TL;DR
Y Combinator’s Summer 2026 startup solicitation lists 15 categories, eight of which require hardware or capital, including agricultural robots, drone defense, space inference chips, lunar manufacturing, and semiconductor supply chain software. The document represents the most dramatic shift in YC’s public investment thesis, noting that the accelerator that built its reputation on software now believes the next decade of multi-billion dollar results will come from AI applied to physical, regulated and capital-intensive industries.
Y Combinator published its Summer 2026 Startup Application End of Aprila few days before the application deadline. The document lists 15 categories of companies that YC partners want to finance. Eight of them require capital, hardware, or both. The list includes AI for low-pesticide agriculture, defense against drone swarms, inference chips for space, lunar manufacturing from molten regolith and semiconductor supply chain software for a process that crosses a dozen countries and takes five months to complete.
Each category is written by a designated partner, and each reads less like a starting tip than a thesis on why the economics of a particular industry just changed. The world’s most influential startup accelerator, the institution that funded Airbnb, Stripe, and Dropbox, is telling its founders that the next decade of multibillion-dollar results will come not from building software, but from using AI to enter physical, regulated, capital-intensive industries that software alone never touched.
the thesis
The RFS begins with Garry Tan, CEO of YC, writing about agriculture. He argues that AI can now identify individual weeds and pests in real time, and when combined with precision robotic treatments, the result is agriculture that uses dramatically fewer pesticides while improving yields. The category is not agtech in the way Silicon Valley has historically understood it, which meant software dashboards for farm management. It is an agricultural technology that involves building physical robots, training vision models with biological data, and deploying hardware in the field.
Tyler Bosmeny’s post on swarm defense compares the companies he wants to fund to Cloudflare instead of Raytheon, software-defined defense systems that neutralize drone swarms at a fraction of the cost of traditional missile systems. The United States Department of Defense proposed more than $70 billion for drone and counterdrone systems in its latest spending plan, and Defense technology is experiencing its strongest investment cycle in decades.. Adi Oltean is looking for founders who will 3D print structures from molten lunar regolith and extract raw materials such as silicon, aluminum, iron and titanium through electrolysis on the moon.
Hard tech categories are not aspirational filler. They reflect a structural change in what venture capital is willing to finance. Defense technology startups raised a record $49.1 billion in 2025, nearly double the previous year. Anduril, the autonomous weapons company, raised $4 billion at a $60 billion valuation in March. SpaceX has shown that hardware-intensive companies can generate returns at risk scale. The old assumption that hardware couldn’t generate the margins or speed that venture capital requires has collapsed, and YC’s RFS is the clearest institutional recognition that the collapse is permanent.
The remaining software
Seven of the 15 categories are software, but none resemble the SaaS playbook that defined the previous decade. The category YC calls Software for Agents asks founders to rebuild every major software category for a world where the next billion users are not people but AI agents. That means APIs, machine-readable documentation, command-line interfaces, identity systems, permissions layers, and payment infrastructure designed for autonomous programs instead of humans. Google renamed its entire AI platform around agents in Cloud Next 2026consolidating Vertex AI into the Gemini Enterprise Agent platform and launching a $750 million fund to finance agent deployments. Gartner predicts that 40 percent of enterprise applications will include AI agents for specific tasks by the end of this year, up from less than 5 percent in 2025.
The Company Brain category calls for a system that extracts knowledge from every fragmented source within a company, structures it, keeps it up to date, and turns it into what YC calls an executable skills archive 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 rebuild software so that agents can operate it natively rather than removing interfaces built for humans.
The SaaS Challengers category explicitly names the targets: ERP, chip design software, industrial control systems and supply chain management. These are the categories where traditional vendors charge more and innovate less, and where AI-native replacements could capture huge markets if they can offset switching costs.
The physical turn
The RFS entry in semiconductor supply chains It may be the most revealing. A single advanced AI chip goes through approximately 1,400 process steps, crosses a dozen countries and takes five months to manufacture. That supply chain is managed, as the RFS says, with spreadsheets, SAP and phone calls. Diana Hu, the YC partner who wrote the post, calls for founders to replace that infrastructure with software that can track, optimize and predict the world’s most complex manufacturing process.
The category sits at the intersection of all the forces currently reshaping the tech industry: chip export controls between the United States and China, the reshoring of semiconductor manufacturing, the explosion in demand for AI chips, and the geopolitical fragility of supply lines that route critical components through Taiwan, South Korea, the Netherlands and Japan.
Spatial categories are also based on economics rather than aspiration. SpaceX and Stoke Space’s reusable rockets are about to deliver a massive increase in the ability to put objects into orbit, which means an equally massive increase in demand for the electronics that operate there.
YC wants inference chips optimized for mass, thermal performance and radiation hardness. SpaceX and Blue Origin are already racing to put data centers into orbitand AI hardware running inference workloads in terrestrial data centers does not survive the thermal and radiation environment of space. The market for spatially classified inference silicon does not yet exist. YC is betting that it will be so.
what a change
Y Combinator’s Spring 2026 RFS, released just three months earlier, listed eight categories. The summer edition nearly doubled that number to 15. The spring list included AI for product management, government AI, AI-native hedge funds, and stablecoins. You can recognize that these are software companies with integrated AI. The summer list includes the manufacture of lunar regolith and anti-drone defense systems. The change between the two documents is the most dramatic reorientation in YC’s public investment thesis since the accelerator began publishing applications from startups.
The change reflects what has happened to venture capital more generally. In the first quarter of 2026, $297 billion flowed into startups globally, 2.5 times more than the previous quarter and the largest venture funding ever recorded in a three-month period. Accel raised a $5 billion fund on the back of Anthropic and Cursor returns.
Andreessen Horowitz raised $15 billion. Thrive Capital closed over $10 billion. The money isn’t looking for the next SaaS enterprise dashboard. It is looking for companies that apply AI to industries where margins are higher, incumbents are slower, and barriers to entry have historically been physical rather than digital. YC’s RFS is the most explicit version of that thesis because it names industries by name: agriculture, defense, space, semiconductors, medicine, manufacturing.
The stablecoin category, one of the few holdovers from the spring list, reveals a different kind of ambition. YC describes stablecoins as falling between the regulated and unregulated worlds, creating space for services that combine the strengths of both: accounts that generate yield, real-world tokenized assets, and infrastructure that moves money faster and cheaper across borders. The AI Personalized Medicine category calls for agents to analyze genomic data, electronic medical records, and wearable outcomes to generate patient-specific treatment protocols rather than population-level guidelines. Neither category requires building physical hardware. Both require operating in industries where regulation, accountability and institutional trust are the barriers, not the code.
the signal
YC’s business formation application is not a prediction. It’s 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 counter-drone systems, and Adi Oltean writes about 3D printing on the moon, they’re telling founders what the next batch of YC will look like. The document is the closest thing the startup ecosystem produces to a forward-looking investment mandate from its most influential institution.
The mandate says that software is now the substrate, not the moat. Models are becoming commoditized. The infrastructure is scaling. Interfaces are being rebuilt for agents. What remains scarce is the ability to apply that substrate to the physical world: build the robot that replaces the pesticide, the chip that survives radiation, the defense system that costs less than the drone it destroys, the supply chain software that tracks 1,400 process steps in 12 countries, the molecular model that designs a drug for a target that the industry considers non-drug.
Y Combinator built its reputation by funding two founders on a garage writing code. Their summer 2026 RFS is a document that says the garage is no longer enough. The founders you want now are the ones who can write the code and then build it.
