Y Combinator has published its Summer 2026 Requests for Startups, outlining 15 categories spanning AI, hardware, defense, agriculture, and space that the accelerator believes represent the most solvable gaps in the market today. The batch will run from July through September in San Francisco, with YC investing $500,000 in every accepted company on standard terms.

This is not a list of trends. According to YC's announcement, "AI has stopped being a feature and started being the foundation." Several ideas came directly from active YC founders describing what they are seeing on the frontier. The 15 categories reveal where the accelerator believes the next wave of billion-dollar companies will emerge.

What Are the 15 Categories Y Combinator Wants Founders to Build?

YC's wishlist spans a diverse range of opportunities, each addressing a specific market gap. The categories include AI-native services that replace human labor entirely, replacements for legacy software incumbents, company knowledge systems, agent-first infrastructure, closed-loop intelligence systems, autonomous scientific discovery, personalized genetic medicine, AI-powered software customization, and semiconductor supply chain optimization.

The opening category focuses on what YC partner Gustaf Alströmer calls a fundamental economic opportunity. The global spend on services dwarfs the spend on software, and most of those services are already outsourced, making them structurally easy to replace with AI. Rather than selling software tools to help people do work, YC wants founders to become the service provider themselves. Target sectors include insurance, accounting, tax, audit, compliance, and healthcare administration.

• AI-Native Services: Companies that do the work themselves rather than selling software to help humans do it, targeting insurance, accounting, tax, audit, compliance, and healthcare administration.
• Legacy Software Replacement: AI-native alternatives to entrenched incumbents like Salesforce, targeting chip design software, enterprise resource planning systems, and industrial control systems that have been functionally untouchable for 20 years.
• Company Brain Systems: Platforms that pull knowledge from fragmented sources like Slack, email, and support tickets, structuring it into executable files for AI agents to act on safely and consistently.
• Agent-First Infrastructure: APIs, Model Context Protocols, and command-line interfaces designed specifically for AI agents rather than humans, enabling agents to discover and use new tools without human intervention.
• Closed-Loop Intelligence: Systems that make entire company operations queryable, turning organizations from open-loop decision-making into real-time monitoring and adjustment cycles.
• Autonomous Scientific Discovery: Systems that propose hypotheses, run experiments, analyze results, and iterate without constant human direction in fields like drug discovery and materials science.
• Personalized Genetic Medicine: Ecosystems combining diagnostic tooling, therapy delivery, and intelligence layers that generate user-specific treatment recommendations from genome scans and wearable data.
• AI-Powered Software Customization: Coding agents capable of giving every user personalized interfaces, potentially shipping source code instead of packaged binaries.
• Semiconductor Supply Chain Optimization: Real-time allocation tracking, multi-tier risk monitoring, and export compliance automation for chip manufacturing and distribution.

Why Is Y Combinator Pushing Founders Toward These Specific Ideas?

YC partner Jared Friedman makes a direct case for why legacy software replacement represents the biggest startup opportunity in a decade. AI has collapsed the cost of producing software by 10 to 100 times. The moat that once protected companies like Salesforce, with millions of lines of code built over decades, is now gone. The real prize is not easy targets like project management tools, but rather chip design software, enterprise resource planning systems, and industrial control systems.

"Every company has critical knowledge scattered across Slack threads, old emails, support tickets, and people's heads. Humans navigate this vaguely. AI agents cannot," explained Tom Blomfield, who founded Monzo.

Tom Blomfield, Founder of Monzo

Blomfield identifies what he calls the missing primitive of the AI era. A "company brain" would pull knowledge from all these fragmented sources, structure it, keep it current, and turn it into an executable file for AI agents. The example he gives is how a company handles refunds, what the rules are for pricing exceptions, and how engineers respond to incidents. All of that should be legible to an AI system that can then act on it safely and consistently.

Aaron Epstein opens with a striking observation: the next trillion users on the internet will not be people. They will be AI agents. And almost no software is built for them. Agents are already browsing the web, doing research, and managing workflows, but they are doing it on top of interfaces designed for humans clicking buttons, which is slow, brittle, and inconsistent.

How Should Founders Approach These Opportunities?

YC partner Diana Hu describes what she is seeing in the best AI-native companies: they have made their entire operation queryable. Every meeting is recorded, every ticket tracked, every customer interaction captured and legible to an intelligence layer that learns from it. This turns a company from an open loop, make a decision, check results weeks later, into a closed loop that monitors, compares, and adjusts in real time. She has seen teams using this approach cut sprint time in half.

For founders interested in scientific discovery, YC is betting on a shift from AI research assistants to closed discovery loops. Systems that propose hypotheses, run experiments, analyze results, and iterate without constant human direction are already beginning in drug discovery, materials science, and protein engineering. Intelligent systems are starting to run full design-make-test-analyze cycles.

In the healthcare space, YC partner Ankit Gupta argues that an agent harness can now take a patient's genome scan and wearable data and generate highly accurate, user-specific treatment recommendations. Five years ago this was science fiction. Today, it is engineering. YC sees an entire ecosystem of startups needed with diagnostic tooling, therapy delivery, and the intelligence layer connecting them.

The semiconductor supply chain represents another massive opportunity. A single advanced AI chip crosses roughly 1,400 process steps, passes through a dozen countries, and takes five months to manufacture. This supply chain is still managed with spreadsheets, legacy enterprise software, and phone calls. With the CHIPS Act standing up entirely new American fabs in Arizona, Texas, Ohio, and New York, each needs a supply chain built from scratch. Almost none of the expected tooling exists.

YC's Summer 2026 batch represents a clear signal about where the accelerator believes the next generation of AI-powered startups should focus. Rather than building better versions of existing products, the accelerator is encouraging founders to identify entire categories of work that AI can now do end-to-end, and to build companies that own the outcome rather than just the software.