Garry Tan says ~80% of YC's W25 batch is AI focused, and the cohort is growing significantly faster than past ones, with actual revenue, thanks to “vibe coding”
Silicon Valley's earliest stage companies are getting a major boost from artificial intelligence.
Context & Ripple Effects
YC's AI concentration has risen through successive cohorts: roughly 35% of a 2023 intake was AI-focused, while investors described the 2024 winter group as notably AI-heavy. By late 2024, YC partners were reporting a surge of applications for vertical AI agents, positioning W25 as an escalation rather than an isolated cohort outcome.
Tan's account adds two economically material claims to that arc: faster growth than prior batches and revenue at the earliest stage. That follows earlier signs of a stronger, consumer-AI-led winter 2024 cohort, but does not establish that the same performance will extend across all AI startups.
First-order effects
- W25 founders focused on AI gain a clearer revenue-and-growth narrative for customer and investor conversations, while YC's current batch becomes still more concentrated around one technology category.
- YC's selection and platform are more directly tied to the outcomes of AI-native companies, since Tan says about 80% of the cohort is AI-focused.
Second-order effects
- Other accelerators and seed investors face pressure to distinguish between AI companies showing revenue and those relying chiefly on technical demos; early traction becomes a more important screening signal.
- The reported speed of cohort growth could intensify competition among AI startups for customers and distribution, particularly in the vertical-agent categories already drawing applications to YC.
Third-order effects
- If revenue-bearing AI cohorts continue to form and grow faster, early-stage startup economics may shift toward smaller teams reaching market tests sooner, raising the bar for traditional SaaS-style build cycles.
- Greater accelerator concentration in AI would make cohort-level results more sensitive to demand for AI products; whether this becomes durable depends on retention and repeatable distribution, not early revenue alone.
The trend: This is part of a broader shift from AI as a startup feature to AI-native companies using faster development cycles to seek revenue earlier.