YC Spring 2026 Companies We Backed
We just finished another intense YC batch cycle, and Team Ignite invested in 17 companies from YC Spring 2026.
We already published a separate batch takeaways post with our broader impressions of the cohort, category-level observations, and what we think this batch says about where early-stage AI and software are heading. You can read that here:
This post is more direct: here are the companies we backed, what they do, and why we were excited to invest.
A few patterns stood out.
First, AI is moving from “copilot” to operating system. Many of the best companies are not building thin wrappers. They are trying to own full workflows: customer success, medical practice operations, construction estimating, pathology reporting, observability, marketing execution, sales development, and more.
Second, infrastructure remains a major theme. GPU utilization, HPC backtesting, app observability, and AI-native workflow systems are becoming more valuable as AI increases the volume and complexity of work.
Third, the batch was broader than generic SaaS. We invested across defense, aerospace, healthcare, fintech infrastructure, devtools, vertical AI, and applied AI agents. That matters because the biggest outcomes rarely look obvious at the start.
Our job is not to wait until consensus forms. Our job is to identify early signal, move quickly, and earn allocation before the round is gone.
The Companies We Backed (Alphabetically)
Arlo Industries
What they do: Arlo Industries is building a passive aerial sensing mesh to track drones, missiles, and other aerial threats without traditional radar.
YC Profile: https://www.ycombinator.com/companies/arlo-industries
Why we invested: Defense is being reshaped in real time by low-cost drones, autonomous systems, and asymmetric warfare. Legacy radar systems were not designed for a world where cheap aerial threats can appear everywhere at once. Arlo’s distributed, passive sensing architecture is compelling because it attacks both the technical problem and the economic problem: how do you create persistent, wide-area coverage without relying on centralized, expensive infrastructure? We were attracted to the severity of the problem, the timing, and the potential for Arlo to become a foundational sensing layer for modern defense.
Dispatch
What they do: Dispatch is building refurbishable reentry vehicles that can host and return payloads for companies manufacturing ultra-high-value materials in microgravity.
YC Profile: https://www.ycombinator.com/companies/dispatch
Why we invested: Space manufacturing is one of those categories that sounds futuristic until the enabling infrastructure arrives. The core bottleneck is simple: if companies can manufacture valuable materials in space, they still need a reliable way to bring those materials back to Earth. Dispatch is attacking that bottleneck directly. We liked the ambition, the technical depth, and the founder-market fit from a team with relevant spacecraft experience. This is a high-risk, capital-intensive category, but the upside case is enormous if Dispatch becomes a core logistics layer for in-space manufacturing.
Expanse
What they do: Expanse helps teams recover wasted GPU and HPC capacity by predicting the resources compute jobs actually need before they run.
YC Profile: https://www.ycombinator.com/companies/expanse
Why we invested: AI infrastructure spend is exploding, but much of that compute is still wasted through over-provisioning, failed jobs, and poor resource prediction. Expanse is building an intelligence layer for GPU and HPC infrastructure: read the job, understand the cluster, predict what resources are needed, and reduce waste before money is burned. The team has unusually strong founder-market fit, having worked directly on the kinds of HPC and GPU workloads they now serve. We invested because compute efficiency is becoming a strategic budget issue, and Expanse has a credible path to becoming a control point for cluster utilization.
InstaAgent
What they do: InstaAgent helps consumer brands create, distribute, test, and learn from large volumes of personalized social creative across personas and channels.
YC Profile: https://www.ycombinator.com/companies/instaagent
Why we invested: AI has made content creation cheap. It has not made content effective. That distinction matters. Consumer brands do not need more generic AI slop; they need high-quality creative variation, fast testing, and a learning loop that compounds across channels. InstaAgent’s wedge is scaled social creative generation and distribution, but the bigger opportunity is becoming a workflow layer for modern brand marketing. We liked the team’s speed, their understanding of social distribution, and the possibility that creative testing becomes more software-like as AI changes how brands operate.
Keyframe Labs
What they do: Keyframe Labs is building visual AI avatars and multimodal agents designed to become more scalable, expressive interfaces for AI applications.
YC Profile: https://www.ycombinator.com/companies/keyframe-labs
Why we invested: Voice AI was one major interface shift. Visual, embodied, multimodal AI may be the next. Keyframe is attacking the cost and scalability constraints that have historically made realistic avatars difficult to deploy widely. We liked the team’s technical ambition and the possibility that avatars become a core interface layer for sales, education, support, entertainment, and agentic software. If AI agents are going to represent companies, teach users, sell products, or guide workflows, they may need faces, presence, and visual interaction—not just text boxes and voice streams.
Klarify
What they do: Klarify is building AI software for therapists and mental health practices, starting with documentation and expanding into broader administrative workflows.
YC Profile: https://www.ycombinator.com/companies/klarify
Why we invested: The best vertical AI companies do not replace the professional; they remove the non-core work that prevents the professional from doing their highest-value job. Klarify’s insight is exactly that. Therapists are overburdened by notes, treatment plans, claims support, scheduling, payments, and client follow-up. Klarify keeps the human therapist at the center and uses AI to automate the surrounding operating burden. We liked the clarity of the wedge, the severity of the administrative pain, and the potential to expand from notes into the operating system for small and mid-sized mental health practices.
Memory Store
What they do: Memory Store is building memory infrastructure for AI applications, helping AI systems retain, retrieve, and use context over time.
YC Profile: https://www.ycombinator.com/companies/memory-store
Why we invested: Memory is one of the most important unsolved primitives in AI applications. Today, many AI products feel impressive in isolated interactions but weak across time because they do not remember enough, structure context well enough, or retrieve the right history at the right moment. Memory Store sits at a foundational layer: persistent memory for AI-native software. We invested because every serious AI workflow eventually needs durable context, personalization, and recall. If they become the memory layer for a meaningful share of agentic applications, the opportunity is large.
Oddpool
What they do: Oddpool is building data and infrastructure for prediction markets and event-based financial markets.
YC Profile: https://www.ycombinator.com/companies/oddpool
Why we invested: Prediction markets are moving from niche curiosity to serious financial infrastructure. As venues, assets, and trading strategies proliferate, institutions need normalized data, symbology, historical records, settlement metadata, and APIs they can build on. Oddpool’s opportunity is to become the neutral data and reference layer for this emerging market structure. We liked the infrastructure angle: as the category grows, high-quality historical data and normalized cross-venue infrastructure become harder to replicate and more valuable over time.
PerfectBit
What they do: PerfectBit is building AI infrastructure for software engineering and code intelligence.
YC Profile: https://www.ycombinator.com/companies/perfectbit
Why we invested: Software engineering is one of the first major labor markets being reshaped by AI, but the tooling stack is still early. The opportunity is not merely autocomplete. The larger prize is understanding codebases, automating complex engineering work, and helping technical teams ship faster with fewer bottlenecks. We liked PerfectBit because the category is enormous, the timing is right, and even small improvements in engineering productivity can create significant customer value. The risk is competition, but the market is so large that multiple durable companies can emerge.
PLAN0 AI
What they do: PLAN0 AI builds AI-native construction cost estimation and analytics software.
YC Profile: https://www.ycombinator.com/companies/plan0-ai
Why we invested: Construction cost estimation is slow, manual, expensive, and error-prone. It also sits at a critical point in the construction value chain: decisions made early can determine whether a project works economically years later. PLAN0 ingests floor plans and elevations, reconstructs projects, and helps produce cost estimates and scenario analysis far faster than traditional workflows. We invested because this is a severe, high-friction vertical problem with a clear AI wedge and a large potential expansion path into analytics, forecasting, and workflow ownership for construction teams.
Plena Health
What they do: Plena Health is building a full-stack operating system for specialty medical practices.
YC Profile: https://www.ycombinator.com/companies/plena-health
Why we invested: Specialty medical practices are operationally complex, understaffed, and buried under repetitive workflows: phones, scheduling, prior authorization, results follow-up, billing, collections, and EMR-adjacent administrative work. Plena’s thesis is that AI can run more of this operational stack end-to-end while working inside the practice’s existing systems. We liked the vertical depth, the wedge into painful back-office workflows, and the potential to become a deeply embedded operating layer for specialty care. Healthcare AI is crowded, but practice operations remain brutally inefficient and highly valuable if solved.
RASPIRE
What they do: RASPIRE builds runtime application security for mobile apps, protecting compiled applications from tampering, exploitation, and reverse engineering.
YC Profile: https://www.ycombinator.com/companies/raspire
Why we invested: Mobile apps are increasingly critical infrastructure for fintech, gaming, consumer software, health, and enterprise workflows. Yet many applications remain vulnerable once they are running in the wild. RASPIRE’s wedge is runtime protection: defending the app while it executes, not merely scanning code before release. We liked the severity of the security problem, the developer-facing adoption path, and the potential for RASPIRE to become a key protection layer for high-risk mobile applications. In a world of more AI-generated code and more automated attacks, runtime protection should become more important, not less.
Sherpa
What they do: Sherpa helps companies improve website conversion by using AI to test, optimize, and personalize growth experiences.
YC Profile: https://www.ycombinator.com/companies/sherpa
Why we invested: Growth teams know their websites leak revenue, but most conversion-rate optimization is slow, manual, and dependent on limited testing bandwidth. Sherpa’s opportunity is to turn CRO into an AI-native workflow: identify friction, propose improvements, run experiments, and compound learning across pages and customers. We liked the clear ROI, low-friction installation, and obvious pain point. The key question is whether Sherpa becomes a durable growth control plane rather than a point solution, but the wedge is strong and the buyer pain is easy to understand.
Superlog
What they do: Superlog is building autonomous observability for software teams.
YC Profile: https://www.ycombinator.com/companies/superlog
Why we invested: Observability has become essential, but it is often expensive, noisy, and painful to configure. Developers do not want more dashboards; they want systems that understand what is happening, collect the right context automatically, and help resolve issues faster. Superlog’s thesis is that observability should become more autonomous: less manual instrumentation, less configuration burden, more agentic debugging and system understanding. We liked the founder urgency, the technical wedge, and the category timing as AI changes how software is built, monitored, and repaired.
Userlens
What they do: Userlens builds AI customer-success agents that detect churn risk, generate account insights, prepare QBR materials, and help teams manage renewals and expansion.
YC Profile: https://www.ycombinator.com/companies/userlens
Why we invested: For B2B SaaS companies, retention and expansion are existential. The problem is that churn signals are scattered across product usage, CRM notes, billing, support tickets, customer conversations, and CSM intuition. Userlens turns that fragmented data into an AI CSM that can monitor accounts, identify risk, prepare playbooks, and support renewal workflows. We liked the repeat-founder angle, the clear pain from the founders’ prior company, and the possibility that customer success becomes increasingly agentic. The best version of Userlens is not a dashboard. It is an operating layer for revenue retention.
Voquill
What they do: Voquill is building an AI coworker for pathologists, starting with voice-driven report generation and expanding toward broader lab workflow automation.
YC Profile: https://www.ycombinator.com/companies/voquill
Why we invested: Pathologists spend an enormous amount of time documenting cases, producing reports, and navigating administrative friction. Voquill’s wedge is highly specific: listen as a pathologist works, learn their reporting style, and generate sign-out-ready reports. We like vertical AI products that start in a painful, repetitive workflow and expand into a larger system of action. Voquill is not a generic medical scribe; it is focused on pathology, where workflow specificity, reimbursement, and lab operations matter. If the company becomes embedded in pathology labs, the expansion opportunity is significant.
Zibra Labs
What they do: Zibra Labs builds high-performance computing infrastructure for quant trading firms running large-scale backtesting.
YC Profile: https://www.ycombinator.com/companies/zibra-labs
Why we invested: AI is accelerating the generation of candidate trading strategies, but backtesting infrastructure is becoming the bottleneck. Quant teams can produce more hypotheses than their systems can evaluate. Zibra Labs is building HPC infrastructure to make large-scale backtesting faster, cheaper, and easier to run across cloud resources. We liked the founder-market fit and the technical credibility of the team. This is a narrower market than generic cloud infrastructure, but it is a high-value buyer segment where performance, scale, and cost efficiency matter enormously.
Why This Matters for Fund III
This batch is a good example of why Team Ignite exists.
YC batches are now large, noisy, and fast-moving. The best companies do not wait politely for every investor to finish diligence. Rounds form quickly. Allocation disappears quickly. Consensus is usually late.
Our Fund III strategy is built around that reality: review the batch early, process signal quickly, meet founders before Demo Day when possible, and invest with disciplined speed.
Team Ignite Fund III is our second YC-focused fund. We have now made 101 YC investments to date in the fund, and the portfolio is already marked up at nearly 3x on 2024 investments. The fund is still young and still actively deploying, but the early signal is strong.
Fund III Snapshot
Updated deck: https://tr.ee/yc_fund
Join the fund: https://teamignite.decilehub.com/pacts?pid=7Na3Qz8D
Closes End of June 2026
Our view is simple: if you want exposure to the next generation of YC companies, the edge is not waiting until everyone agrees. The edge is seeing enough of the batch, moving early, and being useful enough that founders want you on the cap table.
That is what Team Ignite is built to do.