Y Combinator
The 2025
batch
breakdown

YC in 2025, the biggest year in accelerator history

2025 was the year Y Combinator stopped being an accelerator and started operating like an institutional investor with a demo day problem. Four batches. Over 470 companies funded. An acceptance rate holding at approximately 1.5%. And a portfolio that now spans 5,600+ companies with a combined valuation exceeding $600 billion.

Since 2005, YC has produced 82 unicorns, companies valued at $1 billion or more. That's a 4% conversion rate from YC company to unicorn, significantly higher than the ~2.5% rate for comparable venture-backed seed-stage startups. The alumni network reads like a who's who of tech: Airbnb, Stripe, Coinbase, Dropbox, Instacart, DoorDash, Cruise, Deel, Rippling.

But 2025 was different. YC moved from two batches to four. They explicitly leaned into AI, not as a sector bet, but as the defining thesis of the entire programme. And they quietly became the most active fintech investor globally, participating in 151 fintech deals through the year, up 24.8% from 2024.

The survival rate is remarkable: approximately 87% of YC companies remain actively operating, compared to roughly 50% for typical startups after five years. And around 45% go on to raise a Series A, well above the 33% industry average.

Gian's Insider Take

The shift to four batches is the most significant structural change in YC's history. It means more companies get in, but it also means more competition at demo day, more noise in the alumni network, and arguably less per-company attention during the programme. If you're applying, this changes the calculus: batch selection now matters. Winter and Summer are the established batches with the deepest investor attendance at demo day. Spring and Fall are newer, smaller, and potentially less competitive to get into, but the investor networks are still maturing.

The four batches, size, timing & what changed

For the first time in its 20-year history, YC ran four distinct batches in a single calendar year. Here's how they broke down:

The trend is clear: batches are getting larger. W25 had ~160 companies. By F25, that had grown to 180+. YC is scaling the programme while maintaining roughly the same acceptance rate by processing more applications per cycle.

Gian's Insider Take

The Spring batch (X25) is the one to watch strategically. It's new, it's smaller (143 vs 180), and investors are still learning to pay attention to it. That means potentially less competition to get in AND less competition for follow-on capital at demo day. If you're applying to YC and don't need the prestige of Winter/Summer, the Spring and Fall batches could be your edge. It's the same programme, the same $500K investment, the same alumni network, just with fewer companies sharing the spotlight.

The AI takeover, 88% and counting

This is the number that defines YC in 2025: 88% of the S25 batch was classified as AI-native. Not "uses AI somewhere in the stack." AI-native, meaning the core product wouldn't exist without AI.

Of the S25 batch specifically: 141 startups were AI-native, 9 were AI-adjacent, and just 10 operated without any AI integration. Over 60% explicitly referenced AI in their one-liner pitch.

The flavour of AI has shifted too. 2024 was about AI copilots, tools that assist humans. 2025 is about AI agents, autonomous systems that replace workflows entirely. Over half the Spring 2025 batch was building agentic AI. By Summer, coding agents showed the strongest revenue traction of any AI agent category.

The AI Sub-Categories Dominating YC

A surprising number of S25 startups aren't building their own AI agents, they're building tools that enhance other companies' AI agents. Think: evaluation frameworks, observability, testing infrastructure. The AI tooling ecosystem is becoming a category in itself.

Gian's Insider Take

The 88% number is staggering, but don't let it scare you into forcing an AI angle that doesn't exist. I've seen founders bolt on AI language to their applications and it backfires, YC's partners can spot a cosmetic AI wrapper instantly. What matters is whether AI is genuinely core to your defensibility. If it is, lead with it hard. If it's not, own that and make the case for why your market doesn't need it. Authenticity beats trend-surfing at YC.

Sector-by-sector breakdown

Across all four 2025 batches, certain sectors dominated. Here's where YC placed its bets:

Developer Tools & Infrastructure (~30%)

The single largest category. Coding agents, DevOps automation, AI testing frameworks, observability tools. 20 software development companies in S25 alone, making it the biggest sub-sector in any individual batch. Coding agents showed the strongest revenue traction of any AI agent category, which explains why YC keeps funding them.

Healthcare & Biotech

A consistent presence across all batches. The F25 healthtech cohort focused on AI as the operational backbone of healthcare, drug discovery, clinical workflows, payer operations, pharmacy automation, women's health. Healthcare startups consistently raise larger seed rounds ($4.6M median vs $3.1M overall) reflecting longer timelines and regulatory requirements.

Fintech (15%+)

YC became the most active fintech investor globally in 2025, participating in 151 fintech deals, up 24.8% from 121 in 2024. In the Fall batch, 6 of the 16 fintech companies were deploying agentic AI tooling. The standout: Harper, an AI insurance brokerage from W25, which went on to raise $47M.

Defence & Dual-Use

A notable new trend. 5 defence-focused companies in S25, including startups building solutions for drone systems, compliance, and dual-use analytics. This would have been unthinkable at YC five years ago. The defence tech thesis is now firmly established.

Voice AI

16 voice AI companies in S25 alone, an entirely new category that barely existed at YC in 2023. These aren't basic chatbots. They're building enterprise-grade systems managing complex, regulated interactions, particularly in financial services, healthcare, and customer support.

B2B vs B2C Split

The B2B skew is extreme. Only about 15-20% of companies across the 2025 batches were consumer-facing. YC's thesis is clear: enterprise AI is where the money is, and the programme is backing companies that sell to businesses, not consumers.

Gian's Insider Take

The sector data reveals something founders miss: YC doesn't just back good companies, they back markets they believe are about to undergo structural change. Developer tools dominate because AI is rewriting how software gets built. Healthcare AI dominates because the US healthcare system is structurally broken and AI is the first technology that could actually fix the cost problem. Defence is there because geopolitics created a new market. When you apply to YC, you're not just pitching your company, you're pitching that your market is about to fundamentally shift. If you can't articulate the structural change driving your sector, you're at a disadvantage.

Geography, where founders are coming from

YC remains overwhelmingly American, but the international picture is shifting. Here's what the 2025 data shows:

Within the US, there's been a dramatic geographic concentration. San Francisco's share of YC companies rebounded from 21% in 2021 to 73% in 2026, the highest concentration in YC's history. The remote-first experiment is over. YC founders are moving to SF.

European Founders in YC 2025

In the Spring batch (X25), 9 of 143 companies were European-headquartered, but 22 companies had at least one European founder. France was the best-represented European country. This is proportionally up from 11 European companies out of 223 in the S24 batch, suggesting YC is slowly opening up to European founders.

The S25 batch was described as one of YC's most international batches to date, with notable presence from India, Europe, Latin America, Africa, and Southeast Asia. But the numbers still heavily favour US-based companies.

Gian's Insider Take

The SF concentration is the real story here. YC tells you that you can stay remote, but the data says otherwise, nearly three-quarters of recent YC companies are in San Francisco. If you're a European or international founder applying to YC, you need to have a credible plan for US presence. That doesn't mean moving your entire company, but it means being willing to spend significant time in SF during and after the programme. The founders I've worked with who got the most out of YC were the ones who went all-in on the SF ecosystem during batch. The ones who stayed remote missed half the value.

The money, funding, valuations & big raises

The YC Investment

Every company accepted into YC receives $500,000 in funding, structured as two SAFEs: $125,000 on a post-money SAFE for 7% equity, plus $375,000 on an uncapped SAFE with a most-favoured-nation provision. This hasn't changed in 2025.

Post-YC Seed Rounds

The median seed round for a YC company in 2025 stabilised at $3.1 million. But there's significant variance by tier:

Healthcare startups consistently raise larger rounds, $4.6M median, reflecting longer development timelines and regulatory requirements.

The Standout Raises from 2025 Batches

YC Alumni: The Bigger Picture

Looking beyond the 2025 batches, YC alumni continued to raise massive rounds:

Gian's Insider Take

The $3.1M median seed number hides a brutal distribution. The top 10% of each batch raises easily and quickly, they're the ones you read about in TechCrunch. The bottom half struggles. YC's brand gives you a warm introduction to every investor, but it doesn't close the round for you. What closes the round is the same thing that closes every round: traction, narrative clarity, and a well-run process. If you're going through YC, your demo day pitch is not the finish line, it's the starting gun. The founders who raise fastest are the ones who've been running their fundraise process in parallel with the programme, not waiting for demo day to begin.

The YC machine, how it actually works now

The standard deal hasn't changed: $500K for ~7% equity. But the programme structure has evolved significantly:

The Acceptance Rate

Approximately 1.5-2% of applicants are accepted into each batch. With four batches per year, more founders can apply, but the rate hasn't materially changed. YC is reviewing more applications, not lowering the bar.

Gian's Insider Take

The real value of YC isn't the $500K or even the demo day. It's being in a room with 160+ other exceptional founders who are going through exactly the same thing you are. People who can become your first customers, your first partners, your first hires. Founders who'll share what's working in their fundraise, introduce you to their investors, and push you because they're building something just as ambitious. That's the compounding asset. That alumni network, 5,600+ companies deep, doesn't expire. And it's why I genuinely believe YC is probably one of the only accelerators worth doing, full stop. Most accelerators take equity and give you a co-working space. YC gives you a network of the most driven founders on the planet and access to every serious investor in the ecosystem. That's always worth the 7%.

What the data tells you about getting in

Based on the 2025 batch data, here's what actually gets founders into YC right now:

Gian's Insider Take

Here's the thing most founders miss about YC, and fundraising in general. They don't have a lot of time with you. A YC partner reads your application in under two minutes. An investor takes a first meeting for 30 minutes. That's it. So the deck is not the thing. How you talk about your business is the thing. The conversations you have. Your positioning needs to be so tight, so granular, that within 60 seconds the person across from you thinks "this is going to be fucking massive." That's what positioning work is. It's not a marketing exercise, it's the difference between getting in and getting passed on. The founders I work with who nail this spend more time on positioning than on their pitch deck, their financials, or their product roadmap combined. Because if you can't make someone feel the inevitability of your business in a single conversation, none of the rest matters.

What this means for your raise

Whether you're applying to YC or raising independently, this data reshapes how you should think about your fundraise:

Gian's Insider Take

Every data point in this report maps back to the same three pillars I've seen across 120+ funded raises: Network (warm intros, YC alumni connections, SF presence), Narrative (AI positioning, founder-market fit, the one-sentence insight), and Process (batched meetings, demo day prep, post-batch fundraise execution). YC doesn't change the fundamentals, it amplifies them. The founders who get the most out of YC are the ones who already understand these principles and use the programme to accelerate them. The founders who struggle are the ones who expect the YC brand to do the work for them. Which one would you be?