TL;DR
Y Combinator’s 2026 Summer Request for Startups lists 15 categories, eight of which require hardware or capital, including agricultural robots, anti-drone defense, space inference chips, moon manufacturing and semiconductor supply chain software. The document represents the most dramatic point in YC’s public investment thesis, and shows that the accelerator, which built its reputation on software, now believes the next decade’s billion-dollar results will come from artificial intelligence applied to physical, regulated and capital-intensive industries.
Published by Y Combinator Request for Summer Starts 2026 at the end of Aprila few days before the application deadline. The document lists 15 categories of companies that YC’s partners are looking to fund. Eight of them require capital, hardware, or both. The list includes artificial intelligence for low-pesticide agriculture, drone defense, inference chips for space, manufacturing the Moon from molten regolith, and a semiconductor supply chain program for a process that crosses dozens of countries and takes five months to complete.
Each category is written by a named partner, and each reads less like a startup inquiry than a thesis about why the economics of a particular industry have just changed. The world’s most influential startup accelerator, the funder of Airbnb, Stripe and Dropbox, is telling founders that the next decade’s billion-dollar results will come not from building software, but from using artificial intelligence to enter physical, regulated and capital-intensive industries.
Thesis
RFS opens with YC CEO Garry Tan writing about agriculture. According to him, artificial intelligence can now identify individual weeds and pests in real time, and when combined with robotic precision treatments, the result is farming that uses dramatically less pesticides while improving yields. The category is not agtech as Silicon Valley has historically understood it, which meant software dashboards for farm management. It is agtech that involves building physical robots, training vision models on biological data, and deploying equipment in fields.
Tyler compares the companies Bosmeny wants to fund access to countermeasures to Cloudflare, not Raytheon. The US Department of Defense proposed more than $70 billion for drone and anti-drone systems in its latest spending plan, and defense technology is experiencing its strongest investment period in decades. Adi Oltean is asking founders to 3D print structures from molten lunar regolith and extract raw materials including silicon, aluminum, iron and titanium through electrolysis on the moon.
Hard technology categories are not aspirational fillers. They reflect a structural shift in what venture capital is willing to finance. Defense technology startups raised a record $49.1 billion in 2025, nearly double the previous year. Autonomous weapons company Anduril raised $4 billion in March at a $60 billion valuation. SpaceX has demonstrated that venture-scale businesses that require equipment can be profitable. The old assumption that hardware can’t generate the margins or momentum that venture capital demands has collapsed, and YC’s RFS is the clearest institutional acknowledgment that the collapse is permanent.
The rest of the software
Seven of the 15 categories are software, but none of them resemble the SaaS playbook that defined the previous decade. The YC category is called Software for Agents, asking founders to reinvent every major software category for a world where the next trillion users are AI agents, not humans. This means APIs, machine-readable documentation, command-line interfaces, authentication systems, permission layers, and payment infrastructure designed for autonomous applications, not humans. In Google Cloud Next 2026, it rebranded its entire AI platform around agentsIntegrating Vertex AI into the Gemini Enterprise Agent Platform and launching a $750 million fund to fund agent deployments. Gartner predicts that 40 percent of enterprise applications will include custom AI agents by the end of this year, up from less than 5 percent in 2025.
The Enterprise Brain category requires a system that extracts knowledge from every fragmented source within the company, structures it, keeps it current, and turns it into YC’s file of executable skills for AI. This is not a business search. It’s a living map of how a company works: how refunds are processed, how price exceptions are decided, how engineers respond to incidents. The Dynamic Software Interfaces category is its mirror image, asking founders to refactor their software so that agents can run it natively instead of breaking interfaces built for humans.
The SaaS Challengers category clearly names the targets: ERP, chip design software, industrial control systems and supply chain management. These are the categories where incumbent suppliers charge the most and innovate the least, and where local replacements with artificial intelligence can capture large markets if they can eliminate switching costs.
Physical return
RFS access semiconductor supply chains may be the most obvious. A single advanced AI chip goes through about 1,400 process steps, crosses dozens of countries and takes five months to manufacture. This supply chain is managed by spreadsheets, SAP and phone calls, as RFS says. YC partner Diana Hu, who wrote the introduction, asks founders to replace this infrastructure with software that can monitor, optimize and predict the most complex manufacturing process on earth.
The category is at the intersection of all the forces currently reshaping the tech industry: US-China chip export controls, a restructuring of semiconductor manufacturing, an explosion in AI chip demand, and the geopolitical fragility of supply lines that carry critical components through Taiwan, South Korea, the Netherlands and Japan.
Space categories are similarly based on economics rather than desire. SpaceX and Stoke Space’s reusable rockets are poised to massively increase their ability to put objects into orbit, which means an equally massive increase in demand for the electronics that work there.
YC wants output chips optimized for mass, thermal performance and radiation hardness. SpaceX and Blue Origin are already competing to put data centers into orbitand AI hardware running inferential workloads in ground data centers do not survive the heat and radiation environment of space. A market for silicon valued in space does not yet exist. YC is betting it will.
What has changed
Y Combinator’s Spring 2026 RFS, published just three months ago, listed eight categories. The summer edition nearly doubled that to 15. The spring list includes AI for product management, government artificial intelligence, AI-native hedge funds and stablecoins. These are software businesses that are recognizably tied to AI. Summer’s list includes lunar regolith production and anti-drone defense systems. The transition between the two documents is the most dramatic shift in YC’s public investment thesis since the accelerator began publishing requests for startups.
The change reflects what’s happening to venture capital more broadly. Global startups received $297 billion in the first quarter of 2026, up 2.5 times from the previous quarter and the most venture funding ever recorded in a three-month period. Accel raised $5 billion in funding Behind returns from Anthropic and Cursor.
Andreessen Horowitz raised $15 billion. Thrive Capital closed over $10 billion. Money isn’t looking for the next enterprise SaaS dashboard. It looks for companies to apply AI to industries where margins are the highest, incumbents are the slowest, and entry barriers have historically been physical rather than digital. YC’s RFS is the most obvious version of this thesis, as it names industries by name: agriculture, defense, aerospace, semiconductors, medical, manufacturing.
One of the few holders on the spring list, the stablecoin category reveals a different ambition. YC describes stablecoins as sitting between the regulated and unregulated worlds, making room for services that combine the strengths of both: leveraged accounts, tokenized real-world assets, and infrastructure that moves money across borders faster and cheaper. The Artificial Intelligence Personalized Medicine category calls for agents that analyze genomic data, electronic health records and wearables to create patient-specific treatment protocols rather than population-level guidelines. Neither category requires building physical equipment. Both require working in industries where regulation, accountability and institutional trust are obstacles, not codes.
Alarm
YC’s Survey for Startups is not a forecast. It is a commitment. The partners who write the entries are the partners who will evaluate the applications and the categories they describe are the companies they will fund. When Garry Tan writes about agricultural robots, Tyler Bosmeny writes about anti-drone systems, and Adi Oltean writes about 3D printing on the moon, they tell founders what the next batch of YC will look like. The document is the closest thing the startup ecosystem has produced to a forward-looking investment mandate from the most influential body.
Mandate says that software is now the substrate, not the trench. Models are commercialized. Infrastructure is expanding. Interfaces are rebuilt for agents. What remains scarce is the ability to apply that substrate to the physical world: a robot that replaces a pesticide, a chip that survives radiation, a defense system that costs less than the drone it destroys, a supply chain program that tracks 1,400 process steps in 12 countries, a molecular model that designs drugs for a target industry.
Y Combinator built its reputation by funding two founders in the garage writing code. Its 2026 Summer RFS is a document that says the garage is no longer sufficient. What he wants now are founders who can write code and then build things.
