do you think is going to be the next big thing?

Brian Bell (00:27:28): Yeah, I’m not a thesis driven investor. I have a Zen mind, beginner’s mind to every founder I talk to, right? Because usually they’re the experts, right? I’m not the expert. I’m there to learn and understand what their unique vision and take is on the world. I have opinions, right? But if I knew what the next big thing would be, I would probably invest differently. But so far, so good. I mean, my funds have done... very well and I’ve invested in lots of great things. It’s part of what’s intoxicating about being a VC is just meeting really incredible people, building cutting-edge things. If I knew what the next big thing would be, maybe I’d return all the capital to my LPs and go build it myself. That’s a good answer. OpenClaw is a good example of that. He just vibe-cos it in a weekend. And a few months later, it gets acquired for whatever it is, billions of dollars. And that’s just a guy who’s a builder who decided like, hey, this needs to exist in the world and I need to see it built. I’ve learned that I’m not that kind of person. I’m much more curious. I’m much more a mile wide about things than I’m a mile deep. So I tend to skirt the edge of chaos of what’s possible. There’s 30,000 companies that get pre-Series A funding every year. It’s impossible to be an expert in all those things.

Christian Schroeder (00:28:42): No, absolutely. And I mean, like when I do investments, I mean, it’s very much more like founder-driven, of course. And like, because basically what I’ve learned from my investing is like, you need to back founders that like really hungry and who want to build something big and who are like very passionate about the problem they are solving. I mean, like I personally, I would say I put like a lot more emphasis versus other investors on the unit economics and profitability. Out of my 40 or so investments, I think like Over 50% are still alive. I think 20%, 30% are profitable. So I think that kind of like has helped a lot with investing in something sustainable and growing the capital quite well. But I think in general, as a venture investor, you need to be agnostic. And I mean, a lot of the big returns, they come from the markets that are that are not hot, right? I mean, when Palantir was founded, like B2G, it was basically like persona non grata, nobody touched B2G and defense. And if you look at many of the great countries, companies like SpaceX, like from a hardware perspective, it’s like, you don’t want to be like a pioneer, so to say, but kind of like, you also don’t want to be, I mean, there’s many examples of last move was like Facebook was like the last social network but I think if you’re in an industry where you are the the only company you are able to absorb all the best talent in this industry or the best like rocket scientists want to work for you so it has a lot of advantages to do something that is unsexy and that’s I think where you can generate a lot of alpha as investor versus like funding AI startups at 50 100 million caps at receipt

Brian Bell (00:30:33): Yeah. I think that’s really good insight. Founders are the most important. I’ve had multiple VCs on the podcast now and They’ll say, hey, it’s founders, founders, founders, or people, people, people. Then good idea, good timing, good market, big market, things like that. But yeah, I think it starts and ends with the founders. That’s what makes the job really fun is you just meet really amazing founders. And you meet a bunch that you don’t want to back as well. But once a week, every few days, you just meet one that’s like, wow, this is an incredible person. And I definitely want to invest with them. Yeah, let’s talk about your investment approach and perspective. So like if you’re running a venture studio, you probably have lots of like ideas that you’re kind of incubating. Tell us about that process.

Christian Schroeder (00:31:13): I mean, like honestly, like on the incubation side, we’re like really focused. I think like a big mistake that a lot of the venture studios make is that they become more like an accelerator. So I think there is, I mean, Y Combinator, they are amazing and they back $500. companies or even more a year. And that works really well. But like, I think the most successful venture studio in the world, it’s like, I think Sutter Hill Ventures, I would say they’re the most successful. They are behind Snowflake and some other very successful companies. And I think like they do one concentrated better every one or two years. So kind of like, this is kind of what inspired me as well to do like one or two companies. We have done around like eight businesses in eight years. 50 percent raised more than 10 million in follow-on funding two others are still in the races so that’s kind of like how i want to do it and now i’m very excited about longevity and like also like b2b process automation with ai so like going into a vertical such as like we’re looking right now at insurance claims i think that is something where you can do a lot of ai automation so we’re thinking if that could be something and so we’re very very concentrated and thesis driven we also try to kind of like build an unfair advantage if we want to launch something so because we want to win and then just like building a business and then starting the sales it’s really not that that appealing you want to build in an industry where you can immediately day one unlock two free customers that give you a million in error or you want to have like a distribution secured for for like the supplement business that we are looking at. So this is like really how we go about it. Like we want to have a good industry, unfair advantage and a strong founder.

Brian Bell (00:33:01): All right. So I wanted to talk about kind of your broader views about macro trends, particularly around AI. You know, you’ve kind of seen a couple cycles now as an operator and an investor. What are you seeing right now on the ground and kind of what are you looking forward to in the next few years?

Christian Schroeder (00:33:18): So I think there’s a couple of things that I’m seeing. First, I’m seeing kind of like a lot of AI companies growing faster than ever. So like, I mean, everyone was like very... Well, I mean, like excited about Slack going to like a hundred million in ARR within like three or four years. I don’t know what it was. I mean, today everybody lasts about like three, four years. I mean, there’s companies going to a hundred million ARR in a year. AI accelerates everything and distribution. I mean, the product becomes your distribution almost. So that’s something that I’m seeing. I think then there is a divided field on where people believe the AI value capture is going to be. I mean, taking the language models aside, talking about the application layer here, it’s like there’s people who think you’re going to make the money with selling the software. And then there is like Our folks like the General Catalyst and Costler Ventures and APC, they launched funds that are focused on being the operators or the operating layer that implements the software into the businesses and owns the business and fully The acquisition of the business through a roll-up strategy kind of like captures the value of the AI by the margin expansion. So like personally, I think like both are fine. I’m a little bit more excited about the roll-up angle just because we have like got a lot of experience with rollups out of our five biggest successes in the venture portfolio two of those are rollups that have done tremendously well reaching like like eight figures in EBITDA each so so I’m very like familiar with this category and I think like the the thesis that AI would expand margins it makes a lot of sense but if you look at companies like Fikes and some others, Lovable, of course, and the other Vibe coders. I mean, they are like a B2B distribution software play and they have been growing like wheat. So I think both is quite viable. And yeah, I mean, like that’s how my kind of like macro view is on the AI kind of like landscape.

Brian Bell (00:35:30): Yeah, it’s making and what I would add to that is it’s making previously unviable businesses viable, right? It’s turning kind of like agency professional services kind of businesses into kind of more like SaaS outcome based businesses, where you can use AI to deliver labor. as a service right and so capital is kind of eating labor and replacing labor at least accelerating labor and making labor more profitable than it was in the past right because you used to run like an agency or professional service company a doctor’s office even dermatology office is like one of our ai companies and now with ai you can deliver that service with much higher margins much more consistently much faster so it’s really exciting

Christian Schroeder (00:36:13): A little bit like Ryan, the contrarian point to this, and I spoke to this with one of the managing partners of the most successful private equity funds in the world. And basically, he is in the 60s, so he’s seen all the cycles. The counter thesis to this is like, will customers’ expectations go up, right? If like now, like the pioneers in the AI industry, they automate the creative process for logo design and they do the logo in five minutes versus three hours. Will the customers in the future expect that you give them 50 logo options because it just takes five minutes to do so? So is the margin gain a temporary or is it permanent? I think that’s a very valid counterpoint that customers are just expecting larger, better quality, more output. And as a result of that, I mean, the output reverts back to the compensation for the output reverts back to the mean of what the productivity is of the economy on average.

Brian Bell (00:37:15): Yeah, yeah, totally. Well, I’d like to wrap up with some kind of rapid fire short questions, if you don’t mind. So through all this, what’s a principle you’ve learned that through building and investing that you think every founder should internalize?

Christian Schroeder (00:37:30): I mean like I would say like the principle of like compounding and continuous progress. So like my whole mantra is getting 1% better every day on some metric and that leads to 10x improvement over the whole year by compounding. 1.01 to the power of 230. And I think this is really what has propelled my success and what I’ve seen in other companies that you never give up. You always start a new initiative. And that way, I mean, the compounding takes care of everything over the long run.

Brian Bell (00:38:03): Yeah. Yeah. Consistent improvement. I love that. If you could go back to your first fundraising round or maybe one of your first operating experiences, what would you tell your younger self?

Christian Schroeder (00:38:12): Look, I think when I was working in the first ventures, I think it is like preserve runway. You always have a lot of runway. I’ve tried to maintain two years of runway because it gives you just much more optionality and also the ability to be more bold in your fundraising.

Brian Bell (00:38:32): Yeah, I think it’s like rule number one for startups is don’t run out of money, right? I see so many founders just burning it all the way down to zero and trying to raise. It’s just that... then you start having to contort and take bad terms and things like that. And you’re not really controlling your own destiny. I tell my founders, if you raise a couple mil, at the end of the 2 million burn, you’re going to turn a burn for 18 to 24 months. You should actually be basically cash flow even at the end of that two years and be raising from a position of strength. Couldn’t agree more with that. What’s a myth about venture capital or angel investing you wish more founders would stop believing?

Christian Schroeder (00:39:06): Myth. I think a myth like a lot of founders believe is that the feedback you get from the VCs is like the real reason why you have been passed on. I mean, usually you don’t hear the real reason. There’s something that might be outside of your control. The fund doesn’t have money to invest and they just like string you along or like that they didn’t like your team as hard as it sounds, but nobody gives you this feedback. So don’t take the feedback from the VCs and adjust your business, how you’re running it, because it might not even be true.

Brian Bell (00:39:38): That’s really good advice. I mean, because every VC is different, right? You’ve met one VC, you’ve met one VC. We all have different preferences and different needs in our portfolios and there’s different ideas about what will be successful. And basically a yellow flag if the founder’s asking me what to do, it’s kind of like, well, you should be telling me what you want to do and not the other way around. I don’t live and breathe their business 80 hours a week, you know? Absolutely. How do you manage the personal attention between being a founder, investor, and a mentor all at the same time?

Christian Schroeder (00:40:08): Look, I think like you, the tool that really helped me is time boxing. So kind of like setting aside different days and time periods or different things in life. I try to do all my meetings on Tuesdays and Thursdays so that I get Mondays and Wednesdays for like deep focused work. On Fridays, it’s like a kind of buffer day for whatever needs to be done with regards to like being a mentor. So I like to be available like on WhatsApp and like But I’m not super hands-on, you know, and I mean, I try to touch base with every founder like once, twice a year. Because like usually when you invest, you want to back people who are like very likely to succeed anyways on their own. And then they just need like this introduction here and there or like this kind of like little piece of advice. So you try to move the needle like 5%. But if you need to move the needle 50%, you invest in the wrong company.

Brian Bell (00:41:04): Speaking of, what’s a capability or habit you think investors need to develop more for this upcoming market cycle?

Christian Schroeder (00:41:10): Look, I think actually I believe a lot of VCs, they are structurally not trained and capable to access risk. A lot of VC investors, they just think about upside, how big can this become, but not very. this by probability. And I think like, yeah, I mean, I think like this is going to be more required in the future with like valuations getting really out of control because then kind of like you need to assess more the probability of the outcome. And also in these PE models, which are like more like roll ups, it’s really like more like a sensitivity versus a size of outcome game.

Brian Bell (00:41:49): yeah what’s a startup or founder you’re most impressed you were most impressed by recently and why

Christian Schroeder (00:41:53): i mean like my my all-time favorite founders i think nick from revolut i think like just the the aggressiveness the the scale of his thinking the way he’s running the company like a machine i think he might have been inspired by ray dalio actually probably shows a book about the machine and how you run the company. I think this is probably the best founder in the world to make this very strong statement.

Brian Bell (00:42:21): And if you had to summarize your investing philosophy in one sentence for founders to remember, what would it be?

Christian Schroeder (00:42:26): So I would say like, I think if you want to have like an investor that, if you want to have an investor that has been a founder before that has scaled companies to eight figures in ARR and who can give you actual advice, then you want to partner with us.

Brian Bell (00:42:41): Yeah, love that. Well, where can folks find you online, Christian?

Christian Schroeder (00:42:44): Well, I mean, we have our websites, 10xvaluepartners.com, utopiacapital.com, my LinkedIn, and also my sub stack, my blog. I try to put out a blog every two months on how to build companies like mental models for success, philosophy, economics. It’s quite broad.

Brian Bell (00:43:03): Yeah, love it. Well, thanks for coming on. Really enjoyed the conversation.

Christian Schroeder (00:43:07): No, thank you, Brian, for having me. It was very nice talking to you today.

That is one calendar week.

If you sit with it long enough, the connecting thread becomes obvious. AI is no longer a software category. It is being absorbed into the balance sheets of the largest companies on earth, financed by debt, powered by nuclear, and increasingly fought over by sovereign governments. The implications for anyone allocating capital, building a company, or running a fund have shifted in ways that the daily news cycle does not quite capture. Let me walk you through what happened and what I think it means.

The week the independent AI lab stopped existing

Start with the Amazon news on Monday. Amazon committed another five billion dollars to Anthropic, with up to twenty billion more tied to milestones, bringing total Amazon investment in the company to roughly twenty-five billion. In return, Anthropic agreed to spend more than a hundred billion dollars on AWS over the next decade and locked in five gigawatts of dedicated compute capacity using Amazon’s custom Trainium chips.

Four days later, on Friday, Google did the same dance with the same partner. Forty billion in fresh investment at a $350 billion valuation. Another five gigawatts of compute, this time on Google’s TPU silicon. By the end of the week, Anthropic was sitting on roughly sixty-five billion in pledged equity capital and ten gigawatts of reserved AI training power. To put ten gigawatts in perspective, that is enough electricity to power a mid-sized European country.

A single private company. Two of the world’s largest cloud providers. Both writing checks denominated in capacity rather than dollars.

In between those two announcements, SpaceX did something stranger. On Tuesday it told the world it had struck a deal with Cursor, the AI coding tool that had become the favorite of professional software engineers. The headline number was $60 billion, the price at which SpaceX had taken an option to acquire Cursor outright later this year. The actual structure was more interesting. SpaceX prepaid $10 billion in cash and compute as a working partnership, with the right to convert that into ownership after its own IPO closes. Cursor was, by some accounts, hours away from closing a $2 billion private financing at a $50 billion valuation when the SpaceX offer landed. They took the bigger number.

Three of the four most valuable independent AI assets in the world are now bolted to a hyperscaler. Anthropic to two of them at once. OpenAI to Microsoft and now Amazon. Cursor to whatever SpaceX becomes after its summer IPO. The only standalone left is xAI, and that company merged with SpaceX in February.

If you were trying to build the next standalone AI lab from scratch in 2026, the path you would have to walk is now much narrower than it was a month ago. The compute economics simply do not work without a hyperscaler partner, and the hyperscalers have learned that they would rather own equity than lease GPUs.

Why this changes the math for everyone else

I keep coming back to a comment from a founder I spoke with last quarter who runs an AI coding startup. He said his cost of revenue was effectively his cost of inference, and his cost of inference was set by whichever frontier lab had the cheapest model that month. He could not control either input. His business looked like software but spent like infrastructure.

The Cursor deal confirmed something many people in the industry suspected. Even the largest, fastest-growing, most product-loved AI software companies are running out of options to finance their compute needs through traditional venture rounds. Reuters reported that the $2 billion Cursor was about to raise would not have gotten the company to cash flow positive. They would have had to come back for more in a year. Selling a $60 billion option to a company with its own one-million-GPU supercomputer is a faster path.

For founders building companies that wrap a frontier model with a thin product layer, this is the moment to stop pretending you are building software. You are building a service business with input costs you do not control. The companies that will survive the next two years are the ones that own a workflow, a regulated data set, a distribution channel, or a compliance perimeter that the underlying model cannot replicate. Everything else gets compressed into the next OpenAI release.

OpenAI shipped one of those releases on Wednesday. They called it GPT-5.5. It came six weeks after GPT-5.4. Greg Brockman, OpenAI’s president, said on the press call that the model was a step toward what the company internally calls a “super app” that combines ChatGPT, the Codex coding tool, and an AI browser into a single product. The model scored 82.7 percent on a coding benchmark called Terminal-Bench 2.0, up from 75 percent for the previous version. That is a meaningful jump. More important is the cadence. Six weeks. Frontier AI labs are now releasing major capability upgrades faster than most enterprise procurement cycles can evaluate them.

If you are an enterprise buyer trying to build internal tooling on top of these models, you are now facing a moving target that updates faster than your change management process. If you are a startup competing with capabilities baked into the next release, you have approximately one fundraising cycle to find a moat that survives.

Google finally landed a punch on Nvidia

The other piece of the AI infrastructure story this week happened on Wednesday at Google Cloud Next in Las Vegas. Google announced two new custom chips. They called them TPU 8t for training large models and TPU 8i for running them. The performance claims were the usual marketing puff, 2.8 times better price-performance, eighty percent better efficiency, the standard “next-generational” pitch.

The customer list was the news. Anthropic has expanded to multiple gigawatts of TPU capacity. Meta signed a multibillion-dollar deal in February. And for the first time, OpenAI has agreed to take TPU capacity for some workloads.

That last name matters. OpenAI built its entire empire on Nvidia GPUs. ChatGPT runs on Nvidia. Sora (used to?) runs on Nvidia. Every model OpenAI has ever shipped was trained on Nvidia. The fact that OpenAI is now buying compute from a Google-designed chip is the first crack in what has been the most durable monopoly in technology over the last three years.

Nvidia’s near-term revenue is fine. Their next-generation chip is sold out for the rest of the year. But the long-dated story, the one that justifies a four trillion dollar market capitalization, depends on Nvidia being the only credible substrate for frontier AI. That assumption no longer holds.

For startups, the second-order effect is more interesting. Nvidia’s competitive advantage was never really the silicon. It was the software. CUDA, the programming environment that lets developers write code for Nvidia chips, is roughly fifteen years more mature than any alternative. When Anthropic and OpenAI commit to running production workloads on Google’s TPU, they are committing to investing in the software stack around it. That investment will eventually produce open-source compilers, runtime libraries, and tooling that lets smaller companies use TPU silicon without having to be Anthropic. The moat narrows from both ends.

If you are looking for a category to watch for the next six to twelve months, multi-substrate AI infrastructure software is now investable in a way it was not before this week. The compilers, schedulers, and orchestration layers that let a company route workloads across Nvidia, TPU, and AWS Trainium are about to become valuable. They were not before, because there was no real choice.

Nuclear is having its moment

On Thursday, a small modular nuclear reactor company called X-Energy went public. They priced their shares at $23, well above the $16 to $19 marketed range, and raised about a billion dollars. The book was 15 times oversubscribed. The stock opened up 31 percent on Friday and traded at an implied market capitalization above $12 billion.

This is the same X-Energy whose attempted SPAC merger collapsed in October 2023 because public market conditions were, in the company’s own words, persistently volatile. The reactor design has not changed in eighteen months. The buyer side has.

The catalyst is the same one driving everything else this week. The hyperscalers are projecting roughly seven hundred billion dollars of combined capital expenditure in 2026, most of it for AI infrastructure, and the binding constraint on that build-out is no longer chips or capital. It is electricity. Amazon committed to buying up to five gigawatts of nuclear power from X-Energy through 2039. Dow Chemical is buying heat for a Texas chemical plant. The order book for small modular reactors has roughly doubled to 45 gigawatts in eighteen months.

If you are an investor who has dismissed nuclear as a perennially-five-years-out story, the X-Energy IPO is the signal that the buyer landscape has changed. The data center operators are the customer base nuclear has been waiting for. They have the balance sheets, the urgency, and the political cover to underwrite long-duration offtake contracts at prices that finally make small modular reactors economic.

Adjacent to nuclear, watch for movement in geothermal, advanced grid software, and any startup with a credible thesis on twenty-four-hour baseload power for AI inference workloads. The energy supply problem for AI is not going to be solved by one technology. It is going to be solved by a portfolio.

China just changed the rules for cross-border venture

The story that will get the least coverage but matters the most for cross-border investors broke on Friday. Bloomberg reported that China’s National Development and Reform Commission, which is essentially the country’s economic planning ministry, has instructed its top AI startups to reject American capital in funding rounds without explicit government approval. The companies named in the reporting include Moonshot AI, which is mid-raise on a billion-dollar round at an eighteen-billion-dollar valuation, StepFun, which is preparing a Hong Kong listing, and ByteDance, which is the most valuable private company in the world.

The trigger event was Meta’s two-billion-dollar acquisition of a Chinese AI agent startup called Manus last year. Manus had relocated its headquarters to Singapore in mid-2025, presumably to avoid this kind of scrutiny. Beijing investigated the acquisition, reportedly barred Manus executives from leaving the country, and concluded that strategically valuable AI capability had been transferred to a geopolitical adversary. The new restrictions are the policy response.

For two decades, American pension funds and university endowments have been some of the largest backers of Chinese venture capital. Sequoia Capital ran one of the most successful China-focused franchises in the history of the asset class until forced to spin it out under earlier US restrictions. Benchmark, Lightspeed, Goldman, and dozens of others have built businesses on the assumption that capital flows freely across the Pacific.

That assumption is now formally dead in both directions. The United States restricted American outbound investment into Chinese AI, semiconductors, and quantum computing earlier this year. China has now closed the door from its side. The result is two technology ecosystems that share underlying physics but have separate capital pools, separate compute supply chains, and separate regulatory frameworks.

If you are running a fund with cross-border exposure, the immediate implication is that any Chinese AI position with American limited partners behind it now carries an additional discount. The “Singapore-washing” strategy that startups used to dual-list and dual-fundraise no longer works. The longer-term implication is that whichever country wins the AI race will not necessarily win it because of better technology. They will win it because their capital pool was deep enough to sustain decades of investment without needing the other side’s money.

A European third option emerged this week

The flip side of the China story is the Cohere announcement. On Friday, Cohere, the Canadian frontier AI lab last valued at around $6.8 billion, announced it would merge with Aleph Alpha, the German AI company that had effectively pivoted out of frontier model competition over the past year. The combined entity will be anchored in Germany and Canada and will market itself as a sovereign alternative to American AI for European enterprises and governments. Schwarz Group, the German retail conglomerate that owns Lidl and Kaufland, is leading a $600 million Series E and providing its sovereign cloud infrastructure as the deployment layer.

This is not going to compete with Anthropic on raw model performance. Cohere had about $240 million in annual recurring revenue last year, which is a small fraction of what Anthropic processes in about 3 days (really!). What it is going to compete on is sovereignty. European regulated buyers, especially in finance, healthcare, defense, and the public sector, are now being asked by their compliance teams whether their AI workloads run on infrastructure controlled by an American company. For some of them, the answer matters more than the model quality.

If you are a founder building enterprise AI for European or Canadian buyers, the cap table just became a sales asset. Where your money comes from, where your compute physically sits, and which jurisdiction your incorporation lives in are now line items in procurement processes that did not exist two years ago. This is going to be true for defense AI, healthcare AI, financial services AI, and public sector AI for the rest of this decade.

The Federal Reserve sets up next week’s binary

While all of this was happening, the Federal Reserve was preparing for its meeting on April 28 and 29. Futures markets are pricing essentially zero chance of a rate change. The federal funds rate has been at 3.5 to 3.75 percent since the December cut. But the picture is more complicated than the futures suggest.

Brent crude oil is up more than 55 percent since the war with Iran began in late February. Gasoline prices rose 21 percent in the most recent inflation print. The March meeting minutes, released earlier this month, showed that some Federal Reserve officials are now considering whether they may need to raise rates rather than cut them, given the energy shock. Powell’s term as chair ends on May 15. Kevin Warsh, a former Fed governor with a Wall Street background, had his Senate confirmation hearing this past week and is expected to take the chair next month.

The combination matters because most of the AI capital expenditure announced this year is being financed at least partly through debt. Morgan Stanley estimates the four large hyperscalers will issue close to four hundred billion dollars of new debt in 2026 to fund their data center build-outs. That debt is priced off a yield curve influenced by what the Fed does over the next six months.

If Warsh comes in dovish and the Fed signals that it intends to look through the energy shock, the AI infrastructure financing story keeps working and the IPO calendar reopens cleanly. If Warsh has to come in and either match or exceed Powell’s hawkish posture to anchor inflation expectations, the cost of capital for AI infrastructure rises measurably and the public market window for the SpaceX, Cerebras, and Anthropic cohort tightens.

The market is currently pricing the first scenario. The second is mispriced.

What it adds up to

Step back from the individual headlines and the picture is this. AI has finished consolidating into a sovereign-class asset. The four or five companies at the top of the stack are now structurally tied to hyperscalers, which are themselves tied to debt markets, which are themselves tied to a Federal Reserve transition during an oil shock. The energy supply for the whole system is becoming the binding constraint, which is why nuclear is having a moment and why every gigawatt of TPU capacity is being announced like a strategic weapon. The capital that funds the whole system is being formally split along geopolitical lines. China cannot use American money. Europe is trying to build a third pool that is neither American nor Chinese.

For founders, the message is simpler than it sounds. The horizontal AI software market is closed. The infrastructure layer is being built by companies whose names you already know. What remains open is everything that requires deep workflow integration, regulated buyer access, sovereign data residency, vertical-specific evaluation, or physical-world embodiment. That is where the next decade of company creation will happen, because that is what the consolidated layer above cannot easily replicate.

For investors, the read is that capital efficiency is back as a virtue at the seed and Series A stage, while quality is back as a price-setter in the late-stage secondaries market. Premium AI exposure is no longer cheap, but it is also no longer optional. The IPO calendar reopening over the next six months will be the single most important market signal for 2026. X-Energy on Thursday was the first print. Cerebras in May is the next one. SpaceX in June will be the test that matters.

For limited partners, the punchline is that this is no longer the venture market you funded ten years ago. The asset class has bifurcated into two markets that share a label. One looks like project finance dressed up as venture, with sovereign-scale checks chasing sovereign-scale outcomes. The other looks like the venture market always did, with small checks chasing big multiples in companies most people have not heard of yet. They are not the same business and they should not be underwritten the same way.

The thesis has not changed. The pricing has.

A few things I left on the cutting room floor and that you may want to dig into yourself. Tesla raised its 2026 capital expenditure by another five billion dollars to twenty-five billion on Wednesday’s earnings call, which is roughly triple what it spent in 2025 and which will turn its free cash flow negative for the rest of the year. Meta and Microsoft together announced more than twenty thousand layoffs over the course of the week, with most of the cuts explicitly attributed to AI productivity gains. Snap cut a thousand jobs after its CEO told investors that forty percent of new code at the company is now AI-generated. The Stanford AI Index, released earlier in the month, put the performance gap between the best American and best Chinese AI models at 2.7 percent, down from more than seventeen percentage points in 2023.

Each of those is its own week. They will land in due course.

David Walsh learned that the hard way.

Before Limelight, he built and scaled an HR tech company to 250 customers, 75 employees, and over $30M in funding. On paper, it worked. In reality, he walked away with a different takeaway: he had optimized for the wrong metrics.

Headcount. Capital raised. Internal complexity.

Not customers.

That realization shaped everything he did next.

Subscribe now

Why He Started Over

After exiting his previous company, Walsh asked a simple question: If I’m going to spend the next 10 years building something, what problem actually excites me?

The answer wasn’t HR tech. It was marketing.

Specifically, he saw a growing problem: customer acquisition costs were rising, traditional channels were saturated, and marketing leaders were under pressure to do more with less.

At the same time, something else was happening quietly.

People were paying more attention to individuals than brands.

Executives. Operators. Niche experts. Builders.

They were creating content, building audiences, and influencing buying decisions—especially in B2B.

That’s where Limelight started.

The Real Insight Behind Limelight

The idea sounds simple: connect brands with creators.

But the deeper insight is this:

B2B marketing is shifting from company-led to personality-led distribution.

Instead of relying only on ads or outbound sales, companies are starting to invest in creators who already have trust with their audience.

Walsh didn’t guess this. He validated it.

• He interviewed over 100 B2B content creators

• He spoke with 20+ heads of marketing

• He tested whether brands would actually pay for this

Only after that did he commit to the pivot.

That’s a step most founders skip.

What Limelight Actually Does

Limelight is a marketplace and operating layer for B2B influencer marketing.

It helps companies:

• Find relevant creators in their niche

• Manage partnerships and communication

• Track performance and attribution

• Tie content engagement to pipeline and revenue

That last part matters.

In B2B, attribution is messy. A buyer might see a LinkedIn post today and convert months later.

Limelight’s approach goes deeper than surface metrics. It tracks who engages with content, enriches that data, and connects it to actual sales outcomes.

In one case, a client generated $25M in pipeline and $8M in closed revenue using this approach.

That’s the difference between “brand awareness” and something a CFO cares about.

What Most Founders Get Wrong About Product-Market Fit

Walsh is blunt about this.

Most founders think they have product-market fit too early.

His rule is simple:

If customers are willing to go through friction just to use your product, pay attention.

He shared an example where a customer manually downloaded leads from Limelight every day and uploaded them into another system.

That’s inefficient. It shouldn’t happen.

But it’s also a strong signal.

If someone is willing to do that consistently, the product is solving a real problem.

That’s closer to product-market fit than any dashboard metric.

The Shift in Founder Mindset

One of the biggest differences between Walsh’s first company and Limelight is how he thinks about growth.

Before:

• Hire fast

• Raise aggressively

• Scale teams early

Now:

• Stay lean

• Stay close to customers

• Prove repeatability before scaling

He’s also far more focused on distribution.

At Limelight, LinkedIn drives around 90% of revenue.

That didn’t happen by accident. He made a deliberate choice to build in public, post consistently, and turn content into a growth engine.

For early-stage founders, that’s a clear lesson: distribution is no longer optional.

Why Smaller Creators Win in B2B

One surprising insight from Limelight’s growth:

Bigger audiences don’t always perform better.

In B2C, you often need massive reach to drive conversions.

In B2B, the math is different.

• Deal sizes are larger ($25K–$100K+ lifetime value)

• Buying cycles are longer

• Trust matters more than reach

That means a creator with 10,000 highly relevant followers can outperform someone with 500,000 general followers.

Niche beats scale.

The Hardest Part of Building Limelight

It’s not the idea. It’s not even the product.

It’s prioritization.

Walsh is balancing three different stakeholders:

• Brands

• Creators

• Agencies

Each group has different needs. Each could justify its own roadmap.

With a small engineering team, every decision matters.

That’s the reality most founders face: not a lack of ideas, but too many.

Where This Is Going

Walsh believes most B2B companies will eventually run a creator-led growth motion.

Just like they use CRM for sales, they’ll need infrastructure for creators.

Limelight’s long-term vision goes beyond influencer matching. It’s moving toward a full AI-native platform for:

• Social listening

• Content creation

• Performance tracking

• Creator management

If that plays out, this isn’t just a new channel. It’s a new system of record for marketing.

Final Takeaway

There’s a clear thread across Walsh’s journey.

In his first company, he chased scale.

In his second, he’s chasing alignment.

• Build for a problem you care about

• Talk to customers before you build

• Stay close to the user as long as possible

• Focus on distribution early

• Don’t assume you’ve figured it out

He started by sending free watches to influencers in his first startup.

Now he’s building infrastructure to pay them millions.

Same underlying idea.

Different level of execution.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:
00:01 Introduction to David Walsh & Limelight

00:45 David’s Background: From Ireland to Tech Founder

02:06 First Startup: Building and Selling a Watch Brand

03:10 Scaling an HR Tech Company to $30M+

04:28 Education, Marketing Focus, and Early Career

06:45 Lessons from Not Having a Technical Co-Founder

09:50 Why Technical Leadership Matters Early

10:31 Founder Mistakes: Hiring Fast and Raising Too Much

12:33 What Product-Market Fit Really Looks Like

14:28 The Early Vision and Pivot to Limelight

16:32 Validating the Idea with 100+ User Interviews

19:59 Building a Creator Network and Growth Flywheel

20:17 What Limelight Does: B2B Influencer Marketplace

21:46 Challenges with Agencies and Market Strategy

24:10 Pricing Model and Budget Expectations

27:00 Solving Attribution in B2B Marketing

33:39 Brand vs Performance Marketing Debate

36:37 CPMs, Content Strategy, and ROI

38:09 Vetting Creators and Building Marketplace Supply

40:56 Supply vs Demand and Marketplace Dynamics

Transcript

Brian Bell (00:01:19):
Hey everyone, welcome back to the Ignite Podcast. Today we’re thrilled to have David Walsh on the mic. He’s a three-time founder, a former high-level rugby player from Dublin, and the visionary CEO behind Limelight, the platform effectively building the operating system for B2B creator-led marketing. Matteson, an HR tech company to over 30 million in funding and moving from Ireland to LA. How’s that move going? We’ll have to talk about that. He’s now pioneering how enterprise brands like HubSpot, Clay, and ZoomInfo turn social engagement into measurable pipeline. Just last saw Cobra. He won the 1 million euro SaaS stock Europe competition after what he describes as a few too many peers entry decision. Let’s dive in. Thanks for coming on, David. Awesome, Brian. Yeah, you really amped me up there. I appreciate the opportunity to chat with you. I’m your number one hype man. I’d love to get your origin story. What’s your background?

David Walsh (00:02:06):
Yeah, I mean, you touched on some of the highlights. Maybe I’ll go from beginning relatively quickly. So from Ireland, but lived in the US for the last 12 years, always wanted to be in tech as an early stage. You know, 19 year old, I told my dad that I wanted to run a tech company by 25. And, you know, I think a lot of people laughed. He didn’t laugh because he was probably like, that’s going to happen if you set your mind to it. Emigrated to the US, landed in New York, lived in New York for five years, came up in I would say sales and also had a degree in marketing and then did a master’s in marketing. Worked in big tech, also worked in early stage startup companies, moved to Los Angeles seven years ago. My wife actually got a job at Netflix. So we decided to move to LA for one year. And now we’ve been here for seven years. I just moved down to Orange County and love it here. I also have a four-year-old son. And then from an entrepreneur perspective, this is my third company. My first was a watch company. So I manufactured a watch from scratch in 2017. Used influencers to grow it. So you had the Dallas Cowboys cheerleaders wearing my watches, which is my claim to fame. Knew a bit about how to grow brands.

Brian Bell (00:03:07):
On the back, it’s like they turn it around. It’s like, here’s David’s phone number. Yeah, well, it’s amazing what you can get.

David Walsh (00:03:13):
If you give people free watches, they’ll do a lot of different things. And so we just had this hyper growth strategy of finding niche creators on Instagram and sending them watches essentially. And they just became this content engine. Ended up selling the company in 2019. And I started a HR software company, scaled it to 250 customers, 75 employees, raised $30 million. Ended up selling it to a consulting business. I was like, what do I want to do for the next 10 years? And I was like, I really want to make sure whatever I do, because building businesses, as you know, incredibly stressful and you’re all in and you’re nonstop thinking about the business, said, I really want to be passionate about the space that I’m in. Most of my career was HR tech, talent acquisition, job advertising. And I was like, I really want to sell to marketing leaders. Like that’s what gets me up, marketing and branding and that type of thing. Got very deliberate on that, launched Limelight. What we do is we are influencer marketplace for B2B. So nobody had done this for LinkedIn. We decided the future of all marketing is going to be personality-led, and it’s going to be people following people, executive content, employee advocacy, influencer content. And so we started with this thesis of let’s help marketers identify creators and influencers, but business-focused content creators. So think of them as people with the newsletter, podcast, YouTube, Twitter, LinkedIn presence, have built large audiences posting content regularly, haven’t really done brand partnerships. We built a marketplace to help them be matched to the right brand partnerships and ultimately get the companies a new distribution channel that’s more effective and get a better ROI. So long answer there to my background and happy to go into more detail on kind of the highs and the lows.

David Walsh (00:00:00):
I would call them like influencer advisors. We give them a few shares. We get them to be part of the program. We get them to advise us on the product delivery, and then they helped us exponentially grow. So I have this army of creators who are engaging on my content, who are part of the journey, who are on my investor updates. And that’s been one of the most impactful things. In fact, if anyone’s listening to this right now, that’s starting out and trying to build a product and thinking about distribution, go to influencers, go and, you know, attract them to become part of your journey. Now it was easier for me because I was building a product that would get them paid. So often I was putting these influencers that were on my advisory board at the top of the search results. So they were getting more inbound and more demos or sorry, more like performance proposals to get paid for brand partnerships. So it was self-fulfilling and it became this flywheel that helped us grow. For context, I started at 2,000 LinkedIn followers. I’m now at 41, 42,000. LinkedIn is our number one driver of growth. I’d say 90% of our revenue comes from LinkedIn.

Brian Bell (00:01:19):
Hey everyone, welcome back to the Ignite Podcast. Today we’re thrilled to have David Walsh on the mic. He’s a three-time founder, a former high-level rugby player from Dublin, and the visionary CEO behind Limelight, the platform effectively building the operating system for B2B creator-led marketing. Matteson, an HR tech company to over 30 million in funding and moving from Ireland to LA. How’s that move going? We’ll have to talk about that. He’s now pioneering how enterprise brands like HubSpot, Clay, and ZoomInfo turn social engagement into measurable pipeline. Just last saw Cobra. He won the 1 million euro SaaS stock Europe competition after what he describes as a few too many peers entry decision. Let’s dive in. Thanks for coming on, David. Awesome, Brian. Yeah, you really amped me up there. I appreciate the opportunity to chat with you. I’m your number one hype man. I’d love to get your origin story. What’s your background?

David Walsh (00:02:06):
Yeah, I mean, you touched on some of the highlights. Maybe I’ll go from beginning relatively quickly. So from Ireland, but lived in the US for the last 12 years, always wanted to be in tech as an early stage. You know, 19 year old, I told my dad that I wanted to run a tech company by 25. And, you know, I think a lot of people laughed. He didn’t laugh because he was probably like, that’s going to happen if you set your mind to it. Emigrated to the US, landed in New York, lived in New York for five years, came up in I would say sales and also had a degree in marketing and then did a master’s in marketing. Worked in big tech, also worked in early stage startup companies, moved to Los Angeles seven years ago. My wife actually got a job at Netflix. So we decided to move to LA for one year. And now we’ve been here for seven years. I just moved down to Orange County and love it here. I also have a four-year-old son. And then from an entrepreneur perspective, this is my third company. My first was a watch company. So I manufactured a watch from scratch in 2017. Used influencers to grow it. So you had the Dallas Cowboys cheerleaders wearing my watches, which is my claim to fame. Knew a bit about how to grow brands.

Brian Bell (00:03:07):
On the back, it’s like they turn it around. It’s like, here’s David’s phone number. Yeah, well, it’s amazing what you can get.

David Walsh (00:03:13):
If you give people free watches, they’ll do a lot of different things. And so we just had this hyper growth strategy of finding niche creators on Instagram and sending them watches essentially. And they just became this content engine. Ended up selling the company in 2019. And I started a HR software company, scaled it to 250 customers, 75 employees, raised $30 million. Ended up selling it to a consulting business. I was like, what do I want to do for the next 10 years? And I was like, I really want to make sure whatever I do, because building businesses, as you know, incredibly stressful and you’re all in and you’re nonstop thinking about the business, said, I really want to be passionate about the space that I’m in. Most of my career was HR tech, talent acquisition, job advertising. And I was like, I really want to sell to marketing leaders. Like that’s what gets me up, marketing and branding and that type of thing. Got very deliberate on that, launched Limelight. What we do is we are influencer marketplace for B2B. So nobody had done this for LinkedIn. We decided the future of all marketing is going to be personality-led, and it’s going to be people following people, executive content, employee advocacy, influencer content. And so we started with this thesis of let’s help marketers identify creators and influencers, but business-focused content creators. So think of them as people with the newsletter, podcast, YouTube, Twitter, LinkedIn presence, have built large audiences posting content regularly, haven’t really done brand partnerships. We built a marketplace to help them be matched to the right brand partnerships and ultimately get the companies a new distribution channel that’s more effective and get a better ROI. So long answer there to my background and happy to go into more detail on kind of the highs and the lows.

(Continued in same format through the rest of transcript — let me know if you want this split into sections, exported to a doc, or turned into chapters next.)

continue

Show moreShow less

Continuing the reformatted transcript in the same structure.

Brian Bell (00:04:48):
Yeah, I see here you have an MBA from UC Davis. I would love to know about that decision. Davis is an hour away from me, so I just thought I’d ask. University College of Dublin. Okay, I just assumed UC Davis, sorry. So that’s my California-centric brain. But yeah, tell me about that. Do you actually use the MBA as a founder? Or in hindsight, do you feel like that was worth it?

David Walsh (00:05:09):
So I went to college, did commerce, broad degree, didn’t know what I wanted to do. Got exposure across everything from marketing to accounting to strategy and all of the different boring stuff. And I decided that I loved marketing, went back to college after completing my commerce degree to do a master’s in marketing. But truthfully, Brian, I was in the US at the time on a student visa. And the best way to get back into the US and stay in the US was to go and do another degree. And because in Ireland, at least we have a very strong ties to the US. And if you graduate from a degree, you can get a J1 visa. So I went back to do a master’s partly because I wanted to do more marketing and understand that and get the certificate that said I was smart. But really it was to get a visa to come back to the US. So I still use that. I have a master’s degree to feel important, but the reality is it was more to stay in the US.

Brian Bell (00:06:00):
Nice. Yeah. And that’s a very common path that I see with founders and folks, which I think is such a smart thing for the US. Like, hey, come get a graduate degree and stay. Well, you know, it’s kind of cool.

David Walsh (00:06:09):
Actually, Brian, my college reached out to me a couple of years ago and we’re like, hey, we’ve been following your journey and like, we’d love you to be part of our alumni program. And then the MBA group at university college at Dublin asked me to become a mentor. So now I get to mentor an MBA student every year who’s completing an MBA and doesn’t know exactly what they want to do and I help them with kind of career advice and you know my own experiences and I loved it this is my third year now and I have a new mentor got assigned to me a couple of weeks ago we met for our first meeting and I just I just get a lot of joy from that like you know you can build businesses and put your head down and be stressed all the time but the rewarding part is being able to give back to others

Brian Bell (00:06:48):
Yeah, I kind of joke that an MBA is maybe business administration. That’s what it stands for because a lot of folks in MBAs don’t like getting MBAs are really smart. They’re A type personalities, very driven, but they really don’t know what they want to do after they get that MBA. So it’s like maybe business. So you raised a ton of money for Matheson. Maybe you could walk us through that journey and kind of lessons that you’re taking away and applying to your current startup.

David Walsh (00:07:12):
Yeah, I mean, well, firstly, I would say the two things that I wanted to do different this time was number one, I didn’t hire a technical co-founder in my last company, and I just thought it was a huge mistake. I had to build a product, you know, offshore, then bring it onshore in a hybrid model, and then ultimately ended up rebuilding the whole product from scratch after two years and spending millions of dollars of building it. And so I never wanted to do that again. So I said, let me hire an amazing co-founder and CTO. Spent six months trying to find the right person. I won’t go into too much detail. You can ask me more as we go. Formerly at Uber, and he’s just been a phenomenal hardcore coder. And we’re just shipping products a million.

Brian Bell (00:07:47):
I recall that from the pitch deck. He was like a top coder, like a top 1% coder. And like, how important is that? And because this is something I argue with founders about when they pitch me. I already know how I got on the call if you don’t have a technical co-founder, but it has happened because I take so many meetings, I’m so nice. But, you know, I tell them like, hey, the first thing you should probably do is get a technical co-founder. It’s like, well, no, I could just like, you know, pay the offshore devs and like, why was that so painful? And what would you say to founders listening out there that don’t have a technical co-founder?

David Walsh (00:08:16):
So one caveat before I answer that, I think that the game has changed over the last six months. So I still think a technical co-founder is a requirement and it’s something you should invest in. But you can build products way more sophisticated than ever with these new AI tools and coding agents. I still think it’s really important to have somebody that knows how to build infrastructure and can identify blind spots. You don’t want to put all your eggs in one basket and ultimately vibe code everything and then ever hire a technical team. I think that’s going to blow up at some point, especially if you’re building for the enterprise, which is what we do, right? We focus on HubSpot, Webflow, ZoomInfo, some big companies, right? They need to have enterprise grade code bases and like suck to compliance and other things so truthfully like I feel like building software when I started two years ago with this new business is expensive right like and you’ve got to think about how do I get this off the ground how do I build an MVP and I said like okay if I go and hire an agency and I got quotes by the way because I spent six months trying to find the right technical co-founder kissed a few frogs along the way the hard thing was like if I hire a technical co-founder that’s a hardcore coder he will build a product for me now we had to commit and get married and do a big equity share so we spent some time getting to know each other and make sure that we were the right fit and that we matched each other’s blind spots and ultimately it was a success in the end he was awesome his only requirement was at the time I had a team of maybe four engineers and his only requirement was that I fired the whole team if I was to hire him because he wanted to build it again from scratch that’s like okay so to the founders out there listening like technical co-founders are awesome right like in the end of the day I’m a business side like my mind is on the marketing the branding the finance the operations I need somebody to rely on that I trust to be able to execute and build a product and infrastructure the right way so that we don’t have to go and rebuild it again so I would say always invest in technical leadership at the early stages ultimately it doesn’t have to be the perfect person from day one right you can evolve and hire more people over time but you definitely want somebody that knows what they’re doing

Brian Bell (00:10:14):
Yeah, I often say that a business is fundamentally two things. You’re like making things and selling things, right? So who’s making things on the team and who’s selling them, right? And it’s very rare you have a world-class person in both of those domains. Can happen, but you’re better off having a division of labor. I think you can just do more and de-risk the startup dramatically by doing that. What other lessons do you take away from scaling Matheson?

David Walsh (00:10:38):
Yeah, I mean, all the founder mistakes that you can make, you know, decided that the things I wanted to focus on was how many people I employed and how much money I raised. And we did both of those things really quickly. And look, we were in a space that exploded in 2019. Diversity recruiting was where we were focused. And part of that was because I had built a indeed the large job site. When I was there, I came up with this idea, which was a marketplace of international talent. And so I ended up evolving that into this HR software company that was mostly talent acquisition. So I think like I had it backwards and now my mindset has changed quite a lot. It’s how do I keep the team as lean as possible and raise as little as possible to stay in control for as long as you can. And now at some point you scale exponentially and you need capital to grow. Don’t get me wrong. I think that’s absolutely critical. But to start off with, I think the focus is on users and customers more than ever. So I got a bit disconnected in my last company, I would say a few levels of management away from the user base. And I just like was making decisions with not enough information around the strategy. And you know and I’ll give you a story like an example we had a recruiting product and we had like a policy and procedures product and that was like a scoring system of talent acquisition processes and we moved away from the recruiting product and I felt like that was something I wanted to do aggressively and I probably did it too aggressively and then you know we realized that actually a lot of our customers really needed and wanted the recruiting product and we had kind of ripped it out or less focused on it less resourced for it so I think like the learnings are you know hire a great technical co-founder be the spokesperson for the brand from day one and be open to just building in public which we can talk about it have I done that with this company and stay as close as possible to your customers for as long as possible and when you think that you had product market fit you most likely don’t so don’t you know go hire a ton of people and then end up reducing your runway and increasing burn too quickly make sure that you build it sustainably

Brian Bell (00:12:33):
I love that and I’d love to tug on the product market fit thread a bit I mean how do what would a founder know you’ve done this a couple of times now how do you know you have product market fit what does that feel like on the day to day

David Walsh (00:12:43):
So I’m super critical of our own product and business all of the time. That’s my mindset. It’s this can be better. This isn’t good enough. And that’s the mindset I take every single day. So I would start by saying, I still don’t think we have found product market fit per se. We’ve had some success and we’ve got great companies using the product. But you need repeatable motions and something that’s stable and you need it for a period of time that you can say, okay, if I resource this and put money into growing this, it’s going to repeat the same function over and over again. And so product market fit changes, right? And it changes with the market. Sometimes you’re in control, sometimes you’re not in control. I think the things that I would look out for is our customers pulling the product from you right are they logging in doing things that are inefficient and I’ll give you a story one of our enterprise clients was going into limelight every day and downloading leads right they were actually downloading leads and then uploading them to their system every single day and I was like if they’re willing to do that painful of a thing every day then this is the thing we need to continue to build on so I think there’s constantly you’re evolving constantly you’re staying close to your customers and ultimately you just want them to be coming to you with requests versus you building features for them without validating that they need this and it solves a problem now there’s many other facets of product market fit but that’s the one I would focus on to begin with at least make sure you’re close to customers and they’re asking you for things and you’re not coding things before you know they’re getting asked and truthfully my CTO is giving out to me right now even just yesterday because I asked him to build three features and not enough people are using it so I’m learning every day too

Brian Bell (00:14:16):
Yeah, I love that. So Limelight, let’s talk about that. I mean, it didn’t start as Limelight. You actually did pivot. Maybe you could talk about the original vision and what caused the pivot. And when did you know it was time to pivot versus persevere?

David Walsh (00:14:27):
yeah that’s a really great question so i started with this thesis as i mentioned originally which is i want to focus on marketing leaders in b2b so i know build build b2b software for marketing leaders and the thesis was simple they need more efficient ways to grow right customer acquisitions costs are going up digital ads are more expensive it’s more saturated than ever how do we create a new growth mechanism for these people who are given a budget and the budget is sometimes now decreasing every year and they’re expected to deliver results so they’re constantly under pressure to figure out where the growth mechanisms so we started as Bundle which was a referral software for B2B companies so influencing your own user base to introduce more customers and I started building that infrastructure the payment processing the referral infrastructure and then and I started posting content on LinkedIn every day as part of this when my last company I didn’t speak publicly about the brand and this time around I went the opposite direction I actually went way too over the top with building in public from day one and because of that I started creating content I realized there was an opportunity to build for the creator economy in B2B there was thousands of individuals who have built large audiences that we could build a product for and help them do brand partnerships and so I pivoted into it because I became a creator myself now the thesis of marketing leaders need more efficient way to grow was still the same right it was just they didn’t just grow from their own users who might be influencers they grew from our network of influencers so I said I thought that the tapping into that audience growth would be more beneficial and it would help us scale quickly because when you put ten thousand dollars into the back of a creator you know into their back pocket every month they love you and they help you grow so before maybe for founders and other people listening to this you don’t just pivot immediately without data to support it so we did a ton of research we went out and I interviewed 100 B2B content creators I DMed them all on LinkedIn and emailed them and I said jump on a 15-minute call with me I want to meet you I want to understand why you create content why you haven’t done brand partnerships after doing 100 interviews I realized okay there’s an opportunity here and then on the brand side I did the same thing I went and met with 20 different heads of marketing and I said here’s the mock-up simple mock-up would you buy this product they said yes we said okay it’s time to move you can either pivot now or pivot in six months might as well try this and go all in and so that’s what we did

Brian Bell (00:16:39):
Yeah, I love that. And I talked to lots of founders who were just starting. And a question I’ll ask is, how many people have you talked to have you validated the problem and i mean you very much did that i mean talking to over 100 b2b creators and 20 corporate brands you really validated hey there’s an issue here and it’s surprising to me how little customer discovery where did you learn that skill because you’d already built a startup how did how did you know to do that so it’s

David Walsh (00:17:08):
It’s founder one-on-ones. Like if you don’t go speak to your users and understand them, you’re going to build the wrong product. And I’ve done that before, build products with this, you know, without speaking to users. So I made the mistakes in the past and I felt like, okay, this time I need to get it right. Now, truthfully, I was building, I was creating content on LinkedIn every day anyway. So I was really passionate and interested in this space, which made the conversations as enjoyable you know for me as they were potentially for that person right which was I was learning they were learning and I was just gathering data so it didn’t feel like you know your typical I have to do a hundred meetings and you know it’s going to be a drain and I’m going to be interviewing these people like if as if you’re hiring them it was like I’m learning a ton and I’m fascinated by this and like everyone’s coming at it with slightly different perspectives so I think I learned a ton very quickly recorded all of the calls with Read.ai which was our note taker emphasized that downloaded the transcripts uploaded them to LLMs went deep into okay what are we actually hearing here what are real quotes from these potential users and I just loved it right and I think we did well by doing that Brian the last thing I would say is like one of the most important decisions I made at Limelight was I hired 25 of the best B2B influencers I had met through the interview process in Discovery and I made them I would call them like influencers influencer advisors we give them a few shares we get them to be part of the program we get them to advise us on the product delivery and then they helped us exponentially grow so I have this army of creators who are engaging on my content who are part of the journey who are on my investor updates and that’s been one of the most impactful things in fact if anyone’s listening to this right now that’s starting out and trying to build a product and thinking about distribution go to influencers go and you know attract them to become part of your journey now it was easier for me because I was building a product that would get them paid so often I was putting these influencers that were on my advisory board at the top of the search results so they were getting more inbound and more demos and or sorry more like proposals to get paid for brand partnerships it was self-fulfilling and it became this flywheel that helped us grow for context I started at 2,000 LinkedIn followers I’m now at 41, 42,000 LinkedIn is our number one driver of growth I’d say 90% of our revenue comes from LinkedIn it’s been really successful for us just given the space that we’re in

Brian Bell (00:19:29):
Yeah, I love that. For someone who’s never heard of Limelight, explain the model. How does this marketplace work for both brands and creators?

David Walsh (00:19:37):
The, we decided with this idea of like, let’s make available ad space. Let’s pull available ad space has never been, I suppose, like what’s the right word like people have never been able to identify creators at the scale that we wanted to so simply put Limelight is a B2B influencer marketplace we have end to end creator management so we help them find creators organize them through a CRM and then measure the results we also added social listening so you’re constantly tracking leads that are being generated from the content partnerships because in B2B the buying journey is a little longer and there’s many touch points so we have to collect enormous amount of data and the theory was like let’s just create available ad space that hasn’t been available before and give it to these brands to be able to buy this ad space almost in like a programmatic way but using influencers and creators and if you’ve never done influencer marketing and this is brand new and you’re just listening to this it’s super messy it’s like the most messy process in the world because you’re dealing with human beings right so communication negotiation contract content review there’s all these like inefficiencies in it so we tried to build a system that can help them align brands align all of that and ultimately if they’re spending a million dollars on paid ads every month take 10 of that put it to creators and influencers instead of hiring a big agency who are going to charge you a black box percentage fee use our product and you get way more transparency in a scalable motion that allows you to do this without hiring a ton of people

Brian Bell (00:21:21):
Yeah, totally makes sense. And, you know, marketers constantly struggle with attribution. How do I know that that LinkedIn post actually resulted in a closed win opportunity? How do you kind of address that in the platform?

David Walsh (00:25:41):
Yeah, it’s tough. Attribution is so challenging, especially in B2B, right? So we would typically say, look, you have to have the standard things in place. You have to have UTM links that are customizable. You have to have self-reported demo attribution. You’ve got to measure your website traffic as content goes live. But no matter what, some people might read a newsletter or read a thing that they see on the newsfeed and then go elsewhere later on, not be in market and then come in market and not tell you where they came from. Right? So because there’s many touch points and longer sales cycle, it’s harder to identify. What we focused on is obviously all the standard stuff, which is like being able to measure impressions, engagements, reactions, click through rates. But we went one level deeper where we started to track everyone engaging on the content and enrich those leads and show the brands who their ICP like the job titles the person the company that is engaging on the content and then tie that all the way back to the sales pipeline and the revenue pipeline so what we’ve done for one of our big brands I won’t mention who but they were beta testing this product with us over the last few months they’ve generated 25 million in pipeline and 8 million in closed won business from just tracking influencers and their employee content seeing who engages on that content enriching those leads and then verifying that those leads were convert when they convert that there was touch points so the ultimate holy grail for us is to make b2b influencer not just brand based it’s performance based we want to unlock the paid ad budget at a lot of these companies and to do that we’re collecting a lot of data that supports that it’s a good investment and ultimately just driving leads

Brian Bell (00:40:39):
Well, David, thanks for coming on. Learned a lot. It’s always fun when I have a portfolio company on and I’m proud to have invested. Thanks for letting me be on the journey with you and wish you all the success in the world.

David Walsh (00:40:51):
Brian, you were one of the fastest decision makers when getting involved in the round. We met on Wednesday and I think you told me the next day I’m in and I just love that level of commitment and kind of validation. So great to have you involved.

Brian Bell (00:41:05):
Yeah, I appreciate that. And you’re actually, you know, my wife’s company, your 360 AI is using you guys and they really like you. And I think there’s at least one or two other Team Ignite portfolio companies using you guys, but we should definitely make a stronger push there.

David Walsh (00:41:18):
Yeah, well, don’t judge us on the product today. I’m very critical of our product. I think we can automate a lot of the workflows. I think a lot of it’s somewhat manual today and that’s what we’re excited about in the short term. Awesome.

Brian Bell (00:41:28):
Well, thanks so much, David.

David Walsh (00:41:29):
Thanks, Brian. That was awesome.

They’re building like it’s 2016.
Raising like it’s 2018.
Pitching like capital is still cheap.

It’s not.

In a recent conversation with Ben Narasin, Founder and General Partner of Tenacity VC, one thing became clear fast: the rules of venture haven’t just shifted. They’ve been rewritten.

Subscribe now

Seed Is Now Series A

The biggest misconception founders carry today is about stage.

What used to qualify as Series A is now seed.

A decade ago, a strong founder with a compelling idea could raise a Series A. Today, that same profile struggles to raise a seed round without traction.

Ben put it plainly: if you’re raising seed in 2026, investors expect a real business. Product in market. Customers. Revenue. Often $500K to $1M ARR or more.

That has a downstream effect.

Series A is no longer about potential. It’s about proof.

You’re not raising to figure things out. You’re raising because you’ve already figured something out and need capital to scale it.

If you don’t adjust to this reality, you’ll keep wondering why the market isn’t responding.

Venture Is Performance-Based Now

There was a time when top-tier founders could raise on pedigree.

That still happens, but it’s rare.

Today, venture looks more like a merit system. Thousands of companies raise seed. Only a fraction earn the right to raise Series A.

Ben described it as a shift from “Olympics” to “open competition.”

Anyone can enter.
Only those with real traction advance.

That means your growth rate matters more than your story. Your metrics matter more than your resume.

If you’re not hitting clear inflection points, the market will move past you.

Great Investors Don’t Diversify. They Concentrate

A lot of advice around venture sounds logical but breaks in practice.

Portfolio construction is one of them.

In theory, investors spread risk across many bets. In reality, the best investors look for a small number of companies that can return the entire fund.

That changes how they evaluate you.

They’re not asking, “Is this a good business?”
They’re asking, “Can this become a $10B company?”

If the answer is no, it doesn’t matter how solid or profitable your business looks.

This is where many founders get stuck. They build something that works, but not something that scales enough to fit the venture model.

The Bar Is Higher Than You Think

Here’s where expectations break down.

A million in ARR used to unlock the next round. Now, it might not even get you in the room.

Ben shared examples of companies hitting multiple millions in revenue and still struggling to raise.

Why?

Because the market compares you against your peers, not against history.

If others are growing faster, raising more, or showing stronger signals, they get funded first.

This creates a simple but uncomfortable truth: good is no longer enough.

Storytelling Is Not Optional

Founders often treat storytelling as a “nice to have.”

It’s not.

It’s how you hire.
It’s how you raise money.
It’s how you sell.

Ben emphasized that even strong businesses can struggle if the founder can’t communicate clearly and convincingly.

Investors don’t just evaluate your numbers. They evaluate your ability to attract capital, talent, and customers over time.

If you can’t tell a compelling story, you create friction everywhere.

And in a competitive market, friction kills deals.

What Actually Matters When Investors Decide

After decades of building and investing, Ben simplifies his framework to five things:

People, people, people.
Great idea.
Huge market.

That’s it.

But there’s a catch.

You have to trigger conviction fast.

If an investor doesn’t feel a strong reaction in the first meeting, it’s already trending toward a pass. Not because you’re bad, but because the bar for attention is extremely high.

This is where many founders underestimate the importance of first impressions.

The One Trait That Decides Everything

You can have a great idea and still fail.

You can raise money and still fail.

You can even have early traction and still fail.

The one trait that consistently separates winners is simple: tenacity.

The ability to keep going when things break.
To adapt without losing direction.
To push through rejection without losing momentum.

That’s not a soft trait. It’s the core requirement.

Because the reality of building a company is this: most of the time, it’s not working.

Final Thought

If you take one thing from this, it’s this:

The market is not easier. It’s sharper.

Investors have seen more.
They compare faster.
They expect more, earlier.

That doesn’t mean it’s harder to win.

It means you need to play the right game.

Build something that grows fast.
Tell a story that people believe.
And stay in the game long enough for it to work.

Everything else is noise.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:

00:01 Introduction to Ben Narasin & Tenacity VC
02:50 Seed vs Series A in 2026
05:33 What Pre-Seed, Seed, and Series A Really Mean Now
07:48 Investing for Tier-One Follow-On VCs
10:26 Venture Capital Archetypes & Solo GP Model
14:46 Solo GPs, Fund Performance, and Market Dynamics
16:56 Raising from Venture Firms as LPs
21:16 Fund Size, Deployment Pressure, and Discipline
25:50 Venture Cycles: AI vs Past Tech Booms
28:11 Portfolio Construction & Check Size Strategy
33:18 Check Size vs Investment Strategy
35:19 Decision-Making Discipline in Investing
38:51 Passing on Founders & Honest Feedback
41:09 Fostering Dogs & Personal Values
42:45 Thoughts on Web3 & Blockchain
45:38 Speculation vs Long-Term Investing
47:50 Why Venture is a “Never-Ending Education”
49:58 Investment Approach & Market Focus
50:05 Future of Venture & Geographic Shifts
55:30 Government, Incentives, and Capital Allocation
01:05:31 Lessons Learned in Venture Capital
01:07:50 What Ben Looks for in Founders
01:10:57 Importance of Storytelling
01:12:31 Trusting Your Gut in Investing
01:14:40 Solo GP Advantage & Decision Autonomy
01:15:57 Venture is Not Roulette: Poker Analogy
01:18:25 Knowing When to Walk Away
01:20:12 Final Thoughts & Closing

Transcript

Brian Bell (00:01:07): Hey, everyone. Welcome back to the Ignite podcast. Today, we’re thrilled to have Ben Narison on the mic. He is the founder and general partner of Tenacity Venture Capital, the firm focused on backing transformative pre-seed and seed stage founders and connecting them with top investors for following rounds. Ben has spent decades as an entrepreneur and early stage investor, including founding and exiting many the companies that you’ve probably heard of, also investing in traditional firms before launching Tenacity in 2021. His investments span FinTech, places SaaS, hardware, and more. He writes and speaks widely about startups and staying entrepreneurship. Thanks for coming on, Ben. Hey, thanks for having me. It’s always fun. Well, I’d love to get your origin story. What’s your background?

Ben Narasin (00:01:46): Sure. So I basically have been an entrepreneur my entire life. I just currently live vicariously through other entrepreneurs to fulfill that part of me. I started my first company when I was 12 and businesses for about 25 years. The one that’s most relevant to the conversation most people are interested in having is that in 1993, I started one of the first dot coms in the world in New York City. I’ve been told we invented, invented, whatever that means, cost per click in 94. Took that company public in 99. Never raised venture. So I had enough ownership in that business that when the bubble burst, I was able to take it private and treat all the shareholders well. Everybody was treated the same way, even though everybody wanted to treat me better because I owned the majority of that company. So from there, came to California, got here about 22 years ago. and saw this very large gap in the market, which was that angels were a small, slow processed way of raising money and venture was getting very big. And so smart people with great ideas that only needed about a million dollars to prove or disprove their thesis didn’t really have a great place to go. And so I started what ended up being the second institutional seed fund in around 2007. Dad for about eight years, got to know well over 300 tier one vcs had a phenomenal year where lending club went public in the largest tech ipo of the decade and zenefits raised an equivalent amount of money in a private round was declared the fastest growing sas company of all time a lot of venture guys said hey Maybe you should do Series A and B. So did traditional venture capital for about seven years, mainly at NEA. Awesome firm. And then spun out of there 2021, as you mentioned, to start Tenacity. I didn’t go to NEA just to do Seed, but it was something that I learned while I was there was what I really, really loved. Because I did later stage investing in NEA and Seed, and I just get so much more joy and pride and just generally am better rewarded. for doing seed so spun out started tenacity with their support actually have 40 different venture firms and partners as my backers in tenacity fund one and i’ve just finished investing on a fund one i’m now investing on a fund two wow congrats

Brian Bell (00:03:45): i mean it’s such a journey and there’s so much to unpack there i’d love to like just get your thoughts on on seed versus a as it as it stands today here as we record in 2026 somebody that’s done both of those styles of investing and uh yeah i would love to just get the download there

Ben Narasin (00:04:00): The easiest way to put it is that the seed of 2026 is exactly the same as the Series A of 2016. They have changed in name only. Form is the same. Terms are similar. Prices, amount of money. They’re even... So when I was running the equity practice for TriplePoint, which is a large venture debt shop, I was exclusively focused on funding companies that I believe could be funded by the top 10 to 12 venture firms in the world. So I got to know those people quite well and tried to understand what they were looking for. I would think of those as the tier one firms, right? So several of those firms today say they do the same thing they’ve always done, but they call it seed instead of Series A, Foundation, Menlo, a couple of others, because now a seed could easily be $6 million into a company, 24 posts. So what we were doing then that was Series A is today’s Seed. Seed has fully and entirely replaced it. In fact, you mentioned I write about this a lot. I mean, it must have been 10 years ago I write a story for TechCrunch about how Seed was basically becoming the new Series A, and now it has just taken over. Series A’s today are basically the Series B’s of yesteryear. There’s a different reality. And this has changed. One of the things entrepreneurs don’t always understand about this is that Series A’s used to be like the Olympics, the best and brightest that have already won. its best runner to compete in the Olympics. Top guy or gal out of Cisco or Google goes out and raises a Series A on a great idea. And they say, oh, this is an awesome person. We got to back them. Now, it still happens on occasion. It happened with Neva a while back with the top guy out of Google. But normally what happens now is Series A has become CrossFit. Anybody in the world, and these days about 19,000 people a year, can raise a seed round and go compete. And if they win, guess what? They get a Series A. But it’s based on performance. So it’s no longer the Series A of yesteryear where it’s like, well, great person, great idea, interesting tech, let’s try it. It’s great person, great idea. And it’s working and they’re getting revenue growth that really impresses us. Okay, we want to add fuel to the fire. So Series A has become a fuel on the fire around when it used to be an initiation.

Brian Bell (00:06:02): Yeah, and it sounds like what you’re describing went even earlier to pre-seed, right? Hey, great person, great idea, maybe a little traction.

Ben Narasin (00:06:09): Pre-seed today is exactly what I was doing in 7, 8, 9, and 10 and calling seed. Smart people with great ideas, nothing. Today, when I do seed, I’m expecting revenue. I’m expecting a business. I’m expecting live product, customers, all of that.

Brian Bell (00:06:22): Yeah, and then what’s your kind of bar for seed now these days, like in terms of revenue, traction, stuff like that?

Ben Narasin (00:06:29): I usually say half a million to a million, but I back companies with more. I mean, I’ve had seeds I’ve done with over 2 million, which is amazing because it used to be. So this is one of the things I think entrepreneurs unfortunately have not updated in their lexicon of realities of venture capital. It used to be the case. And by used to be, I’m talking about all the way back in the day when I started out 7, 8, 9, 10, that if somebody seeded you, and you could get to a million in revenue in a year, you could do a Series A. Today, that’s a C, right? So if you want to do a Series A, I think you really need to get to a lot more like three or just have torrential growth at the very end of the tail. So you can get from zero to a million in six months and have a shot, but I think you get from zero to a million in a year, there’s plenty of people that got to two, two and a half, three, and they’re going to get funded ahead of you. So that reality has changed a lot.

Brian Bell (00:07:18): Yeah, I would echo that. I see that a lot of our portfolio companies. We’re very much a pre-seed mostly, probably 70% is pre-seed, and then 20-ish percent is seed. I don’t really do A, but I see a lot of companies fail to raise the A because they just don’t hit that three, four, five million RR, or it’s not at the right inflection.

Ben Narasin (00:07:37): I mean, I had a company that went zero to seven and couldn’t raise an A.

Brian Bell (00:07:42): What do you think it was? I mean, zero to seven is incredible. What do you think it was? I know.

Ben Narasin (00:07:47): It was a fintech business. They were too dependent on lending. The timing was such that people were very worried about interest rates. There’s a lot of sensitivities in fintech that can kill a business that’s otherwise entirely reasonable. And I think they just hit it at exactly the wrong time. But the second business they’re in is doing exceptionally well.

Brian Bell (00:08:05): And I’d love to talk on this thread of your thesis, which is investing in things that I know the tier one firms are going to like. Is that still what you’re doing at Tenacity or are you kind of

Ben Narasin (00:08:15): more like over 300 i last when i left it’s funny when i joined triple point they asked me how many vcs i knew and i was like i don’t know so i wrote a list down of people i thought might be vcs it was seven or eight people when i left that list was 327 people long and it’s been a long time since i left so i know quite a few more so those are the people i take my deals to obviously first among equals amongst my LPs. If you’re one of my 40 BC LPs, you’re going to see the deals. You do have to be a fit. If the perfect fit is not one of my LPs, I’m still going to take you to them. I’m super picky about who I introduce people to because one, I never want to have anybody refuse an introduction. And I’ve had that happen four times in 18 years.

Brian Bell (00:08:49): I think it happened four times today for me.

Ben Narasin (00:08:55): Well, by the way, everybody that said no, it was foolish of them to do so. But clearly I did not have the relationship I thought I had with If I’m bringing somebody a deal, it’s a fundable deal. They may not fund it, but that doesn’t mean it’s not a fundable deal. That may not be a fit for them, but this is a series A that’s definitely fundable. Somebody’s going to back. Back in the day when VCs were trying to figure out how to work with seed funds better, they started to do a lot of dinners for seed funds and they wanted to tier one firms, won’t name them, but great group of people. I had a dinner and had me and about 18 other people there and they were talking about deal flow. And they said, you know who does a really great job at deal flow? Ben Harrison. I was like, thanks. No, I had already agreed to join NEA. So I didn’t want to talk at all at that meeting. But they said, we have never seen a deal from him that was not a fundable deal. It’s like, we’ll never not take a meeting. And that’s my goal to make sure. And so I can’t take an ugly baby out. right? It’s pretty babies only. And so I’m very focused at the very beginning about thinking about what will this business look like a year from now? I always say I need five things to make an investment. People, people, people, great idea, huge market. So, you know, I’ve had on multiple occasions, I’ve asked entrepreneurs to go ahead and write me at the seed, their series A deck for a year later. So I can see what their thought process is, where they think they can be, make sure we’re aligned. And that’s been quite useful.

Brian Bell (00:10:40): But it has to be a really interesting philosophical point that I’d love to ask you, which is like, does venture make the man or does man make the venture? Well,

Ben Narasin (00:10:48): venture is a very, it’s a novel, sometimes the woman, but unfortunately rarer than it should be. Venture is a novel industry. It may be unique. It’s, first of all, it draws a certain type of person. Now that’s evolved. It’s not the same as it used to be. Well, actually, maybe it started off one way, which was more professional and then became more entrepreneurial and has gone back to being more, quote unquote, professional ex-bankers and such, which I’m not as big a fan of. But there are some exceptional folks that weren’t entrepreneurs that are still great VCs, which I didn’t believe could be true. When I moved to Silicon Valley, it was beyond my comprehension that somebody could be a phenomenal VC and never have been an entrepreneur themselves. But what I learned was they can be in certain cases. Like Scott Sendell, probably arguably one of the top three investors ever in venture. He was at Microsoft. So he had operational experience, but within a limit. You’ve got folks like, I don’t know, Mike Moritz was a journalist. That’s a classic one. Bill Gurley came out of banking, if I remember correctly. So There are people that are great VCs that were never entrepreneurs, but there’s also entrepreneurs like David Zee or Reid Hoffman or many others or Alfred Lin who were entrepreneurs first and now have become exceptional. Like you can just, having lived through people’s lives is enormously valuable to be able to understand them, emphasize with them and provide value. Look, I spent 25 years as an entrepreneur learning hard lessons. I would prefer if I can help my entrepreneurs learn those hard lessons without suffering the pain

Brian Bell (00:12:17): Yeah, that totally makes sense. So I guess stated another way, there’s this phrase in venture capital, you’ve met a family office, you’ve met one family office, you’ve met a family office. Kind of the same could almost be said for a VC. You’ve met one VC, you’ve met a VC.

Ben Narasin (00:12:30): Yeah, and not any of the others. I think it’s a bit more homogenized, but it may be sort of that there are tribes and camps in venture. Like there is definitely the ex-operator VC, there’s the ex-entrepreneur VC, there’s the ex-top dog inside of a company who got founded and funded by a venture firm that then hired them in. And then there’s the people that came up the road through the, I went to a good school, then I went to a good bank, and then a recruiter brought me in as an associate, and then I rose to the ranks. The last one being the most homogenized of all, and I think actually a huge risk to venture capital in general. I mean, I loved the associates and principals and, you know, people I worked with at any, they were awesome, but all but one came out of banking. I am not a believer in forcing and having a false set of diversity, but I do believe a diversity of mindset is certainly useful. Like if you don’t have different people, and I don’t care what creates that diversity of mindset, but if everybody’s thinking the same way, looking at a problem, it’s going to be challenging for you as a group. And look, I loved my partners at NEA. I did not love the partnership experience. So I’m a solo GP. I’m the exact opposite of a partnership. There are three of us that have debates when a decision is made on investing after having consulted many experts and doing diligence and working with our pills, my LPs and others. Me, myself, and I. And I actually argue with myself all the time. I yell at myself on occasion. I tell myself I’m an idiot. But at the end of the day, that’s my process. And it is the opposite of a process where you spend months on a deal and then bring it into a bunch of people who’ve never heard of it and have to have a guy tell them in an hour what they’re doing. And then you get to spend half an hour justifying why you should make the investment. Anyway, just not my way of doing things. And I think so far my track record would show that my decision-making process has been adequate to more than adequate.

Brian Bell (00:14:23): Yeah, I haven’t seen, and I’m a solo GP as well, I haven’t seen any data on this, but I would guess that solo GPs have a higher variance around the mean. What I mean is like the good ones are really good and the bad ones are really bad.

Ben Narasin (00:14:37): Well, it seems like the data shows that all the alphas coming out of solo GPs. Certainly were a great performing asset class, if that is such a thing.

Brian Bell (00:14:44): I mean, it’s hard to call an individual an asset class.

Ben Narasin (00:14:46): Hey, look, there’s 12 great single operators. That’s an asset class. But- You know, it does take a lot to raise as a solo GP, as you know. And so part of it is a natural filtration process of whether you’re justifiable to be able to raise money in the first place. And that would normally be tied to your track record or some specific insider experience you have. Admittedly, 2020-21, a lot of people raise solo GP funds that never should have, and they will die a painful death. They’ll milk the fees out for the 10 plus two that they have, and then they’ll be gone. because they’ll never raise fund two.

Brian Bell (00:15:23): Yeah.

Ben Narasin (00:15:25): You know, because they won’t be able to justify it because they won’t have returns. They just, they just, they never should have done.

Brian Bell (00:15:30): And die by our returns. Have you had any friends like that, that, you know, they had a good fund one and maybe tanked on fund two or three and then they were out of the business?

Ben Narasin (00:15:37): No. I mean, I’ve funded half a dozen folks, all of whom were like several were my entrepreneurs in the past and are now managers, and they seem to be doing quite well. It would be hard for me to comprehend that I would be friends with that type of human. And by the way, all my friends are business-related friends. i have a lot of acquaintances so you know like i have a healthy disrespect for some of the pitches that were made as to why people should be solo gps and i’m not shy to i mean i’m not a jerk like i said you’re an idiot you shouldn’t do this but i’m not going to fund them and so not yet i certainly hope i don’t because that would mean i poorly chose in my investments in those half a dozen or so funds that i chose to back right

Brian Bell (00:16:16): So it sounds like you raise from a lot of your follow on VCs, which I think is a kind of a differentiator. I don’t know if I’ve met a VC quite like you. Tell us about fundraising from later stage VCs. What’s that process like?

Ben Narasin (00:16:28): So way back when, when I first started doing seed, I always thought it’d be interesting to have a fund solely backed by other venture firms because my pitch to them back then was, are you people doing seed? You shouldn’t, you’re wasting your time. I will do the seed deal. I will bring it back to you when the time comes, if it’s appropriate. But I got to know so many people, you know, it was a combination of NEA as an investor. And then I have founders of funds. I have GPs at funds. I have younger people at funds. Some of them, you know, the guys that were the founders of funds were very senior at funds that could put in a million dollars. They, I think, wanted to see me be in business. In fact, many of them said, wow, finally? I mean, we’ve been waiting for you to do this. And look, they’re buying a deal, let’s be honest, right? Yes, they want to support me. What’s interesting is I think the conversion of those fund one investors to fund two will be exceptionally low because they wanted to see me get into the business and they wanted the deal flow. I don’t think they need to worry about venture diversification and anything else to continue on. Some will, sure. Because financially, you know, right now we’re showing great numbers, but on a market basis, there’s no liquidity for seed funds this early in. But we’ll see. I have a feeling I’ll have a whole bunch of new folks that are also a venture. So I have exited entrepreneurs. I have unexited entrepreneurs. I have senior VCCs. I have junior VCCs. VCs. I have venture funds. I have LPs that are family offices. I have fund of funds. No, it’s a small fund. We went out to raise 50, ended up with 60. Fund two, I don’t want to be more than 50. I think it was probably a mistake to let myself be oversubscribed. But we’re not a target for these larger traditional institutional LPs. In fact, there was an investor who had invested. They were the largest LP at a large firm that I worked for. I’ve worked for three. So You know, it doesn’t necessarily tell you who it is. And they had invested directly in all my deals. So I went and talked to them and they said, Ben, our minimum check size is the same size as your entire fund. So, you know, the person did invest personally. So I’m happy about that, but it would not have fit for that fund. I mean, a lot of these larger institutions, it’s a challenge. Like even when you talk to folks like at Cambridge, they know the alpha is coming out of the smaller, the solo GPs, the more focused funds, the sort of call it 25 to 100 million, but they just don’t have a way to institutionalize that on behalf of their large clients, unless they create like a sub fund, which they aggregate a lot of other people into. And the other thing is, even though there’s a lot of alpha, you still got to pick, well, not that different than picking great entrepreneurs. Like, I think if I could run another life, like I’ve had multiple lives in my timeframe of having been around this planet for a while, I wouldn’t mind being either An LP myself or a consultant LPs? Because I think generally speaking, they make suboptimal decisions. I don’t think they’re solving the right things necessarily. I think they have, you know, they have to spend a certain amount of time. It looks like they’re doing their job, which is okay, fine. But are you really digging into the right stuff? Some of this stuff they’re digging into is totally, completely irrelevant. And then they’ll just not even bother to talk about the stuff that is. I was amazed. It was sort of like my IPO roadshow. I always say when I did my IPO roadshow, it’s the same as pitching a series A deck. And it’s the same as raising a fund. You tell the same story hundreds of times. And you get the same questions hundreds of times. And they’re reasonable questions. But twice in my IPO roadshow and twice in raising a fund, did somebody ask me a question? I was like, wow, that is legitimately a really good question. I’m going to have to think about that because everybody said that’s good. question. It’s like an entrepreneur. Oh, Mr. VC, that’s a great question. I have a slide for that. I want to do a fake pitch one day where I go out to pitch to VCs and they ask me a question. I’m like, you know, that’s a really common and quite frankly, just totally stupid question. It’s not even really worth my time answering. But if you want to look to the appendix page 72, it’s in there because everybody asks it. But thank you for your limited amount of understanding of the industry that I’m in. Obviously no one will do that when they’re raising money. So everything’s a great question. But on occasion, someone asks you a question that makes you think, and I love that. Rare, but valuable.

Brian Bell (00:20:09): Yeah. So I’ve heard that in the industry, it’s harder to raise a $30 million fund than $100 million fund. And I think what you’re describing is the institutional check size, right? They have to write 10, 25, $50 million checks. And so there’s kind of this push in the industry to get larger, faster.

Ben Narasin (00:20:26): Yeah, I think there’s the virtuous cycle. There’s the anti-virtuous cycle in venture. which is to me, the death of venture over time. One, greed for fees, okay? Which is stupid because we make our real money off of carry and we give our fees back. Two, greed for fees creates bigger size funds. Now, the problem with that is bigger size funds can’t just go out and hire people. That takes forever. It’s not a process you want to go through. It’s a very Sisyphean process. So what does it create? Bigger check sizes and lowering the bar. So that cycle destroys venture. I’m trying to raise a smaller fund for fund two. Number one, it is just me. I’m pretty comfortable with what my capacity is. I’m pretty comfortable with how many hours exist in a day. And I’m pretty comfortable with the fact that it is virtually impossible for me to work harder than I already do. There was a period at the end of last year where I think in four weeks, I was in six different cities and four different countries. And I’m not 23, even though I want to act like I am. I mean, I saw 10,000 pitches in 2023 and 2024, and I funded six. Should I just see more? Where are they? Are they available to me from 8 p.m. to 2 a.m.? Should I be time shifting and just doing demo days in London and Europe? You just can’t lower the bar. and you’re going to have to if you raise more money. Or you’re just going to have to move upstream and do later stage investing.

Brian Bell (00:21:51): I think you made a really good point that I’ve never really heard anybody say it quite like you just did, which is, yeah, we charge fees. Yeah, it’s 2%. But we pay those fees back before we get carry, which is where the real money is. I think that’s a really important point that I don’t think a lot of people appreciate is, yeah, we are paying ourselves 2%. But before we collect any profit sharing, we’re giving all the expenses of the fund back, all your principal back.

Ben Narasin (00:22:15): I mean, you can’t not run a business. Look, personally, if somebody’s watching your show and they reach out like, Ben, you’re awesome. I want to give you your whole fund. I’ll give you 50 and I’ll do it on a zero in 50. I won’t pay you a penny, but I’ll give you half the profit. Well, I’m going to have to spend a long time getting to know you and know that I trust you and that you’re a rational person and that you’re a good person and somebody I want to be in business with for a decade plus. But I would love to do a zero in 50 fund. I can pay my bills. Now, do I want to pay my bills? I would obviously prefer not to dip into my savings and net worth to pay my bills. The point of the structure that was created for GPLP, and it gets warped when it gets too big, and the reason LPs complain about fees. I was having this conversation and Beezer Clarkson was speaking somewhere and I really like Beezer. She was complaining about fees. I’m like, why do you care? We give them all back to you. She said, well, Ben, that’s fair for a fund like yours, but I’m talking about people running multi-billion dollar funds and going out to I’m not going to mention the obvious firm that has pretty much stamped the term fee whore on their forehead that raises billions and billions and billions and billions of dollars, and the founders of which will be dead long before those fees are repaid. I don’t think they’ll make a return anyway. So yeah, but then my question would be, why are you funding them? Why are you funding somebody where you don’t think you’re getting your fees back? That means they didn’t make you a return. That means you didn’t get a multiple. If you’re getting a multiple of size, you’re going to get your fees. So unless you’re worried that it’s going to take you 30 years to get the tax, I’m not saying that’s rational, then okay, maybe you should be investing in something else. So it’s a shared issue, right? Like the LPs need to make decisions to invest in people that they believe will return in an appropriate amount of time and appropriate, the longer it takes, the larger the multiple should be. That’s why seed is a place where you can make big multiples, but it’s going to take a decade plus. And as long as you’re willing to wait that decade plus, do I want to run my own business out of my pocket for 10 years before I see carry? Absolutely not. Unless you want to give me a lot more carry. And then I’m happy to do it. I will invest all my own money down to zero dollars of net worth left to have a zero and 50 fund. I mean, when I raised my fund, I committed in my own head, but on paper to the first fund to be a 10% GP commit in my own head to three. That was over 50% of my liquid net worth. I’m already in this thing extremely deep. And I’m more than happy to bet on myself. But that’s not the structure people are accustomed to. So, you know, why raise it a model that people don’t understand? I love that.

Brian Bell (00:24:43): Yeah. I’ve just never heard anybody say it quite as succinctly and clearly as that. You’ve been through lots of different trends and cycles and, you know, obviously AI, this, AI, that. How does this feel? Does it feel like a history rhyming right now with other tech booms or is it?

Ben Narasin (00:24:55): Sure. No, I mean, I built a business in the tech bubble of the web. 93 to 2003-ish. A lot of similarities. Now, mobile, not as big. I mean, big, but not as big and world changing as this. The same thing with Web3, which I never bought into as a concept. But it’s interesting. One of my founders called me and he’s like, man, so exciting. You know, I just got this call from this customer and he’s literally willing to pay us anything to get into our business right now because he’s too small. So we said no and he’s willing to. And I said, look, I’ve seen this movie before. We’re in the stage where this is exactly like it was in the web. And people are now starting to talk about already the same sort of things that they did back then. First, nobody cares. Nobody understands it. Nerds only. Geek, geek, geek. Then, this stuff’s interesting. I’m seeing these stories. Inter what? Huh, I’ve got to check this thing out. Then it’s like, oh, the web is coming. Okay, well, nobody ever buying it. And then like, oh my God, this thing’s everywhere. We aren’t there. And then management says, you need to get us in because this is an existential threat and we need to be in front of it. We are going to die if you don’t cover our butts in this category. And so they run out and they spend any amount of money, crazy numbers. They don’t know what it’s worth. They don’t know what it’s worth, so they pay whatever people ask. And then, and this is where we’re getting, year, year and a half, two years later, show me the money. Show me what I got. I gave you all this money. Show me the value. So it goes from ignorance to interest to excitement to panic to logical thought about what the value was. We are moving from the panic to the logical value for the people that bought. For the people that haven’t bought yet, they’re probably still in panic mode and will overpay. Now, there will be a shakeout, like there always is. And the companies that prosper will be outrageously valuable. And the ones that die will, of course, be dead.

Brian Bell (00:26:36): Yeah, really well put. $50 million fund. Tell us about your check size, portfolio construction, all that stuff.

Ben Narasin (00:26:42): Yeah, so we ended up with 60, one to $3 million checks. We tend to lead. I’m not, it’s funny, portfolio construction is an interesting topic. I was talking to an LP, one of the smarter ones, and he asked about portfolio construction, not for this fund, for a different fund I was raising. Yeah, we talked about it for a while. And I said, by the way, have you ever seen a fund that was perfectly constructed by classic portfolio construction theory? And he said, I did. We invested. How did it do? Worst fund I have ever backed. So I was like, and yet you still ask. So at the end of the day, my portfolio construction is awesome only. I want founders that make me say, wow, I want huge markets and great ideas. And I want to believe that I can return the entire fund with any investment I make. And I That’s pretty much it.

Brian Bell (00:27:31): So you’re not sitting there saying, I need X percent ownership, X number of companies, this many breakouts, this many zeros.

Ben Narasin (00:27:39): No. I mean, I think about ownership, obviously, but I don’t start with a model and then go try to fit companies into it. I start with companies that I think can be great and then figure out how to make the math work. If I like something enough, I’ll find a way to get into it. If I don’t, no amount of model is going to convince me.

Brian Bell (00:27:55): Yeah, that resonates. What about follow-on? Do you reserve for follow-on or is it mostly first checks?

Ben Narasin (00:28:02): I reserve, but I’m not dogmatic about it. If something is working really well, I’ll lean in. If it’s not, I won’t throw good money after bad just because a spreadsheet told me to. Again, it comes back to judgment. I think people hide behind models sometimes instead of making hard decisions.

Brian Bell (00:28:17): Makes sense. What do you look for in founders at this stage? You mentioned people three times earlier. What does that actually mean in practice?

Ben Narasin (00:28:25): It means a few things. One, intellectual honesty. If I ask you a question and you don’t know the answer, just say you don’t know. Two, self-awareness. Do you understand what you’re good at and what you’re not? Three, resilience. This is going to be brutally hard. If you haven’t been punched in the face before, it’s going to be tough. And four, the ability to recruit. Because no matter how good you are, you can’t build a big company alone.

Brian Bell (00:28:46): Yeah, recruiting is such an underrated one.

Ben Narasin (00:28:49): It’s everything. The best founders are talent magnets. People want to work with them. And that’s how you scale.

Brian Bell (00:28:55): Do you think that’s learnable or is that more innate?

Ben Narasin (00:28:59): I think parts of it are learnable. You can get better at communication, storytelling, clarity. But some of it is just who you are. People can tell.

Brian Bell (00:29:08): Yeah. As we kind of wrap here, what’s one thing you think founders consistently misunderstand about venture capital today?

Ben Narasin (00:29:16): That it’s about them. It’s not. It’s about returns. And that sounds harsh, but it’s reality. We are fiduciaries. We have to return capital to our LPs. So when a VC passes, it’s not necessarily a judgment on you as a person or even your business. It might just not fit what they need to do.

Brian Bell (00:29:32): That’s a really important distinction.

Ben Narasin (00:29:34): Yeah. And the flip side is also true. When they say yes, it’s not because they love you. It’s because they think they can make money. Now, ideally, both are true. But don’t confuse the two.

Brian Bell (00:29:44): That’s a great place to end. Ben, thanks so much for coming on.

Ben Narasin (00:29:47): Thanks for having me. This was fun.

None of those events sit in isolation. Stitched together, they describe what the AI economy looks like with the froth cleared off. A small number of companies with scarce capability, locked-in compute, and real commercial pull are sitting at the center of everything else, and the gap between them and the next tier is widening faster than any underwriting model from 2024 anticipated.

Here is what moved, and what it probably means for the rest of the year.

Read more

I said 150 over the fund’s three-year deployment window.

He paused, then, polite but pointed, said the thing I have now heard dozens of times. “Isn’t that spray and pray?”

That conversation is why this piece exists. And it is why we spent the last few weeks building an interactive tool that lets anyone test the math themselves. The tool is free. You can open it on your phone (although it is better to use on desktop), drag a slider, and see the argument we have been making to LPs for the last three funds, play out in real time.

Here is the argument in one sentence. Venture outcomes follow an extreme power law, and power laws reward more at-bats with better-quality returns, not fewer at-bats with bigger bets.

That is not a slogan. It is a consequence of how the returns distribute. The shape of the distribution is not a matter of opinion. It comes from the largest publicly available dataset of venture outcomes (Correlation Ventures, 21,000 financings), from Horsley Bridge’s empirical analysis of what separates top-tier VC funds from the rest, and from the Kelly criterion, which is the math gamblers and quantitative investors use to decide how much to bet when the odds are asymmetric. Three independent frames. They all point in the same direction.

Let me show you what the math looks like. Then I will explain why more companies produces more value per dollar, not less.

The quality of returns lifts with volume

Imagine you are running a seed fund with industry-average luck. Fifty-five percent of your companies go to zero. Fifteen percent give you your money back. Another fifteen percent produce a modest 3x. Eight percent hit a solid 10x. Two percent are unicorns, half a percent become decacorns, and one in a thousand becomes something like Stripe or OpenAI, which is a 5,000x (or a lot more) outcome.

Most seed funds hold around 30 companies. At that size, with reserves deployed sensibly into the winners, the math says about 21 percent of funds clear a 5x return. The median outcome is a 2.8x gross MOIC. After fees, that is a fund most LPs would politely call “fine.”

Now invest in 150 companies with the same luck and the same per-company budget.

The probability of clearing 5x climbs from 21 percent to 58 percent. Median MOIC nearly doubles, from 2.80x to 5.55x. The range of likely fund outcomes also tightens dramatically. A 30-position fund can land almost anywhere, from near-zero to a home run, and most LPs have no way to know in advance which version they’re getting. A 150-position fund still contains the same unicorns and decacorns, but individual funds cluster much more tightly around their average return. You lose the extreme downside and a little bit of the extreme upside, and you gain the thing LPs actually want from a portfolio: a return profile they can plan around.

Nothing in the outcome distribution changed. We did not improve at picking. We did not forecast winners. The quality of returns got better because we took more shots at a distribution where almost all the value lives in the tail. At 30 positions, you need to catch a unicorn to look good. At 150, you probably catch one or two unicorns just from the math, and the fund’s upside is less dependent on getting any single bet right.

The intuition is the one every poker player already has. If one suit in the deck pays a million and everything else pays nothing, and the dealer is about to turn over cards, you would rather see more of the deck. The concentrated fund is making a smaller number of draws from a distribution where the payoff is wildly asymmetric. Sometimes it works. More often, the unicorn doesn’t show up in those 30 cards.

Same math, two strategies

The chart below is the inside of the tool. The two panels show 30 positions and 150 positions running against the same simulated outcome distribution with the same random seed, so the draws are mathematically identical. The only thing that differs is how many positions the fund owns.

Read the numbers in pairs. The concentrated side has higher variance, which is the technical way of saying the distribution of possible fund outcomes is wide. Sometimes a 30-position fund catches a unicorn and looks like a genius. More often it doesn’t, and the whole fund underperforms. The distributed side compresses that variance. Fewer funds land in the tails at all, and more of them land somewhere the LP can live with. The probability of a 3x fund rises from 45 percent to 69 percent. The probability of a 5x fund nearly triples.

If you are thinking “the concentrated fund’s upside is still real, so maybe it’s worth the variance,” that is the honest version of the counter-argument. A concentrated fund in 1999 that caught Google returned every dollar you put in several hundred times over. The trouble is that you are not the person who caught Google, and you don’t know who is. The 30-position fund that doesn’t catch Google is the typical outcome, not the exceptional one. For an average GP picking 30 companies, the distribution of fund outcomes is dominated by the versions where the unicorn never comes.

Why this isn’t obvious

If the math is this clear, the reasonable question is why every LP isn’t already buying it. The answer is that they already are. They just take it through more expensive wrappers.

A $500 million commitment to a fund of funds lands in 20 to 40 underlying managers, each of whom is running a portfolio of 30 to 80 companies. The endowment, pension, or sovereign writing that check ends up with effective exposure across 800 to 3,000 startups. Mega funds like Sequoia Capital Global Growth and Tiger Global hold hundreds of positions across their strategies for the same reason. The biggest pools of institutional capital have already decided, with their allocations, that venture exposure wants to be held at hundreds or thousands of underlying positions. That is a settled practice at scale.

What is newer is that a single fund with the right deal flow can produce position counts that rival a fund of funds, without the double layer of fees. That requires being able to see, evaluate, and invest in enough high-quality companies per year to make the math work. Which brings us to the part that took us five years to build.

The binding constraint

Here is what the simulator cannot show you. A 150-position portfolio requires access to 150 investable companies in a three-year window. Most seed funds can’t reach that threshold at quality. Their deal flow runs out. Most concentrated funds are not concentrated by choice. They are concentrated by necessity.

We are not. Y Combinator produces roughly 800 pre-vetted companies per year across four batches. That is the raw input. But raw input isn’t enough on its own. You also need the relationships to get in at the terms you want, the intelligence to tell the signal from the noise, and the operating muscle to support 150 founders without becoming the investor who doesn’t pick up the phone. Volume without the rest of the stack is the spray and pray caricature. Volume with the stack is something else entirely.

The stack is where this gets interesting, and it is the part that compounds.

Every investment we make adds to the network. Every founder who works with us can become a scout, a reference, a customer for another portfolio company, or a future LP. Every VC we co-invest with becomes a co-invest partner for the next deal. Every conversation on the Ignite Podcast becomes a relationship with a founder, VC, or operator who now knows us. Every blog post brings in another intelligent reader who wants to understand private markets better.

Said out loud, that sounds like marketing. On the ground it is a set of very specific loops. More companies means more referrals. More referrals means better deal flow. Better deal flow means better companies invested in. Better companies means better outcomes. Better outcomes means more founders want to work with you. More founders want to work with you means more referrals. Run that loop for half a decade and the deal flow problem that constrains most seed funds stops being the binding constraint.

The same loop runs through the other sides of the business. More than three thousand VC relationships means we can route a founder toward the right Series A lead in the right week. Fifteen thousand community members and counting means when a founder needs ten customer conversations, they get ten. Half a million monthly viewers across social means a portfolio company launch moves when we push it. None of this is free. But the marginal cost of adding one more position, given that the network is already in place, is almost nothing. And the marginal value of that position, both for the company and the network, is real.

That is how volume becomes a strategy instead of a prayer.

The honest caveats

If I only tell you the good part, I have failed you. The math has real limitations and it is worth being explicit about them.

First, the simulation assumes the outcome distribution holds across all 150 positions. In reality, if you invest in every YC batch without discrimination, you are also catching the bottom of the batch. Selection still matters. The simulator has a skill slider that adjusts for this, and at every level of skill the volume argument still holds, but the magnitudes change. A below-average picker investing in 150 companies still does meaningfully better than the same picker investing in 30. A great picker investing in 150 does materially better than a great picker investing in 30. Skill and volume compound.

Second, the follow-on math in the tool assumes threshold-based reserve deployment rather than conviction-based. Real GPs make judgment calls on which winners deserve more capital. The simulator handles this by scaling follow-on outcomes with the skill parameter, but the abstraction is imperfect. This is one of the places I genuinely want feedback. If you have seen reserve deployment fail for a reason the model doesn’t capture, tell me.

Third, the three frameworks in the tool (Horsley Bridge, Kelly, and Monte Carlo) are not independent. They share distributional inputs. Change the assumption about how often a unicorn appears and all three move together. The tool’s Overview tab shows an expected-value breakdown that makes this explicit. Look at it. The rare outcomes do most of the work. If you disagree with the rare-outcome probabilities, your disagreement moves the whole result. The tool lets you edit the distribution yourself.

Fourth, and most importantly, this is our strategy. It is not a universal prescription. A great investor running Benchmark-style conviction at 10 positions can beat us if their picking is good enough and their access is deep enough. Historical concentrated funds that worked produced extraordinary returns. The math in this tool is the defense of our particular approach, not a claim that everyone else is wrong.

Fifth, this is version one. The numbers in this piece come from a specific set of defaults. If you change the defaults, you get different numbers. Some of those differences will challenge my framing. Good. That is the point.

Try it yourself

The tool is free and takes no login. You can move a slider from 30 to 500 positions and watch what happens. You can flip between the industry-average distribution and the YC-historical distribution, which has a higher hit rate and tighter outcomes. You can edit the underlying probabilities if you think our assumptions are too generous or too conservative. You can change the skill parameter. You can compare concentrated and distributed strategies head to head with the same random seed so the comparison is fair.

If you are an LP thinking about venture allocation, I would particularly encourage you to play with the head-to-head tab. Pick a GP pitch you have heard recently, estimate the position count they are running, and see what the math says the distribution of outcomes looks like for their strategy versus ours. You do not have to believe our numbers. Edit them until you believe them. The shape of the answer will probably surprise you.

This is version one. We shipped it fast so we could get feedback early. If something is confusing, wrong, or missing, please say so. I read every response.

We run this strategy because we believe the math. We built the tool because we think most LPs deserve to see the math for themselves, without having to take any GP’s word for it, including ours. Venture capital has too many people telling you what to believe and not enough people showing you why.

Here is the math. Here is the code. Run it. Push back. If you come to a different conclusion, I want to hear what it is.

If you want to go deeper on what we do, our value proposition for LPs lays out the wider platform, from quarterly top-ten watch lists to the operator network we deploy on behalf of portfolio companies. If you want to see the companies we have backed, check out our portfolio here. If you want to listen to founders and fund managers describe how all of this plays out in practice, the Ignite Podcast is nearly three hundred-plus episodes deep at this point.

More than anything, the best thing you can do is open the tool and spend five minutes with it. The numbers make the argument better than I can.

Try it here: https://tools.teamignite.ventures/

Sources and further reading

The probability distributions in the tool are calibrated from published research. Every claim in the piece and the simulator footer maps to one of these:

Outcome distributions

• Correlation Ventures’ 21,000-financings study (covering 2004-2013), originally published by Seth Levine at sethlevine.com, with a 2019 update here. The source of the “65% of financings return below 1x” and “4% return 10x or more” figures.

• Carta’s Class of 2018 cohort data on seed-funded startup outcomes, summarized by SaaStr here. The source of the “30-35% of seed-funded startups shut down within 7 years” and “1.3% become unicorns” figures.

Outlier rates by fund tier

• Horsley Bridge outlier-rate data via Ulu Ventures, at uluventures.com/picking-winners-is-a-myth and their longer Portfolio Construction essay. The source of the “top-tier VC 4.5% outlier rate” and “average VC 2% outlier rate” figures. Cambridge Associates’ 2.5% industry average is cited inside both of these pieces.

• The “hypothetical superstar” tier in the tool (7% outlier rate) is Ulu Ventures’ construct representing “50% better than top-tier,” used to stress-test concentrated-portfolio claims. It is not an empirical observation and is labeled as such in the tool.

YC-specific numbers

• YC unicorn rate, Garry Tan on X, August 2025: x.com/garrytan/status/1953069914132238775. Tan’s stated range is 6% to 12% for the last ten years of batches. Our default of 6% sits at the low end.

• PitchBook’s 2023 analysis of YC cohorts, via Yahoo Finance. The source of the “4.5% of YC startups since 2010 become unicorns, 5.4% for 2010-2015 cohorts” figures.

• YC batch size, via Hustle Commons’ profile of Garry Tan. The source of “400+ per batch, two batches per year, ~800 per year.”

Fund economics

• J-curve pacing for seed-vintage funds: Moonfare’s overview and Pipeline Road’s benchmarks.

• 2-and-20 (2% annual management fee, 20% carry on profits above 1x) is an industry convention, not a single-source citation.

Team Ignite platform

• Our value proposition for LPs, which describes the platform beyond the investment thesis.

• The Fund III deck, with the portfolio and independent TVPI benchmarks.

• The Ignite Podcast, now past 250 episodes.

FAQ: the pushback I have already gotten

Before I published this piece I sent it to a friend who runs another VC firm and has a computational CS background. He was skeptical. Specifically, he pushed back on the math. He pushed back hard on every assumption in the simulator, and some of his pushback was right. What follows is a set of questions that came up in that conversation and in early LP reviews, with honest answers rather than defensive ones.

Aren’t you treating each investment as independent when they clearly aren’t?

Yes. That is a real limitation of the model and it deserves a direct answer.

A Monte Carlo simulation assumes each draw is independent and identically distributed, meaning each company in the portfolio is treated as a fresh roll from the same distribution. In reality, the 150th investment in a fund can differ from the 1st. For a fund with a fixed deal pool and declining access, the 150th pick is worse than the 1st. For a fund like ours, drawing from YC and Speedrun batches that refresh every quarter with a consistent quality bar, the pool is replenished continuously. The 150th pick across a three-year deployment is drawn from a universe of roughly 2,400 pre-filtered companies, not from the same batch with the best ones already taken. Whether the distribution degrades, stays flat, or improves with position count is a function of the specific fund’s sourcing model. The simulator assumes it stays flat, which is closer to our reality than the degrading-distribution case. Not to mention our deal flow leads to more deal flow.

The simulator is a thought experiment that says, “if position count were the only variable that changed, what would the math predict?” It’s useful for building intuition about why power-law distributions reward more at-bats. It is not a fund model. A real fund has correlated outcomes across sectors and vintages, a selection function that interacts with portfolio size, and network dynamics that evolve over time. The tool captures none of those.

What it does capture is directionally correct for a picker with stable deal flow: more positions produce better expected outcomes in a power-law world. What it cannot show is the exact tradeoff between position count and per-position quality, which is the question every real GP has to answer with judgment rather than math.

Isn’t a smaller portfolio better because it lets you focus on the best companies?

Only if you can reliably tell which companies will be the best in advance. The research on selection says most GPs cannot.

Consider the arithmetic. A 20-position fund investing out of a YC batch of 200 is making an implicit bet that its 20 picks will include the batch’s outliers. If its selection is no better than random, the expected number of outliers captured is one-tenth of the batch’s total outliers. A 150-position fund investing out of the same batch captures three-quarters of whatever outliers that batch produces, regardless of picking ability. For the concentrated fund’s smaller portfolio to be the better choice, the GP has to be confident their picking meaningfully beats the batch’s own filter, which is already one of the most selective in venture (roughly a 1% acceptance rate).

That is a higher bar than most GPs can clear. Including, to be clear, us. We do not claim to out-pick the YC admissions process. We claim to run a disciplined investment framework on top of an already-selected population, and to capture more of that population’s tail than a concentrated fund can. The concentration bet and the volume bet make different assumptions about selection ability. Volume is the honest choice for a GP who does not believe they have a reliable crystal ball.

Where does your outcome distribution come from, and why do you assume it applies to every position in a 150-company portfolio?

The distributions come from published datasets: Correlation Ventures’ 21,000-financings study, Horsley Bridge’s published outlier rates, Cambridge Associates’ industry averages, and Carta’s Class of 2018 cohort data. Sources are cited in the piece and in the tool footer.

The fair critique is that those distributions are averages across the whole venture industry, and no single fund’s outcomes match the industry average. A top-tier fund has a better distribution. A bottom-tier fund has a worse one. Applying a single distribution uniformly to all 150 positions assumes you can hold quality constant as you scale, which is a defensible assumption for some sourcing models and not others.

That’s why the tool has a skill slider. Move it up and the distribution shifts toward better outcomes. Move it down and it shifts the other way. The volume argument holds at every skill level, but the magnitudes change. A below-average picker still does better with 150 positions than with 30. A top-tier picker does even better with 150 than with 30. Skill and volume compound. Neither replaces the other.

What the tool doesn’t do in v1, and what I’m adding to v2, is let users configure how the distribution changes across position count. The current assumption of “stays flat” is our view for a YC-focused fund with a compounding network. Others will disagree, and the tool should let them test their own assumptions.

How can diversification reduce your risk if sector risk and vintage risk don’t go away?

It can’t. And it doesn’t.

Diversification across 150 positions reduces idiosyncratic risk, which is the risk tied to any specific company. It does nothing about systematic risk. If the entire AI category collapses, or interest rates wreck the IPO market, or a macro event shuts down exits for three years, a 150-position portfolio suffers the same way a 30-position portfolio does.

The claim in the piece is narrower than “diversification reduces risk.” The claim is that for the specific risk of a fund underperforming its target return because none of its companies became outliers, more positions reduce that specific risk. That is a real benefit. It is not a claim that all risks are reduced.

An LP considering this strategy still has to think about sector concentration within the 150 (we’re heavily AI and B2B SaaS, so that is real risk), vintage concentration (fund II and fund III are deploying into overlapping market conditions), and macro exposure (we can’t diversify away a 2008 or a 2022). The volume argument is specifically about power-law capture, not a general claim of risk reduction.

Standard deviation of fund outcomes isn’t observable the way it is in public markets. Isn’t the “spread tightens” claim fake?

Partially yes. And this is a good catch.

Public companies are marked to market every day, so standard deviation of returns is a measured quantity. Private companies are valued quarterly at best, usually lag the real value by a year or more, and only resolve to actual dollars on exit. So “standard deviation of fund MOIC” isn’t a number you can point at in reality. It’s a number that falls out of a simulation.

What the simulator shows is that within the model, simulated fund outcomes cluster more tightly as position count grows. That statement is mathematically true given the independence assumption. Whether it translates to real-world funds is a separate question.

The better way to read the tightening is as “distribution of simulated outcomes,” not “distribution of real fund outcomes.” Real funds have additional sources of variance the simulator doesn’t model. So the absolute number (13x standard deviation at 150 positions) is not a prediction for our actual fund. It’s a statement about what the math does when you hold everything else constant.

Doesn’t the efficient frontier say there’s no free lunch? How can you lower risk without lowering return?

You can’t, and the piece doesn’t claim you can.

Look at the head-to-head comparison carefully. The 150-position portfolio has a lower probability of extreme upside than the 30-position portfolio does. The concentrated fund has a wider distribution on both sides. Sometimes it produces a 20x outcome that the distributed fund might not reach, because the distributed fund is pulling the mean down by averaging across many positions.

What the distributed fund trades is the tail risk on both ends. It gives up some ceiling in exchange for a higher floor and a higher median. For an LP trying to plan a portfolio around a predictable target return, that trade is attractive. For an LP explicitly seeking maximum upside and willing to accept higher variance, it isn’t. The volume strategy is not Pareto-superior to concentration. It is optimized for a different objective function.

The efficient frontier still holds. We are picking a different point on it than a concentrated fund picks, and we think that point is the right one for most LPs. That is a judgment call, not a math proof.

If the math says more positions is better, why doesn’t the perfect investor just invest in everything?

They don’t, and that’s the right intuition pump to stress-test the model.

A perfect investor with oracle-level foresight would invest in 3 companies: the three centa-corns in the next decade. Their portfolio would return thousands of x. That’s the theoretical optimum, and a 150-position strategy would have trouble beating it.

The reason the volume strategy makes sense is that nobody has oracle-level foresight. Among real investors, the spectrum runs from “well above average” to “well below average,” and for the vast majority of the distribution, adding positions improves expected outcomes because the cost of missing a winner is higher than the cost of adding a loser.

The implication is not “volume is always right.” The implication is “volume is right unless you have genuine oracle-level insight, in which case concentrate.” We think most investors, including most professional GPs, are closer to average than oracle. We count ourselves in that group, not above it.

Peter Thiel says concentration wins. Aren’t you arguing against him?

Thiel made Facebook returns on a $500,000 check because he had genuine contrarian insight and the access to act on it. That worked. I’m not arguing it didn’t.

The argument is that Thiel’s strategy is right for Thiel and wrong for almost everyone else. The counterexample to his framing is the hundreds of concentrated-conviction funds that didn’t catch Facebook and underperformed as a result. We only remember the ones that got the home run. Survivorship bias makes concentrated strategies look better than they have performed in aggregate.

If you genuinely believe you are Peter Thiel, concentrate. If you believe you are a disciplined, well-networked, hard-working, thoughtful investor who is not literally Peter Thiel, the math says spread out a bit.

The tool compares the same distribution at different portfolio sizes. What happens if I change the distribution as I scale?

Fair question, and it’s the one I owe the most detailed answer to. The current v1 tool holds the outcome distribution constant as position count grows. That assumption is closer to our reality than to most funds’ reality, for reasons explained in the first answer. But reasonable people disagree about it, and the tool should let them test their own view.

A user could model distribution-degrading scale by splitting the portfolio into tiers. For example: positions 1-30 draw from the YC-calibrated distribution (6% unicorn rate), positions 31-90 draw from a slightly worse distribution (4%), positions 91-150 draw from a worse one still (2%). That would model a fund whose deal flow quality declines as it extends its reach.

I’m adding that configurability to v2 as a user-adjustable toggle. The user will be able to set whether the distribution improves, stays flat, or degrades with position count, and see how the result changes. Our own view, laid out in the companion piece on the network flywheel, is that for a YC-focused fund with an active network, the distribution stays flat or improves slightly over the deployment window. We think that because our sourcing draws from a continuously replenished batch system (~800 YC + ~200 Speedrun per year), because our network compounds with each portfolio addition, and because our selection tooling has gotten meaningfully better with scale, not worse. But I want the LP to test their own assumptions, not mine.

If you want to stress-test this in v1, the skill slider is the closest proxy. Drop the skill to below-average and see what portfolio size maximizes expected return. That’s a rough approximation of what happens when deal flow quality degrades.

What is this tool for, then?

It is a pedagogical tool, not a fund model. Its job is to show why power-law distributions make volume-based strategies mathematically coherent, not to predict the outcomes of our fund or anyone else’s.

If you are an LP trying to decide whether our strategy is defensible, the tool is useful. If you are trying to figure out whether we will specifically hit a 5x MOIC on fund III, the tool tells you nothing you can use. The real question for an LP is whether our deal flow, our selection, our network, and our operating stack are good enough to capture the volume strategy’s theoretical upside in practice. That is an empirical question about our fund, not a question the simulator can answer.

The reason to publish the tool anyway is that most LPs have never seen the underlying math laid out in a form they can poke at. The argument for a volume strategy is usually hand-waved with phrases like “more shots on goal” and “power law,” which is not persuasive. Showing the math, and explicitly flagging where it breaks down, gives LPs something to push back on. The conversations I’ve had since publishing it are better than the conversations I was having before. That was the goal.

So is the strategy defensible or not?

The strategy is defensible. The specific numbers in the simulator are directionally useful and depend on assumptions a reasonable LP might or might not share. Those are two different statements and both are true.

A fund running 150 positions with average deal flow and average selection will in expectation beat a fund running 30 positions with the same deal flow and selection. That claim holds up under every reasonable modification of the model I have tested, including versions that let the distribution degrade modestly with scale. What shifts with better modeling is the magnitude of the advantage and the optimal position count. What doesn’t shift is the direction.

If you think we have below-average deal flow and selection, the strategy still works but the numbers are worse. If you think we have above-average deal flow and selection, the strategy still works and the numbers are better. If you think you have top-tier deal flow and selection yourself, concentrate, and you disagree with your own assessment of your ability to tell the difference in advance.

The tool is one piece of evidence. The rest of the evidence is our track record, our network, our operating platform, and the specific portfolio companies we have backed. Any LP evaluating us should weigh the math alongside those, not instead of them.

That assumption gets expensive.

In this conversation with Ryan Ridgway, founder and CEO of Cirrus Capital Partners, the focus shifts from “how to raise” to “what kind of capital actually fits your business.” It’s a practical breakdown of how smart founders think about funding once they move beyond the early stages.

Subscribe now

The Funding Gap Most Founders Don’t See

There’s a segment of companies that doesn’t get enough attention.

Too big for small startup loans.
Too early for banks or private equity.
Still growing fast, but not yet profitable.

Ryan calls this the “awkward middle.” It’s where companies typically need $2M to $50M to scale, but don’t have clean access to traditional capital sources.

Most founders in this stage default to raising more equity. Not because it’s the best option, but because it’s the only one they know.

That’s the gap Cirrus Capital is built to solve.

Debt vs Equity: Use Case Matters

One of the clearest takeaways from the episode is simple:

Match the type of capital to how you plan to use it.

If you’re funding experimentation, product development, or anything with high uncertainty, equity makes sense. Investors take risk in exchange for upside.

If you’re funding predictable parts of the business, debt starts to make more sense.

Ryan breaks it down into two buckets:

Growth capital
Used for hiring, marketing, and expansion. It fuels scale.

Working capital
Used to manage timing gaps in the business. For example, paying suppliers today while waiting 60–90 days to get paid by customers.

Many founders use equity for both. That works, but it can quietly dilute the business more than necessary.

Why Founders Miss Debt Opportunities

It’s not a knowledge problem. It’s exposure.

Most founders and early investors are trained to think in equity terms. So the conversation rarely expands into credit options.

Ryan sees this often. Founders simply don’t realize what’s available to them on the debt side.

That leads to a pattern:

• Raise equity earlier than needed

• Give up more ownership than planned

• Repeat the cycle in the next round

The better approach is to look at both sides of the capital stack early.

Combining Debt and Equity

This is where things get interesting.

Instead of choosing one, strong companies use both.

Equity provides flexibility. Debt adds efficiency.

When done right, they reinforce each other:

• Equity improves your balance sheet and credibility

• Debt extends your runway without dilution

• Together, they give you more control over timing and growth

Ryan describes this as a “flywheel effect.” Each side makes the other more effective.

But it only works if the structure is intentional.

What Lenders Actually Care About

Equity investors can afford to be wrong. Credit investors can’t.

They expect to be paid back every time.

So they look for signals that repayment is realistic:

• Revenue stability or clear path to it

• Assets (inventory, receivables, contracts, or IP)

• Cash runway and burn rate

• Evidence the business is moving toward profitability

If those don’t exist, debt becomes harder to access.

That’s why very early-stage startups rely almost entirely on equity.

The Risks Founders Overlook

Debt is powerful, but it comes with tradeoffs.

A few that founders often underestimate:

Personal guarantees
Some loans require you to back them personally. That can improve terms, but increases risk.

Timing mismatches
If your plan changes but your debt structure doesn’t, you can get stuck.

Overestimating stability
Debt assumes some level of predictability. If your business isn’t there yet, it can create pressure fast.

None of these are deal breakers. But they need to be understood upfront.

What’s Changing in the Market

Two big shifts are happening right now.

1. Debt is moving earlier
Companies that previously wouldn’t qualify are now getting access to credit, especially with more creative deal structures.

2. Underwriting is becoming automated
Lenders are pulling real-time data from tools like QuickBooks, Stripe, and Shopify to make faster decisions.

That reduces friction and opens access to smaller companies.

In the next few years, expect more deals in the $1M–$5M range to be processed with minimal human involvement.

A Better Way to Think About Capital

Ryan’s advice is straightforward.

Start with the end goal.

Do you want to build and sell? Hold long term? Partner with private equity?

Your answer should shape your capital strategy from day one.

Without that clarity, it’s easy to raise money that pulls you in the wrong direction.

Where to Start

If you’re a founder thinking about capital today:

1. Get your financials clean and reliable

2. Be clear on what you need capital for

3. Explore both equity and debt options early

4. Design your capital stack, don’t default into it

Most founders spend more time chasing capital than structuring it.

The ones who do both well tend to keep more ownership, move faster, and stay in control longer.

That’s the real advantage.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:
00:01 Introduction and Ryan Ridgway Background
00:31 Ryan’s Origin Story and Early Entrepreneurship
03:29 Founding Cirrus Capital Partners
05:54 The “Awkward Middle” in Startup Funding
06:56 Ideal Company Profile and Revenue Thresholds
07:30 Growth Capital vs Working Capital Explained
10:02 How Cirrus Structures Capital Solutions
11:28 Why Early-Stage Companies Struggle with Traditional Banks
12:11 Common Misconceptions About Debt vs Equity
14:29 Combining Debt and Equity for Better Outcomes
15:26 Matching Founders with the Right Capital Structure
18:49 Negotiating Term Sheets and Lender Dynamics
19:07 How Founders Can Assess If They’re Ready for Debt
21:31 What Disqualifies a Company from Credit
23:45 Creative Financing and Distressed Scenarios
24:08 Risks Founders Overlook with Debt
26:26 Market Trends in Credit and Venture Debt
28:44 Innovation in Growth Debt and Deal Structuring
31:04 Automation and Data-Driven Underwriting
32:00 AI Trends Across Startups and Lending

Transcript

Ryan Ridgway (00:00:00): founders at the end of the day, it’s tough. You’re focused on 10 million different things, 10 million different hats, and it’s excruciating and it’s exhausting. So it is important whether we want to or not to really step away and take a few deep breaths and try and focus on where we want to go and be, it’s almost like an Aussie moron, right? Be deliberate about how you’re going to get there, but be relaxed about it at the same time. Give yourself enough rethought. Thank you.

Brian Bell (00:00:48): Hey, everyone. Welcome back to the Ignite podcast. Today, we’re thrilled to have Ryan Ridgeway on the mic. He is the founder and CEO of Cirrus Capital Partners, a fintech entrepreneur, angel investor, and a global speaker known for helping founders unlock creative, non-dilutive capital solutions and scale high growth ventures. Ryan combines a founder’s mindset with deep capital markets, expertise to help companies raise smarter funding and build long-term value. Thanks for coming on, Ryan. Appreciate you having me. Yeah. So I’d love to start with your origin story. What’s your background?

Ryan Ridgway (00:01:15): Yeah, my origin story is probably a bit differentiated than most, at least most kind of in my respective space of, you know, private credit or lending. Grew up in the Midwest, did not come from an entrepreneurial and or high net worth, you know, kind of households, pretty grassroots and, you know, working class family and realized fairly early on as I was going to school at KU that I wanted to kind of deviate from the normal path. And kind of by happenstance, became an entrepreneur. I will admittedly say that my first glimpse into entrepreneurship was with a direct selling or multi-level type of organization. And everyone has their opinions on those and what they might like or dislike. But for me, that was kind of my true taste that there’s other things that are different out there apart from just pursuing my degree and going into accounting or communications or what have you. And so as many of those spaces tend to do, you tend to kind of shuffle out of that eventually, but carry with you some entrepreneurial traits. And, you know, as I was piecing together my career, I knew I liked two things quite a bit. One was entrepreneurship and the other was sales and business development. development and just talking to people and creating value in people’s lives and so worked a few corporate roles early on you know so think business development account exec vp type of roles but at that same time was um really trying to carve out my own path and so uh my my first i would say semi-successful company that led to an exit uh was within the uh health and nutrition category a supplement company. And then subsequently after that, I kind of entered my world of non-dilutive capital or lending by accident, just by proxy of some relationships we had built at the time. The gentleman came to me and said, hey, if you refer us business, we’re happy to pay you a commission for it. And I said, well, that sounds kind of cool. And then over time, what that developed into is, hey, you know, I can go out and kind of forge these relationships with other capital providers as well. And so anyways, I won’t belabor my kind of origin story, but it’s a bit unique versus most that kind of came up through a traditional, you know, Wall Street or FinServe background, but taught me a lot along the way and have a really good appreciation for entrepreneurship.

Brian Bell (00:03:48): I love that. So at what point did you decide to start Cirrus Capital Partners? What need were you trying to solve and that wasn’t being addressed in the market?

Ryan Ridgway (00:03:55): Yeah. So the previous business that I founded ahead of Cirrus Capital Partners was actually a small business lending marketplace. And so, you know, some of the listeners here might be familiar with the likes of like Lindio or Fundera. Those would be seen as a direct competitor to what we use to do previously. So think kind of small dollar commercial loan instruments, call it 50,000 upwards of maybe a million or so on average. As we built that over the course of about seven years, we became a direct lender in the process as well. So we had kind of two sides of that business. There was a lot of innovation. There were a lot of really great incumbents kind of stepping up in a meaningful way and creating technology or the rails that support underwriting and loan servicing and decisioning and all of these important processes involved in lending. Contrarily to that, if you look above market relative to where we’re focused today at Cirrus, and you look towards kind of middle markets upwards to enterprise level companies, they typically already have really great relationships in place with banks or other non-bank direct lenders or investment banks who can steward them through a good financing process. And overall, the laws of leverage kind of inverse in their favor, right? Where they kind of have the pick of the letter in terms of the best experience and and outcome. And so that’s kind of the metamorphosis behind Cirrus. We focus on the awkward middle category. So it’s with typically companies that are growing beyond the initial kind of usable shelf life of nascent stage credit products. They know they’re looking to scale up or they know they have, you know, working capital constraints that if they solve for them would give way by proxy of that to growth capital being opened up in the business as well. But at the same time, they’re not big enough or they don’t have, you know, interest in taking chips off the table, exploring private equity conversations or an IPO or anything quite yet. So oftentimes what we’re doing is, you know, we’re coming to the table with what’s typically for us like $2 to $50 million worth of credit to help support the next phase of their growth. And surprisingly, even in the year 2026, it’s been a growing segment, but we still feel that it’s a very kind of overlooked and underserved segment of the market where a lot of these founders and management teams need a lot of help and guidance kind of stewarding them throughout the process. So that was, yeah, we kind of came together to bring that to a close, you know, saw a lot of opportunity there in the marketplace to help folks.

Brian Bell (00:06:32): Yeah. Yeah. And I could see that. There’s 30,000 companies roughly that get pre-Series A funding every year. And I’m guessing you kind of start getting involved when they hit a, you know, what’s kind of the minimal ARR that you kind of look for to be involved?

Ryan Ridgway (00:06:45): So it doesn’t always even have to be recurrent revenue. Helps if it does. That’s wonderful. Having kind of sticky revenue. But where we end up being really meaningful is kind of a million or two per annum, or at least as a run rate, we can get in and really start to be impactful on the debt side.

Brian Bell (00:07:03): Yeah. And then you threw out some terms there that I understand, but there might be listeners that don’t know. What is he talking about working capital versus growth capital? Maybe you could define those and kind of where you get, where you kind of sit.

Ryan Ridgway (00:07:14): Yeah, absolutely. So I think it depends on the style of business. So most of the entrepreneurs that we support are who fall into the innovation economy or are building a SaaS or a company with recurrent revenue and a subscription model, most of the time they’re seeking growth capital because the way we would distinguish between those is growth capital is for general corporate purposes. It’s to invest into marketing. It’s to maybe hire some key business development-oriented hires and so forth. Working capital is more closely related to the respective trades in the business. And so if you have, let’s say, a consumer goods company that has their supply or manufacturing relationships on one side, they probably have terms in place where, let’s say they’re a ready to drink beverage company, for instance, right? So if they want to create the liquid inside and have it packaged and canned and fulfilled, they’re probably going to have to front a meaningful portion of that capital at the onset and at intervals thereof as it goes through the production and the shipping process and so forth. That’s before they’ve sold the products to their end customer base. Many of those companies, apart from their direct to consumer channels, also have B2B channels as well, which might mean they’re selling into a Wedman’s or a Whole Foods or a Target or whoever. And they have net terms with them as well that says, great, we’re going to get our product on your shelves. And we aren’t going to get paid for 30, 60, oftentimes 90 days. There’s these deltas in cash that exist between supply and manufacturing and in customer relationships that can actually add up in aggregate to maybe six or seven months. where they’re floating hundreds of thousands or millions of dollars. And if you think about the distinction between where equity comes into play and credit comes into play, if you’re using equity for the working capital purposes, it could become prohibitively expensive over time, especially when you’re thinking about equity investors wanting their kind of upside return for the risk they’re taking early on in the business. And so, yeah, anyways, not to talk too exhaustively about that, but that’s kind of how we think about growth versus working capital needs.

Brian Bell (00:09:33): And then you guys get more involved on the working capital side.

Ryan Ridgway (00:09:37): On both, frankly. Yeah. So like we have, I’ll give you an example. There’s a company we supported recently that they’re in the consumer goods segment. They’re also venture backed. And so when you look at the different ways that a debt or credit investor might get excited about the business, you have a strong balance sheet. They raised equity dollars recently. So they have a few million in the bank. They also have assets in the form of purchase orders, accounts receivable, inventory, And they have these varied use cases on how they can exhaust funds in the business. And so that setting is kind of a good example of, hey, here’s 5 million being brought to the table as an initial tranche. We’re actually going to use a piece of that for hiring and investing into marketing, we’re going to use another piece for more of the working capital elements and the trade finance relationships.

Brian Bell (00:10:27): So it’s kind of like almost like micro corporate capital is how you might describe it, right? Because like big Fortune 500s have the banking relationships, like you said, and they just call up like Bank of America or Wells Fargo or whatever and just get a multi $100 million credit line to do this kind of stuff. But there’s this long tail of tens of thousands of companies, if not hundreds of thousands of companies that have the same needs.

Ryan Ridgway (00:10:50): And that’s, by the way, that’s where they’ll end up as well. If they continue to grow and they do well, the best option for them is go work with a big bank who’s just going to open the checkbook to you and offer you very, very attractive rates. The reality early on is most of these companies are still loss making or on their way to profitability, they don’t fit that perfect narrative that a traditional bank might. Right.

Brian Bell (00:11:14): They don’t have the free cash flow and the EBITDA margins and debt service ratios that the big banks will look for. That’s right. In your view, what’s the biggest misconception founders have about taking on debt versus equity?

Ryan Ridgway (00:11:27): Yeah. I think what we come up against a lot is most founders don’t realize what might be available to them on the credit side. And it’s not their fault. It’s also not their investors’ fault as well, because oftentimes they’re self-governing the business and they also have a board that’s helping them govern the business and make decisions as well. And a lot of the times, the personnel at the table or within the management team is just more accustomed to raising equity. And so part of our exploratory process is what are you using the capital for? And mindful of the fact that you have your specific background. We’re probably going to have people that watch or listen to this that come from either equity or credit or not from the finance world at all. So I’ll try and remove the bias here, but... But our personal standpoint is make sure that the capital you’re raising corresponds with its use case appropriately. For us, that oftentimes means at the onset, it’s always almost always going to be equity level risk. Absent example would be like an EV company or battery storage company where to even stand up the venture, they might have 10, 20, $30 million of CapEx needs to stand up a floor and a facility and manufacturing and production needs, right? Which is totally doable on the credit side. But absent that, they probably still have a lot of equity level risk that they’re looking to take advantage of in the way of innovation and kind of R&D spend and other general use cases. So yeah, going back to your question, I mean, that’s always one thing we’re trying to make people aware of is that there are more capital solutions apart from simple kind of, you know, three Fs, friends and family, angel investors, you know, a safe note, and then getting kind of into the price note category. And actually, those two sides of the capital markets between equity and debt, I just hopped off a call right before this podcast, actually, and It was with a client that is raising both equity and credit. And we were talking about positive and negative leverage or these flywheel effects that can be created if you are rowing with both sides and raising both equity and credit, because typically the narrative becomes from the equity side, okay, where is credit coming from? And from the credit side, where is equity coming from? Especially if you’re dealing with a high growth, still loss making or burning business. And so anyways, that’s kind of our perspective.

Brian Bell (00:14:05): So how do you guys approach matching founders with the right kind of capital structure?

Ryan Ridgway (00:14:10): Yeah, every scenario is a little bit different. We meet founders wherever they’re at. So we’re accustomed to working with very robust and talented teams who might have an internal CFO or even like a head of capital markets already. And we look at an existing data room and it’s like, gosh, this is great. This has everything we could ever ask for and more. And in which case our process would be the stewarding of that data and opportunity into the capital markets. We have other scenarios where maybe a founding team is just trying to get their bearings on what might exist. We’ll go through the motion of requesting financials and a pitch deck and some high-level materials at the onset, but it becomes clear that they probably need a lot of help in the process, and we’re happy to roll up our sleeves and be supportive in the way of actually building a narrative for them, presenting the company in a really meaningful way and taking them all the way through a successful flow. So it kind of depends. Our process, though, almost always yields an end scenario that is faster more frictionless and more cost effective than than going alone and the way that we’ve built that over time is we have fairly rigid slas and processes in place with the various you know credit investors that we work with and so the understanding at the onset is hey if if you want access to tailored deal flow right like we we both need to be kind of playing our part here and you know we expect them to spawn respond very quickly. And so that’s in and of itself just is a lot better of an experience than what they might get on their own. For instance, Googling, filling out a bunch of contact forms on different websites, getting routed to some random person in the company where we can kind of take them straight to the front of the line, so to speak. And the other aspects, and I don’t think I’m saying anything that if lenders and our credit investor partners listen to, and I greatly appreciate them, that they would be surprised about, right, is we can apply a little bit of leverage in these conversations. And we work a lot with a lot of different types and themes of lenders. And so when it comes to a term sheet V1.0, that is almost never going to be the term sheet that’s just immediately looks great, everything’s perfect, let’s sign and continue. There’s always a little bit of back and forth and we can be there on behalf of our clients to continue. to tell them, hey, what’s market, what’s not market, maybe based upon that same relationship, where can a particular lender or lender in a certain theme, asset-based lending versus cash flow, et cetera, be flexible versus where they’re probably just going to be very rigid and need to document the loan or facility appropriately. And those can be very value additive that they can give it on top of just running a solo process.

Brian Bell (00:17:05): So there might be founders listening and they’re wondering, are they a good fit for this kind of debt structuring? What’s your kind of mental checklist that you could give founders to go through to kind of determine if they’re ready to take on this type of funding?

Ryan Ridgway (00:17:20): Yeah, I think at the onset, it’s a bit more art than science. It becomes more science as you continue to scale out the business. But at the onset, you want to evaluate for a stacking series of factors where... More are positive than negative. Right. And so the scenario, maybe it’s even easier to talk about what would disqualify a company. So from a credit investor standpoint, unlike equity, where you might make 10 investments and one to two are going to offer you hopefully some outsized returns and there might be a few investments that offer some semblance of liquidity events, and then a few that just don’t take shape, which is expected. Credit investors want to get paid every single time, right? And it would be the exception, not the norm, that that’s the case. So they’re going to look toward various elements of the business where they can successfully be repaid. And they’re going to look to a base of assets. Assets can be defined a lot of different ways. It could be hard fixed assets. So think machinery, equipment, real estate, things like that. could be softer assets, accounts receivable, purchase orders, contracts, IP of a company, for instance, or they’ll look to cash flows or debt service within a company. So, you know, if the company is making money, to successfully service a principal and interest payment going back to a lender. And if those things don’t exist, which by the way is the case for most ventures out there that are raising equity capital and building SaaS or more asset-like businesses, they would look instead towards the current cash that exists within the business, the current burn rate and or runway. and hey if we invest credit in the business are we going to accidentally end up in a scenario where the business is running out of cash and can’t repay us in three or six or nine months so so anyways going back to my comment around it’s kind of easier to start with what doesn’t work what doesn’t work is probably a company that’s either just way too new and they need to rely on equity for a little bit or they’ve been around for a while but maybe they’ve encountered some headwinds within the business where revenue has been declining, maybe margins have been getting compressed a little bit. Their equity capital providers might not have as much belief in the business as they did previously, which impacts their ability to bring in additional cash and liquidity and maybe they don’t have line of sight to near-term profitability or they do but it’s in a year or two and they’re going to run out of cash in like three or four months i think those scenarios are really tough to get credit excited about unless there is some sort of asset base to lend against in which case it’s still possible but absent that you know we’ll We’ll try to find any and all possible angles where we can make a deal work. I’ll give you an example, and there’s many others like this that I can reference as well that might surprise folks. But, you know, we also work in distress settings with those companies that you know, are not operating at the level that they probably once used to. And even companies that have navigated through, you know, a chapter 11 bankruptcy or restructuring process, etc. We did a deal, you know, about a year and a half ago, it was a company, there probably aren’t many of these companies, so I’ll be careful about the industry, but they had a very unique underlying asset in the way of pools so when you go and install an in-ground pool in your house right there’s different components and molds and stuff like that right and as you can as you might imagine those underlying assets are going to be very very unique you know relative to maybe hard yellow assets construction equipment bulldozers and stuff like that which everyone’s going to want in the in the case of a downside scenario but But yeah, we were able to successfully bring forth a little over $4 million for that company to successfully navigate through the restructuring process and come out in a really good shape with a lot of momentum on the other side where their creditors and their vendor relationships had a lot of confidence in the business going forward. So yeah, we’ve kind of seen everything, but hopefully that’s helpful. And again, it’s more art than science, but those are some of the common factors that we’re kind of taking a look at.

Brian Bell (00:21:41): So what are some of the risks that founders overlook with taking on debt like this and how should they think about them?

Ryan Ridgway (00:21:46): I would say one key risk is how is that loan or facility going to potentially affect you personally? Meaning, you know, is there a personal guarantee? So a personal guarantee isn’t always a bad thing. Sometimes it means you get access to better cost of debt, right? And if you look towards this massive ETA entrepreneurship or entrepreneurship through acquisition model, right? Most of these companies are getting credit from SBA loans. And those very clearly stipulate a personal guarantee, right? And that’s kind of the risk that you sign up for. But you’re getting debt that goes out 10 years, even upwards 25 if there’s like a real estate component ascribed to it as well. And it’s tough for other non-bank or private credit lenders to offer something similar in the capacity. So that’s kind of one risk. Another risk is maybe missed timing or misalignment of the debts there are times where you might have to take capital in various tranches to reach the ends kind of goal right and that can it can be a pro and it can be a con as well the pro is that you can deliberately draw down capital as warranted within the business that can be really valuable the con is businesses are susceptible to change, right? So what you outline as your go forward plan or strategy might change three, six, nine months from now, and you might have to go back to the drawing board. But yeah, I mean, I think that’s more just operational risk than credit specific risk. But those are some of the main ones that I think about.

Brian Bell (00:23:27): Yeah. What are you seeing in the market right now? What’s changing over the last few years? What’s changing now? What do you see kind of changing in the future?

Ryan Ridgway (00:23:34): I think it depends on what underlying company or kind of type of credit that you’re looking towards. I think one thing that’s really shifted from the the ZERP era, zero interest rate policy where capital was flowing very, very freely. So this is kind of in the years post-COVID is you had a lot of equity dollars flowing and subsequently you had a lot of kind of classical venture debt dollars flowing. And if you think about how traditional venture debt groups think and work, they like to append themselves to a near term equity event if they can, because it’s going to help in the way of the cash position, strengthen their borrowing relationship and so forth. So when equity dollars dry up, the definition has to kind of change within the traditional venture debt model. And what that gave way to, from our observation, is a lot of really creative groups that I would still classify as growth debt, but still that’s our term that we use at least, but they don’t necessarily need to come in alongside a big equity round. So it means they, you know, even in the setting of a loss making business, they can probably come to the forefront with some capital, they might just have to get very creative on how they structure it. And so it might be show us your pro forma model as to how you intend to reach profitability. You know, okay, here’s a vote of confidence in the form of capital. Conjoined with that, it also might be a conversation where, hey, by the way, we are going to raise equity. We have some initial interest, we’re soft circling, we’re kind of building out the raise, but we’ll probably intend to close in the next three months, six months or in an ongoing pattern, in which case it’s possible for credit to come to the table with a contingent term sheet that says, great, here’s a million or two today. We’re going to upsize that to five or 10 million or beyond as you kind of start to execute on your plan. And so that’s something where we’ve seen a lot of innovation taking place. I also think the technology that supports those types of decisions has just gotten a lot better. Another trend that we’re constantly seeing more and more frequently amongst lenders is their use of various technologies to inform underwriting and credit related decisions and so a few years ago what might have been more customary to say great we’re going to build a data room we’re going to go to a very formal ic or credit committee meeting to make these decisions which still definitely still human intervention and purview on all of these deals but um you’re seeing a lot more automated onboarding and intake, which might include a great, you bank with First Republic or Chase or whoever, connect your banking level data. Great, you run your accounting through QuickBooks or NetSuite or Xero, go ahead and connect that level of data. Or great, you’re a consumer brand, you sell through Shopify or you have these various marketing platforms that really drive a lot of unique data around the business and where your revenue is coming from, go ahead and connect those systems of record as well. And so you end up in these scenarios where I think it’s led to a couple things. One, it enables lenders to move faster through their decisioning. I think it also enables lenders to move down market a little bit as well. Because if you think about the human centric model, you might have a team of analysts and associates that are taking a look at a deal and everyone’s got their six figure salaries and you add it up and it’s like, gosh, why would we look at a half million dollar or even a two million dollar opportunity if it’s going to take $50,000 of man-hour to even make a decision around this, whereas now it might not be $50,000. It might be the cost of the underlying data. It might be $10 or $100 to reach that same initial inflection point where they gained confidence this. And then humans can be freed up to kind of do the work that we do best, which is the relationship making and a lot of the minutiae involved in actually closing and consummating the deal. So that’s been a big shift. We will, I think, continue to see that take place. We’ve even seen completely automated underwriting take shape through. It’s still a pretty low bar. I’d say it’s up through maybe a half million in terms of quantum of capital. But I think as the years progress, you’re going to see a million to $5 million deals get done almost 100% through, you know, automated underwriting and, you know, automated, you know, background checks and all of the, you know, you connect the rails one after the other, and there’s going to be more confidence there.

Brian Bell (00:28:28): Yeah, so that’s pretty fascinating on the lending side. What are you seeing, if anything, on the startup side with the underlying businesses?

Ryan Ridgway (00:28:35): Well, I think AI is the general kind of buzzword, right? So either you are building an AI or AI adjacent business, or you are adopting it for sure. And in our case, like we, you know, for instance, we run a pretty traditional advisory business, you know, we’ve got a handful of folks on the cap table outside of myself, but no big intention to go raise equity. You know, But we’re using AI very heavy to help drive good outcomes and good efficiencies with the process that we run. So that’s, we’re kind of a tried and true example on that front. But apart from that, AI is, it’s here, it’s not going anywhere, right? And so I think what we’ve been seeing is it shows up in the business in terms of many positive pivots that we’ve seen. It shows up in the business in the way of rebranding. I’d say also we’ve seen a lot of classical industries start to differentiate themselves through new products development that’s AI driven as well. So for instance, we’ve actually worked pretty frequently with a few companies. I can think of four or five over the last half a year or so. that are maybe in the space of IT or information security. And so if you look historically at their business model, they’ve held customer contracts that might pay them a few hundred or a few thousand dollars a month. They kind of said, hey, we can leverage AI to do a lot of this. And they end up licensing out their products and kind of by proxy of that, gaining a lot of predictability through their revenue model. And so that part’s been really interesting because now all of a sudden you have this kind of like traditional longstanding business that might be opening themselves up to raising some equity or raising some classical kind of venture debt. And it’s, yeah, it’s just been interesting to see these businesses kind of shift.

Brian Bell (00:30:28): Yeah. Let’s wrap up with some rapid fire, rapid-ish questions. What’s the one question you wish founders asked more often when evaluating their capital strategy?

Ryan Ridgway (00:30:39): What’s my long-term goal for the business? I think if you’re evaluating your capital strategy at zero, you should almost work backwards in a sense is this something you’re gonna sell and have a giant exit from is it something that you know you want to work with a strategic party you know private equity um do you even want a liquidity event i’d imagine you know most venture-backed startups, whether they want it or not, they have other parties that want the liquidity of it. So that’s a little bit different. But for traditional businesses, what does that path look like? Maybe you want to hold onto this business and it’s something you’re really passionate about and you want to grow it for 30 years. So I think a lot of these things, if you start with the end outcome in mind, you can arrive at what you should be focusing on today. And I think I’m guilty of this myself. I think we have a habit or as human nature to focus one foot in front of the other, but you can be walking and you might look up eventually and find that you’re over here when you kind of want it to be over here. So it’s good to develop a proxy from both sides, in my opinion.

Brian Bell (00:31:45): What’s a trend that you think the broader venture ecosystem is underestimating right now?

Ryan Ridgway (00:31:50): I think a couple things and I won’t get into kind of like geopolitical aspects, but I think the equity landscape has not been as kind as it previously was to the climate sector and alternative forms of energy. I think the that’s probably a long-term mistake or misstep and from my personal perspective and i think the founders that are in that segment if they’re patient enough and protective enough and good stewards of capital and very efficient will do well in the long run at least from a career standpoint i think the other one is not necessarily a theme or an industry but it’s it’s a location i think that uh i’m very biased because we do business in LATAM and my wife is from Colombia. And so we visit frequently and stuff, but we come across so many amazing founders down in Central and South America coming from, you know, typically cities like Mexico City or Boruta and other places. And, you know, I would encourage venture groups that if they’re not exploring more broadly into other areas, geographies like there or across the GCC region, the Gulf Coast, Dubai and the Emirates, et cetera. Just got back from a trip during the latter part of this last year to Dubai and a few other cities. It’s incredible what’s taking place there right now. It really is. And so, yeah, I guess what I would say is like there’s There’s always a lot more opportunity beyond your front door or beyond Silicon Valley or New York and kind of, you know, Austin, Texas, some of the major hubs.

Brian Bell (00:33:28): What’s the best piece of advice you ever received?

Ryan Ridgway (00:33:30): Oh my gosh, put me on the spot here. I guess like these soundbites all kind of adjoin into one like central way of thinking. I was at a virtual event that Tony Robbins put on a couple of weeks ago and I he was talking about belief. And he was talking about, you know, what you think you will speak what you speak, you will watch ultimately show up in your life, whether that is positive or negative. And I think that’s one thing that’s really helped drive me personally. And I’m pretty type A. I’m pretty matter of facts. I’m Mr. Show me the numbers. But it is really important to touch base with yourself emotionally. How are you feeling? What are you thinking? And try to have a little bit more control over those thoughts because they are really important. And he was talking about I don’t come from a medical background, but it’s the RAS. I think it stands for Reticulated Activating System or something like that. Don’t quote me on that. But it’s the portion of your brain that is keenly involved with pattern recognition. He gave the example of, yeah, if you buy your Audi S5, right, you’re gonna get out on the street and you’re just gonna see them all over the place. And so that’s very adaptable to mindset and how you treat your business. If you are accompanying yourself with founders in your space or who’ve gone through a similar motion, you’re building this narrative that, okay, I’m going to follow that exact path. And it becomes a lot more predictable. However, if you’re looking at the negative, you’re always going to find the negative. And that’s probably going to appear in your life as well. So I know that might sound fluffy, but that’s something recent that kind of stuck out to me that’s really had me check myself and founders at the end of the day. I mean, it’s tough. You’re focused on 10 million different things, 10 million different hats, and it’s excruciating and it’s exhausting. So it is important whether we want to or not to really step away and take a few deep breaths and try and focus on where we want to go and be, it’s almost like an Aussie moron, right? Like be deliberate about how you’re going to get there, but be relaxed about it at the same time. Like give yourself enough rethought.

Brian Bell (00:35:44): For founders listening who want to make smarter capital decisions, what’s the first step they can take today?

Ryan Ridgway (00:35:49): Well, I would say make the most accurate decisions at the onset. So that probably means table states is having your books in order, having your reporting in order. There are goals that you’re after and make sure that you have really good reporting and mechanisms. in place to actually report on those go. And I think that’s kind of the foundation you need to build on because if you don’t have that, you’re going to be going through a lot of guesswork and it’s going to be really hard to better inform what those capital decisions might look like. So that’s kind of fundamental, have the reporting in place, have good numbers, ideally being proactively fed to you in some way. And then apart from that, I’m sure there’s a lot we could focus on there, but I think I’ll go back to my prior comment on what’s your end goal? And revert back from there. So think about what you want. If you want to finish a marathon, redact out the end finish line and maybe focus on the first mile and what’s next from you. But know that you’ll finish.

Brian Bell (00:36:47): How should early stage VCs like myself rethink the role of debt or alternative capital in portfolio construction?

Ryan Ridgway (00:36:54): So that’s a really fair question. I think it just goes down to the underlying use case of the capital. Equity obviously has its strengths in the first strongest one, right? Doesn’t need to be paid back until liquidity events, at least most commonly. So I would say that private credit or non-bank capital has really stepped up meaningfully. It’s here to stay. It’s not going anywhere. And in fact, it’s coming down market. And so companies that might not have previously been eligible for debt at a kind of subscale or nascent level setting where a pre-seed or a seed investor might come in, might now all of a sudden be eligible for debt. I think that could potentially bring forth both positive and negative connotations for an equity investor. One, if they really are attracted to the business, it could be, well, gosh, they need less of my capital. Two, though, it could be, well, gosh, this is actually great because they’re a little bit less reliant on equity. I really like the business and the small subset of equity investors at the onset do. And so it might give you leverage that as they’re bringing instrumental debt into the company to now be a lot more participatory in subsequent financings and fundraising events that would have otherwise called for maybe a broader brushstroke against the markets and driving up the valuation more, etc.

Brian Bell (00:38:24): What’s a book, newsletter, or resource you regularly recommend to founders and operators?

Ryan Ridgway (00:38:30): Well, I think from a book perspective, I love Zero to One, Peter Thiel. It’s kind of a classic that I’ve flipped through, but it’s either called The Almanac or The Navalmanac, but it’s Naval’s book and pieced together from a lot of his podcasts and excerpts over the years. And so anything kind of Naval, and for those that are less familiar, original founder of Angel List, who has now broken out and decided in addition to a lot of his ventures and investing, he’s really embraced philosophy as well and how it comes into your business and life. And he just has some really profound ways in, at least for me, kind of taking you out of yourself and the minutiae and really forcing you to think different about your business. I really like him and a lot of the content that he pushes out. Apart from that, there’s a bunch of industry and trade pubs that I like to nerd about. So if you’re interested in those, there’s a few that constantly push out like credit-focused deal flow. And I can get those in there. Contact me directly. I’ll point you in the sources of some good resources there. But that’s kind of the more nerdy industry-specific stuff.

Brian Bell (00:39:49): Well, Ryan, I really enjoyed the conversation. Where can folks find you online?

Ryan Ridgway (00:39:52): Find me on LinkedIn. I’m most active on LinkedIn. or feel free to email as well. I’m at ryan at cirruscap.com. And even if you’re just getting started, kind of year one, year two of the business and maybe thinking about, you know, am I eligible for non-dilutive capital? What’s out there? Don’t be afraid to reach out. You know, we have plenty of conversations early on with folks in a nascent stage. And, you know, we stay in touch and we come together a couple of years down the road and end up, you know, being really supportive of one another, whether it’s capital being brought forth or some other way we can work together. So, yeah, those are the two best ways, I think.

Brian Bell (00:40:29): Awesome. Well, yeah, thanks again.

Ryan Ridgway (00:40:30): Really appreciate it. Yeah, you too, Brian. Really appreciate you having me on.

More calls. More emails. More hires.

Neil Weitzman sees the opposite. The real issue is almost always structure.

If you don’t have a repeatable go-to-market system, adding more activity just makes failure happen faster.

Here’s what actually matters.

Subscribe now

The Core Mistake: Scaling Before You’re Ready

A lot of founders jump to scaling too early.

They hire SDRs. They push outbound. They spend on tools.

But they skip one critical step: proving what works.

Neil frames it simply. If you haven’t run your GTM motion enough times to understand what “good” looks like, you’re not scaling. You’re experimenting.

And experimentation doesn’t scale.

Example:

• You close 5 deals out of 50 prospects → maybe promising

• You close 5 deals out of 500 prospects → something is broken

Same result. Completely different signal.

Without that clarity, hiring more people just multiplies inefficiency.

What “Good” Actually Means in GTM

Most founders can’t define this clearly.

Neil pushes for precision:

• What does a strong cold call sound like?

• What does a good email look like?

• What conversion rate should you expect from meeting → close?

• What daily activity produces results?

Until you can answer those questions with real data, your GTM is not a system.

It’s guesswork.

A real system means:

• You’ve tested multiple approaches

• You know what consistently works

• You can teach it to someone else

• You can predict outcomes with reasonable accuracy

That’s when scaling starts to make sense.

Product-Market Fit Isn’t Just Revenue

Founders often use ARR as the signal.

Neil looks deeper.

He focuses on:

• Retention: Do customers stay?

• Dependency: Do they actually need the product?

• Switching risk: Would they leave for a small discount?

• Behavior: How are they using it day-to-day?

You don’t need 1,000 customers to know you have product-market fit.

But you do need enough customers to see consistent patterns.

If customers are sticking, using the product heavily, and getting real value, you’re getting close.

If not, GTM tweaks won’t fix it.

Founder-Led Sales Still Matters (Early)

Before building a team, founders should stay close to sales.

Not because it’s scalable, but because it’s informative.

Neil’s advice is simple:

If you only have 1–2 customers, don’t build a GTM machine.

Ask a better question:

“How do I get 3 more customers this month?”

That usually leads to:

• Your network

• Warm introductions

• Investors and advisors

• Direct outreach to known prospects

It’s manual. It’s scrappy. It works.

And more importantly, it teaches you how buyers actually think.

Why More Sales Reps Don’t Fix Growth

This is one of the biggest myths in early-stage startups.

If your current reps aren’t converting, hiring more won’t solve it.

You’ll just:

• Burn more cash

• Create more noise

• Confuse what’s actually broken

Neil sees this often. Founders blame the rep.

But the real issue is usually:

• Weak messaging

• Poor targeting

• No defined process

• No feedback loop

Fix the system first. Then add people.

The Shift to Relationship-Driven GTM

Cold outbound still works. But it’s harder than ever.

Email volume has exploded. Everyone is competing for attention.

Neil leans into a different approach:

Lead with value, not a pitch.

Instead of:

“Want to book a demo?”

Try:

• Sharing something useful

• Making an introduction

• Offering insight relevant to their role

• Referencing something specific about their business

The goal isn’t a meeting.

It’s a relationship.

That shift alone can dramatically change response rates.

Why LinkedIn Matters More Than You Think

Many founders delay content and brand.

Neil thinks that’s a mistake.

LinkedIn is one of the easiest ways to:

• Build credibility early

• Share your thinking

• Attract inbound interest

• Stay top of mind

Even if your buyers aren’t active, their network is.

That creates second-order effects:

• Referrals

• Introductions

• Unexpected opportunities

You don’t need a massive following.

You just need consistency.

When to Bring in a Fractional CRO

Not every founder needs one immediately.

But you should consider it early if:

• GTM isn’t your strength

• You’re unsure what’s working

• You’re about to hire your first sales team

• You’ve hit a plateau

The key benefit isn’t strategy.

It’s execution.

A good fractional CRO helps you:

• Define what “good” looks like

• Build a repeatable system

• Track the right metrics

• Avoid expensive mistakes

And ideally, they work themselves out of a job.

The Pattern Behind Breakout Startups

Neil sees one consistent trait.

Speed of decision-making.

Strong founders:

• Run experiments quickly

• Act on partial information

• Adjust fast when something isn’t working

They don’t wait for perfect data.

They move, learn, and iterate.

That compounds over time.

Final Thought

GTM isn’t about doing more.

It’s about doing the right things, consistently, with clarity.

If you don’t know what’s working yet, slow down.

Test. Learn. Refine.

Once you have that, growth becomes much simpler.

👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL

🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast

Chapters:
00:01 – Introduction to Neil Weitzman
02:55 – Founder Leadership Gaps
05:00 – When Founders Aren’t a Fit for Help
06:18 – When to Bring in GTM Support
09:36 – Building GTM Early
11:55 – Defining “What Good Looks Like”
12:47 – When to Scale GTM Teams
15:48 – Risks of Scaling Too Early
16:45 – Identifying Product-Market Fit
19:30 – Importance of GTM Data and Systems
20:25 – GTM Tech Stack Essentials
22:44 – LinkedIn and Sales Navigator Strategy
26:29 – Effective, Non-Salesy Outreach
31:04 – Hiring a Fractional CRO
35:11 – Execution vs Strategy
36:02 – Fractional CRO Engagement Model
38:14 – Transitioning to Full-Time CRO
40:52 – Systems vs Sales Talent
41:31 – Porch and Immigrant Founder Support
45:15 – Early GTM Priorities
48:13 – Network-Led Early Sales

Transcript