Requests for Startups
RFS is our tradition of sharing ideas wed like to see founders tackle. These represent just a fraction of what we fund — if one excites you, take it as extra validation to dive in, but you dont need to work on these ideas to apply to YC.
Fall 2026Summer 2026Spring 2026Fall 2025Summer 2025Spring 2025Winter 2025Summer 2024
Fall 2026
AI is moving into the physical world. Were excited about a new wave of startups rebuilding the systems that power the real world, from education and healthcare to defense, finance, infrastructure, and work itself. This batchs requests come from YC partners and founders building on the frontier, and include, for the first time, one from the sitting U.S. Secretary of the Army.
The Primer#
By Andrew Miklas
In Neal Stephensons The Diamond Age, a young girl is given an interactive book called A Young Ladys Illustrated Primer. It looks like a tool for learning to read, but its far more. It adapts to her completely, and through stories tuned to her life, it teaches her not just to read but to think, to reason, and eventually to grapple with the hardest questions of ethics, meaning, and character. She returns to it day after day for years, and as she grows, it grows too. Continually reshaping itself around who she is becoming and the life she is living. For the first time in history, something like the Primer is starting to feel possible. The best education has always come from one-on-one tutoring. Aristotle taught Alexander. But that privilege has been reserved for the few. AI can bring it to every child. And great tutors do more than drill facts. Over years, they learn a childs mind, and with it how to teach what cant be drilled at all: thinking, reasoning, even wisdom. In Stephensons story, the Primer is an AI tutor that never runs out of patience or time. Were a long way from building one, but we can start today. What wed love to see now is a product that adaptively teaches young children to read, write, and do arithmetic, at the quality of a devoted private tutor and at consumer scale. Not a replacement for teachers, but a supplement that makes them more effective. While we think this begins as something a parent buys to help their child learn basic skills, it is also the entry point to far greater ambitions. A company that gets it right could build toward something like the Primer, and even a fraction of that vision would have a profound effect on society. If youre building this, wed love to hear from you.
The Future of American Defense#
By Daniel P. Driscoll, US Secretary of the Army
Warfare is at an inflection point, and the old ways of Army acquisition simply dont keep pace with modern threats. Modern combat demands commercially developed, modular open-system solutions, and thats why we ripped up the old acquisition playbook. We need hungry, innovative founders to build for the crucible of ground combat. The threats we face change daily, and the Army must dominate everywhere, from the Arctic to the archipelago and from space to subterranean environments. So here is our request for startups. We are actively funding low-cost interceptors or any component that helps us lower the cost per kill. We need next-gen sensors, software, payloads, and other hardware that plugs directly into our open system architecture. We need cutting-edge drones, resilient logistics, and advanced manufacturing, and we need it all to survive the most extreme climates on Earth. Bring us your ideas, and we will give you the capital and the proving ground to scale. There has never been a better time to build for the Army. The door is wide open, so lets get to work.
A Cloud for Small Software#
By Pete Koomen
Using agents to build personal software to solve your own problems is a lot of fun. This is Small Software. Purpose-built tools that will only ever have one or a small handful of users. Small software is useful for teams too: Every team does things differently, and theres unlimited demand for bespoke tools for running workflows, tracking important numbers, managing sprints, sharing prototypes, and so on. Software like this is now very easy to build, but still hard to deploy and share. Incumbent clouds like Azure and AWS were designed for shipping Big Software that scales with many users… at the cost of complexity. A cloud designed for small software could delete most of this complexity and unlock a big set of new use cases that agents have only recently made possible. At the same time, there are other hard problems to solve: every company will want to customize the environment this software runs in, auth permissions are hard, and allowing nontechnical users to share arbitrary code is tricky to do securely. Small software should be as easy to share with your colleagues as a Google Doc. If youre working on something like this, wed love to hear from you.
Multiplayer AI#
By Aaron Epstein
The best work tools of the last two decades won by going multiplayer. Google Docs replaced Microsoft Word. Figma beat Photoshop. And they turned solo tools into places where teams do their best work together. But AI hasnt had its multiplayer moment yet. AI agents are the most powerful new tool a team has, but its the one thing people still use by themselves. Thats because right now, working with AI is largely single-player. You open a chat, type a prompt, and get an answer, in a box only you can see. When you want to collaborate with your teammates and agents, the best you can do is send a link to a read-only transcript they cant touch. Thats about to change. Agents are starting to run tasks that take hours, days, even weeks. Work at that scale was never meant to be done alone, and pulls in many people across a company. Anyone on a team should be able to drop into the same live agent session to watch it work, redirect it, and hand it off, the way theyd work with any other human team member. This turns the work a team does with agents into a shared, living thing instead of a thousand private threads. We think theres a version of this for every kind of work. Shared agents for engineers coding together in real time. For sales teams working a deal together. For support teams resolving a ticket. For lawyers drafting a contract, analysts building a model, and marketers shipping a campaign. Anywhere a team already crowds around one problem, there should be multiplayer agents they all share. So if youre building AI thats multiplayer by default, wed love to hear from you.
Compute at Sea#
By Francois Chaubard
Artificial intelligence is running out of compute. And data centers are running out of electricity and land. The demand for data centers is insatiable. But new data centers can take years for approval and enough megawatts, and still could get killed by local government intervention. While communities increasingly oppose the land, the water is open. It sounds crazy, but we think part of the answer may be to move compute offshore. The ocean is 70% of the Earths surface area, has abundant sunlight, no permitting process, and is an enormous natural heat sink that is already getting hit by the sun all day anyway. Think of them as compute flotillas: many standardized, modular vessels operating together as one global cloud. We want to fund founders who are bringing the worlds compute to the oceans.
AI-Powered Consumer Products for 1 Billion People#
By Raphael Schaad
Every platform shift mints consumer giants. The web gave us Google + Airbnb. Mobile gave us Instagram + DoorDash. AI is the biggest shift YET. But three years in, the only new icon on your home screen is ChatGPT. So why do we think NOW is a great time to build AI-powered consumer products? Intelligence just got good enough: you can treat an agent like a person. And its about to get cheap enough, too: today, the magic can run $1,000 a month in tokens for each user, but that is falling 10x a year. Follow the curve, and you can predict the consumer moment: it lands very soon. Whoever builds now, owns it. What do you build? How we get things done, get around, learn, stay healthy, manage our money, play, connect with friends. It all opens up again. CONSUMER is going to be so back. If youre building for a billion people, wed love to hear from you.
AI for the Aging Population#
By Max Kolysh
By 2030, one in five Americans will be over 65, and theres nowhere near enough people to take care of everyone. The US is projected to have millions of unfilled caregiving jobs within the decade, and 53 million family members are already doing this work unpaid. Meanwhile, almost no technology is actually built for older people. Even Alexa and Google Home are frustrating for most seniors to use. AI finally makes a new class of products possible: voice interfaces that can hold real conversations, monitoring that helps older adults stay safe and independent, robotics that can assist with physical tasks around the home, and software that helps family caregivers coordinate care, appointments, and emergencies. This is one of the largest, most underserved markets in the world, and its growing every single day. If youre building for the aging population, wed love to hear from you.
New Operating Systems for the Physical World#
By Charlie Warren
80% of the global workforce doesnt sit at a desk. But the software for the physical world hasnt really changed in over 20 years. In construction, maintenance, and fleet operations, the software basically does some combination of: dispatching people, tracking them, managing assets, and billing the customer. Were excited about how AI changes that entire model of work. There are now three kinds of workers: - AI agents that can quote complex work and schedule teams. - Robots actually deployed in the field. - And humans who increasingly use wearables to record everything theyre doing. Todays operating systems werent designed to manage all three types of workers. And this creates all sorts of cool new opportunities for a startup. How do you route a job between an agent, a robot, and a person? What does safety look like when humans and robots work literally side-by-side? And how do you measure reliability? The opportunity here is heck of a lot bigger than the existing software. The incumbents charge for human coordination and visibility. The new operating systems will manage the robot and human labor together. And these industries spend 10 to 100x more on labor than software. Theres another reason you should build here. Youll record all the work as it actually happens. The frontier models, robotics startups, or software incumbents wont have that kind of end-to-end data. So if youre building this new kind of operating system for the physical world, wed love to hear from you.
The Best Time to Build in Crypto#
By Nemil Dalal
Its a dispiriting moment in crypto: prices are down, hot narratives have fallen flat, and many builders are leaving. It sounds crazy, but at Y Combinator were more optimistic than ever. Weve invested in more than 100 crypto startups, but we expect that number to go up a lot. Eventually, we expect every YC startup to use crypto rails from capital raising to payments, though most will probably never even know. Heres why were bullish. Many YC-funded fintechs way outside crypto are building on crypto, like Deel and Gusto. Regulatory clarity is finally here, stablecoins are being adopted by every major financial institution, tokenized stocks are transforming trading. Projects like Hyperliquid, with a tiny team, are making the top stock exchanges squeamish about their edge. And it feels inevitable that agents are going to use crypto networks as financial rails. Just as importantly, bear markets let the real projects build and thrive while bull markets are the worst time to build. In crypto, prices are decoupled from reality. In bear markets, you dont need to compete with a criminal offering infinite yield. They attract a different type of founder, more focused on building than getting an early liquidity event. Weve funded major teams building in crypto like Stripe, Coinbase, and Axiom, but also lots of infrastructure and bets on the future. For example, BlindPay and Infinia are building the developer interface and ramps for Latin America. Aspora is building the easiest way to transfer money to India. Just some of the things were excited about are capital raising, new stablecoins and stablecoin applications, agentic commerce, trading, institutional products, and scalable and private blockchains.
Data for the Real World#
By Austin Tindle and Diana Hu
AI has gotten remarkably good at learning from data. We now have superhuman models for code, language, and images. But for the physical world? Were still working with sparse data from remote sensors designed for humans, not AI. With improving foundation models and plummeting sensor costs, dense physical-world data collection is now feasible. And its happening. Gecko Robotics uses robots to collect data in hard-to-reach places and builds predictive models. At Sorcerer, we use autonomous weather balloons to collect data about the atmosphere, which the US government uses to make better weather forecasts. And the opportunity is much bigger. The worlds biggest industries are in energy, agriculture, logistics, and construction. They rely on limited data and intuition-based models. More real-world data enables precise modeling. And once you can model a system, you can control it. Steering hurricanes. Reversing desertification. Cooling the planet. Were interested in more companies like Sorcerer building new ways to collect physical-world data. If youre working on this, wed love to hear from you!
Proving Youre Human#
By Max Kolysh
Recently, a finance worker joined a video call with his CFO and several colleagues, and wired out $25 million. Every other person on that call turned out to be a deepfake. This isnt science fiction anymore. Voice clones and fake video calls are getting cheap and ultra-realistic, and fraud like this is exploding. The scary part is we dont have a good way to tell whos real anymore. It used to be that if you saw someones face or heard their voice, that was enough. Not anymore. Every trust signal we have was built for a world where faking a human was expensive, and that world is gone. So we think one of the most important problems of the next decade is rebuilding the trust layer of the internet: knowing theres a verified human on the other end of a call, a message, a transaction. We dont know exactly what the solution looks like. Ideally its one that doesnt make everyone give up their privacy. And its not just about stopping scams. Imagine Twitter with no bots in the replies, dating apps where every match is a real person, and reviews written by people who actually bought the thing. Whoever builds this becomes the layer every bank, app, and video call checks before it trusts anyone. If youre working on this, wed love to hear from you.
AI-Native Compliance Infrastructure#
By Daivik Goel
Financial compliance is still stitched together with spreadsheets, siloed tools, and expensive headcount. Companies hire chief compliance officers and assemble stacks of point solutions just to understand whats happening under the hood. As businesses expand into new markets, the complexity compounds and the cost of staying compliant grows faster than revenue. This is an AI-native problem. Most compliance work is monitoring regulatory changes, flagging anomalies, generating reports, and keeping audit trails. These are tasks that AI can handle faster and cheaper than humans. Yet most solutions today are still built around manual workflows and human review bottlenecks. The pain is especially acute for businesses navigating state-by-state licensing, renewal cycles, audits, and regulatory patchwork that varies widely across jurisdictions. The current process is slow, fragmented, and expensive, often requiring dedicated legal teams just to maintain what you already have. Were looking for founders building compliance infrastructure that consolidates fragmented tools, reduces reliance on specialized headcount, and gives finance teams real-time visibility across regulatory regimes. The best version of this doesnt just automate existing processes, but rethinks what compliance operations look like when AI is the default. The companies that get this right will become essential infrastructure for any business operating globally. If youre building in compliance infrastructure, please apply!
Self-Maintaining APIs#
By Harsha Gaddipati
Over the past year, Ive worked with over 50 API vendors, mostly early-stage startups. One pattern is consistent: API communication is broken. Breaking changes ship with little warning. Useful features quietly launch and go unnoticed. Changelogs dont get read. Heck, when I worked at AWS, over 30% of our service downtime was due to external api/package changes going unnoticed. This friction made sense before agentic coding tools existed. However, now it doesnt. Agentic coding tools like Claude Code, Devin, Greptile, etc prove that developers and enterprises are willing to give codebase access to external tools, provided theyre valuable. Two years ago, this was unthinkable. Now its standard practice. The infrastructure for automated code changes exists. Whats missing is the application layer connecting API providers to their customers codebases. API providers shouldnt just announce changes; they should apply them. When Stripe ships a breaking change or a new feature, an agent should scan customer codebases, identify affected usages, and open a PR with the fix. This could work as per-provider agents. Install Stripes update agent, or as a neutral third-party service tracking changes across vendors, like Dependabot but for APIs. If youre working on this, consider applying to YC.
Footer
Y Combinator
Make something people want.