YC’s Summer 2026 RFS bets on agricultural robots, drone defense, and lunar production as software loses its groove

IY Combinator published its application for Summer 2026 Startups in late April, a few days before the application deadline. This document lists 15 categories of companies that YC partners want to fund. Eight of them require capital, hardware, or both. The list includes AI for low-pesticide agriculture, counter-swarm drone defense, space inference chips, lunar production from molten regolith, and semiconductor supply chain software for a process that crosses twelve countries and takes five months to complete.

Each section is written by a named colleague, and each one reads less like an introduction than a thesis on why the economics of a particular sector have recently changed. The world’s most influential startup accelerator, the institution that funded Airbnb, Stripe, and Dropbox, is telling founders that the next decade of multibillion-dollar results won’t come from building software but from using AI to penetrate physical, regulated, and capital-intensive industries that software alone has never touched.

The thesis

RFS opens with Garry Tan, CEO of YC, writing about agriculture. AI can now identify individual weeds and insects in real time, he argues, and when combined with precision robotic treatments, the result is farming that uses far less pesticide while improving yields. The category is not agtech as Silicon Valley has historically understood it, meaning farm management software dashboards. It is agtech that involves building portable robots, training vision models on biological data, and deploying hardware in fields.

Tyler Bosmeny’s foray into counter-swarm defense compares the companies he wants to fund to Cloudflare to Raytheon, software-defined defense systems that reduce 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 counter-drone programs in its latest spending plan, and defense technology is facing its strongest investment cycle in decades. Adi Oltean is asking inventors to print 3D structures in molten lunar regolith and extract raw materials including 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 are expected to raise a record $49.1 billion by 2025, nearly double the previous year. Anduril, a private arms company, raised $4 billion at a $60 billion valuation in March. SpaceX has shown that hardware-intensive businesses can generate returns of scale. The old assumption that hardware could not generate the margins or speed required by venture capital has collapsed, and YC’s RFS is a clear admission by the institution that the collapse is permanent.

The remaining software

Seven of the 15 categories are software, but none of them resemble the SaaS playbook that defined the last decade. The YC category calls for Software for Agents is asking innovators to reinvent all of the world’s major software categories where the next billion users are not humans but AI agents. That means APIs, machine-readable scripts, command-line interfaces, identity systems, permission layers, and payment infrastructure designed for autonomous systems rather than humans. Google renamed its entire AI platform to Cloud Next 2026 agents, integrated Vertex AI into the Gemini Enterprise Agent Platform and launched a $750 million fund to fund agent deployments. Gartner predicts that 40 percent of business applications will include task-specific AI agents by the end of this year, up from less than 5 percent by 2025.

The Corporate Brain category calls for a system that extracts information from all disparate sources within the company, organizes it, stores it, and converts it into what YC calls an actionable AI skills file. This is not a business search. It’s a living map of how the company works: how refunds are made, how pricing exceptions are determined, how engineers respond to incidents. The Dynamic Software Interfaces category is its mirror image, asking developers to redesign software so agents can use it natively rather than scratch human-built interfaces.

The SaaS Challengers category has clear targets: ERP, chip design software, industrial control systems, and supply chain management. These are the segments where incumbent vendors charge more and innovate less, and where AI replacements can capture large markets if they can wipe out switching costs.

Body turning

The inclusion of RFS in semiconductor supply chains can be very revealing. One advanced AI chip goes through about 1,400 process steps, crosses a dozen countries, and takes five months to produce. That supply chain is managed, as RFS puts it, through spreadsheets, SAP, and telephones. Diana Hu, the YC partner who wrote the event, calls on the innovators who will replace that infrastructure to install software that can track, optimize, and predict the entire world’s most complex manufacturing process.

The division sits at the intersection of all the forces currently reshaping the tech industry: US-China chip export controls, the restructuring of semiconductor manufacturing, the explosion of AI chip demand, and the geopolitical fragility of supply lines that move key components to Taiwan, South Korea, the Netherlands, and Japan.

Space classes are similarly focused on economics rather than aspiration. The reusable rockets from SpaceX and Stoke Space are about to produce a huge increase in the power to put things into orbit, which means a correspondingly huge increase in the demand for electronics that work there.

YC is looking for inference chips optimized for size, thermal performance, and radiation hardness. SpaceX and Blue Origin are already racing to put data centers in orbit, and the AI ​​hardware that runs workloads in terrestrial data centers doesn’t survive the heat and radiation. A space-limited silicon market does not yet exist. YC is betting it will.

What has changed

IY Combinator’s Spring 2026 RFS, published just three months earlier, outlined eight categories. The summer program nearly doubled that to 15. Spring’s list included AI for product management, AI for government, AI-native hedge funds, and stablecoins. Those are virtual software businesses with AI on board. Summer’s list includes lunar regolith production and anti-drone defense systems. The transition between the two documents is the most dramatic reorientation of YC’s public investment thesis since the accelerator began publishing startup applications.

The change reflects what has happened to business finance in general. In the first quarter of 2026, 297 billion dollars came in globally, 2.5 times more than the previous quarter and the most financing ever recorded in a three-month period. Accel raised $5 billion in funding after returns from Anthropic and Cursor.

Andreessen Horowitz raised $15 billion. Thrive Capital has closed more than $10 billion. Money isn’t looking at the next SaaS dashboard for the business. It looks for companies to use AI in industries where margins are high, stakeholders are slow, and barriers to entry have historically been physical rather than digital. YC’s RFS is a more specific version of that thesis because it names industries by name: agriculture, defense, space, semiconductors, pharmaceuticals, manufacturing.

The stablecoin category, one of the few securities on the Spring list, represents a different kind of ambition. YC defines stablecoins as sitting between the regulated and unregulated worlds, creating an environment for services that combine the power of both: yield-generating accounts, tokens for real-world assets, and infrastructure that moves money quickly and cheaply across borders. The AI ​​personalized medicine category calls for agents that analyze genomic data, electronic health records, and wearable outputs to generate patient-specific treatment regimens rather than population-level guidelines. Neither class needs to build physical hardware. Both need to work in industries where regulation, credit, and institutional trust are constraints, not code.

The signal

The YC Startup Application is not a prediction. It is an obligation. The partners who write the entries are the partners who will evaluate the applications, and the categories they describe are the companies they will sponsor. When Garry Tan writes about agricultural robots and Tyler Bosmeny writes about counter-drone systems and Adi Oltean writes about 3D printing on the moon, they tell the founders what the next batch of YC will look like. This document is the closest thing produced by the first ecosystem to the authorization of forward investment from its most influential institution.

The mandate states that software is now a substrate, not a moat. Models are for sale. Infrastructure is growing. Interfaces are rebuilt by agents. What remains is the ability to use that substrate in the virtual world: to build a robot that replaces a pesticide, a chip that survives radiation, a defense system that costs less than a drone that destroys it, a supply chain software that tracks the steps of a 1,400 process in 12 countries, a molecular model that designs a drug that is targeted by the industry called undruggable.

IY Combinator made its name by funding two inventors writing code from a garage. The Summer 2026 RFS is a document that says the garage is no longer enough. The inventors you want now are the ones who can write code and build something.