YC Winter 2027 RFS · November 2026
YC's Winter 2027 startup requests, explained for founders
Y Combinator published its Winter 2027 Requests for Startups in late October 2026. The list moves away from pure software toward the physical, regulated, and public parts of the economy: launch, robots, power equipment, city hall, and AI that runs where the cloud can't reach. For each request, we cover what YC actually said, why the timing is now, three concrete starting points, and how to validate one before you build.
Based on YC's official Requests for Startups page.
The pattern
Compared with the 2026 lists, crypto, AI-native agencies, AI compliance, and proof of humanity are gone. In their place: the bottlenecks that appear once AI leaves the screen. Robots need domestic parts, data centers need transformers, physical assets need financing, and regulated buyers need AI that never leaves the building. These markets sell slower than SaaS, but the problems are concrete and the buyers have budgets.
Startups in a World of Abundant Space Access
by Ankit Gupta
YC is asking founders to build for a near future where getting mass to orbit is radically cheaper. The bet is that business models which make no sense at today's launch prices start to work once fully reusable rockets fly at high cadence.
Read founder breakdown →
The $15-an-Hour Expert
by Brad Flora
Instead of AI that replaces workers, YC wants AI that upgrades them: software that lets someone without a professional credential do work that used to need a specialist. The target is any field where the credential, not the work itself, is the bottleneck.
Read founder breakdown →
Robots Made in America
by Diana Hu
Robot software is improving fast, so the bottleneck has moved to hardware, and most of that hardware is imported. YC wants companies that build the American robot supply chain itself, not just assemble robots from foreign parts.
Read founder breakdown →
LLMs That Write Well
by Jared Friedman
As large language models got better at coding and math, their writing got worse: flatter, more robotic, and more uniform. YC thinks the big labs will keep optimizing for code and agents, which leaves room for a startup that builds a model people actually enjoy reading and talking to.
Read founder breakdown →
Plaid for Government Benefits
by Gustaf Alströmer
People applying for government benefits prove the same facts again and again to agencies that do not share data. YC wants a consent-based verification layer that pulls income, tax, and residency records from the source, so one verification works across many programs.
Read founder breakdown →
Wall Street for the Physical Buildout
by Brad Flora
Robots, GPU clusters, and data centers cost a lot up front, and the financing for them barely exists. YC wants the financial companies that the physical AI buildout needs: lenders, leasing companies, securitization, and hedging products built for these assets.
Read founder breakdown →
Vertical Foundation Models
by Diana Hu
YC wants research labs that build specialized foundation models for industries like manufacturing, energy, and biology. The models should work zero-shot or from a few examples, and match or beat general models while using far less compute.
Read founder breakdown →
AI for City Hall
by Gustaf Alströmer
Building permits take so long that some projects never get built, and city staff are buried in manual code checks and correction cycles. YC wants AI that clears that backlog, starting with permits and expanding to 311, public records, licensing, and inspections.
Read founder breakdown →
Data Center Supply Chain
by Jared Friedman
YC calls the AI data center buildout the largest capital project in US history and wants companies that unblock every part of its supply chain, from power equipment to cooling to the materials those depend on.
Read founder breakdown →
AI for the Air-Gapped Economy
by Brad Flora
Defense contractors, banks, hospitals, and nuclear plants keep their most valuable data offline and cannot use cloud AI APIs. YC wants companies that bring AI fully inside those buildings: on-prem models, deployment tools, security accreditation, and hardware.
Read founder breakdown →
The eleventh request: Own Your Intelligence (by Jon Xu). Helping companies own and improve specialized open-weight models instead of only renting frontier models: evals, training data from expertise, and carrying improvements to new base models. It overlaps with Vertical Foundation Models; read YC's original.
Idea valuation
Building toward one of these?
Paste your idea and see what 28,000 Hacker News threads say about its market, timing, technical risk, and distribution.
Your first evaluation is free with Google sign-in.
Earlier RFS breakdowns