Y Combinator has released its Summer 2026 Request for Startups with a noticeable transition away from placing AI use as its top criterion. The world’s most influential startup accelerator is now asking founders to rebuild software, services, and more from the ground up.
“AI has stopped being a feature and started being the foundation. We’re excited about a new wave of startups rebuilding software, services, and silicon, and pushing AI into the physical world,” says the Y Combinator team.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
The timing is important here, too. While the AI hype cycle peaked in 2023-2024 with chatbot wrappers and “copilot” features, YC’s Summer 2026 list highlights a maturation. The accelerator is no longer interested in AI as an add-on. It wants companies building AI-native products that fundamentally rethink how work gets done, how services get delivered, and how physical infrastructure operates.
For founders in smaller ecosystems like Cyprus, where early-stage capital and institutional accelerator programs remain limited, the RFS provides a glimpse into what top accelerators believe the next wave of transformative companies will look like and what problems founders should be solving at present in order to access support and funding. YC’s program primarily targets U.S. startups, but Cypriot founders working on AI infrastructure, enterprise software, hardware, or specialized services can use this list as a strategic guide to international markets and emerging opportunities.
The 15 Ideas That Made the List
This batch is split between three major themes. AI is transforming services and software. Hardware and physical infrastructure. Enterprise reinvention.
AI-Native Services and Software
In no particular order, the first request is within the AI-native services and software market for companies that don’t sell software. This category is in partnership with Gustaf Alströmer. These companies sell the finished service. Instead of an accounting tool, they want startups that just do your accounting. The total spend on services is many times larger than the spend on software, and many services are already outsourced, making them easier to replace. YC specifically gives the example of insurance brokerage, accounting, tax, audit, compliance, and healthcare administration.
Company Brain takes that concept internally with Tom Blomfield. YC defines the “Company Brain” as a living system of how a company works, capturing every meeting, customer interaction, and ticket. This system would turn a company’s scattered knowledge into an “executable skills file” that AI agents can use to run operations. Think of it as making an entire company queryable and legible to AI.
Software for Agents changes the infrastructure layer. According to partner Aaron Epstein, “The next trillion users on the internet won’t be people, they’ll be AI agents.” Instead of visual interfaces, those agents need machine-readable interfaces like APIs, MCPs, and CLIs., as well as thorough documentation to allow them to discover, sign up for, and start using new tools instantly, without a human in the loop. Every major software category, therefore, will need to be rebuilt explicitly for agents as first-class citizens.
Dynamic Software Interfaces goes even further. Before AI, every user of a piece of software saw the same interface. YC wants startups to rethink that concept entirely. According to partner Ankit Gupta, “We think that coding agents have now gotten good enough to allow users to become their own forward-deployed engineers and more radically customize the software they consume.” Imagine an email client that looks like a task list for one person and an events calendar for another. Coding agents would allow users to customize interfaces on the fly, but in order to achieve this, the whole stack of software delivery will need to be reconsidered.
SaaS Challengers is going for the big, hard SaaS categories. While investors have wiped trillions off software market caps on fears that AI coding will kill SaaS, Jared Friedman, from YC, argues this creates the biggest startup opportunity. If AI has collapsed the cost of producing software by 10-100x, legacy SaaS products are suddenly vulnerable. ERPs, chip design software, and industrial control systems. The moat is gone.
AI for Low-Pesticide Agriculture addresses a pressing problem. Modern agriculture runs on chemicals, but pesticide residues are everywhere, and pests are evolving resistance. YC, on behalf of Garry Tan, wants startups combining AI vision, precision robotics, and biological solutions to cut pesticide use by 90% while helping farmers grow more food. Biological solutions include microbes, peptides, and RNA-based treatments. The company that does this isn’t just a good business but a generational one.
AI Personalized Medicine leverages intelligent agents to analyze genomic scans, wearables data, and diagnostics to deliver hyper-personalized care. As the cost of genome sequencing and personalized genetic therapies plummets, Ankit Gupta sees a revolution in care delivery where abundant data and intelligence democratize access to treatments for serious illnesses.
Hardware and Physical Infrastructure
Counter-Swarm Defense tackles an urgent military problem. Tyler Bosmeny wants startups building the counter-swarm stack, high-capacity interceptors that neutralize fifty drones, not one. Software that fuses sensors into real-time defense systems. Non-kinetic defenses like aerosols or streamers. As Tyler Bosmeny writes, “Drone defense is looking less like operating a weapon and more like running a real-time distributed system. The winning companies will look more like Cloudflare than Raytheon.”
Hardware Supply Chain addresses a bottleneck for every hardware startup that YC funds, according to Nicolas Dessaigne. In Shenzhen, teams go from design to physical part in a day. In the US, that loop takes weeks. YC wants startups building the infrastructure to match China’s iteration speed. Tools that produce parts faster, enable rapid prototyping, and tightly integrate design, manufacturing, and logistics.
Electronics in Space and Industrial Capabilities in Space reflect SpaceX and Stoke Space recently making launches dramatically cheaper. YC, alongside Adi Oltean, wants inference chips optimized for mass, thermal constraints, and radiation to support the massive increase in compute capacity needed for orbital operations. On the industrial side, YC is calling for infrastructure to extract raw materials like silicon, aluminum, and iron from the moon and use 3D printing with molten regolith to build complex structures in space.
Inference Chips for Agent Workflows recognizes that current GPUs hit 30-40% utilization on agent workloads. Agents loop, calling tools, branching, backtracking, and holding context across dozens of steps. Diana Hu wants to see purpose-built silicon that handles fast context switching, native speculative decoding, and memory built for persistent KV caches, which could unlock incredibly large gains.
Enterprise AI and Operations
The AI Operating System for Companies wants to see every company’s operations legible to AI by default. As highlighted by Diana Hu, the best AI-native companies have made their entire company queryable. Their meetings are recorded, every ticket tracked, every customer interaction captured. This, ultimately, turns a company from an open loop into a closed loop where the system monitors, compares, and adjusts in real time.
Startups That Want to Sell to Huge Companies capitalizes on a trend that is highlighted by Harshita Arora and Brad Flora. For the first time, Fortune 100 companies are actively seeking AI solutions from startups. “In the last 3 years, for the first time ever, we have seen YC companies land pilots and actual multimillion dollar deals within their first year, if not during the actual YC batch,” say Harshita Arora and Brad Flora. AI has made it possible for 2-3 person teams to ship products to large orgs in months, not years.
Supply Chain 2.0 for Semiconductors takes on the 1,400-step global journey of advanced AI chips, currently managed with spreadsheets and phone calls. In 2021, a $300 chip held up a $50,000 car, and $210 billion in vehicles didn’t get built. Diana Hu wants real-time allocation tracking, multi-tier risk monitoring, and export compliance tooling. None of which exists today.
In the Context of Cyprus
Several themes stand out as particularly relevant for the Cypriot startup ecosystem.
Agriculture presents an immediate opportunity. Cyprus agriculture faces the same chemical dependency YC describes. Mediterranean pests, water scarcity, and rising input costs. A startup combining AI vision with precision application could pilot on Cyprus’s smaller farms before scaling to Southern Europe and beyond. The regulatory environment in the EU, while complex, is increasingly favorable to solutions that reduce chemical use.
Service company transformation is accessible. Cyprus already has a strong base of professional services. Accounting firms, insurance brokers, and so on serve both local and international clients. The move from selling software to selling AI-native services doesn’t require extensive technical infrastructure. A two-person team can build an AI-native accounting service targeting small businesses and international companies using Cyprus as a base.
Defense tech is newly relevant. Cyprus is positioned favorably at a strategic crossroads and has ongoing security concerns, especially in consequence of recent political upheaval in the region. Cypriot startups with software, robotics, or systems integration expertise could find early customers in European defense procurement or dual-use applications in critical infrastructure protection.
Selling to an enterprise is viable. Cyprus-based startups can now sell directly to Fortune 500 companies piloting AI solutions. Geography is no longer a limiting factor as remote work has normalized. A team in Limassol can land a pilot with a major European or American corporation in their first year.
Talent arbitrage still exists. Cyprus offers lower operating costs than Western Europe or the US while maintaining access to EU markets, strong internet infrastructure, and an English-speaking workforce. This is especially attractive for hardware companies needing to iterate quickly or software companies building AI-native services.
Y Combinator’s Summer 2026 RFS is open now. Applications for the next batch close in late May. You can explore the full Requests for Startups list at https://www.ycombinator.com/rfs.
