Does Y Combinator Steal Ideas From Applications?

YC applications force you to be specific. You describe the problem, the wedge, your unfair advantage, early traction, and sometimes the exact tactic working. For a paranoid first-time founder, handing that to anyone — let alone the most powerful accelerator in the world — feels risky.

Then add the surface-level red flags: YC funds ~250 companies per batch, several often look similar, and post-batch you might see a YC-funded company doing 'your' idea. The conclusion looks obvious. It isn't.

1. The business model. YC makes money when its portfolio is worth more. Stealing one idea from an applicant would torch the inbound flywheel that produces the next 10,000 applications. The math is catastrophically bad.

2. Idea overlap. YC partners read every batch's worth of applications. Most batches have 5–20 teams attacking the same wedge. Convergent thinking is the norm, not theft.

3. Execution > idea. Sam Altman, Paul Graham, and Garry Tan have all said publicly: ideas are cheap, founders are rare. YC bets on people who can outrun copycats — including copycats they fund.

4. Reputation cost. If even one credible 'YC stole my idea' story went viral with proof, founder applications would drop 50% overnight. The asymmetry is enormous.

Applications are read by 2–4 partners. Rejected applications stay in YC's internal system but are not shared externally, not sold, and not used to brief other founders. There is no public 'idea library.'

Partners may, in office hours with funded companies, mention 'we've seen others try X' — but this is pattern recognition across thousands of apps, not directing one team to copy another.

If you genuinely have proprietary tech (patentable IP, a trade-secret algorithm), don't write it in the application. Describe what it does, not how. YC has never required protocol-level disclosure.

For everything else — the market, the wedge, the customer — write the whole truth. Vague applications get rejected. Specific ones get interviews.