The Blubber

Silicon Valley’s youngest startups are seeing a significant lift thanks to advances in artificial intelligence.

At this week’s annual demo day in San Francisco, Y Combinator — the startup accelerator behind companies like Airbnb, Dropbox, and Stripe — showcased its latest batch of early-stage companies to a room full of potential investors.

Y Combinator CEO Garry Tan told CNBC that this cohort is expanding faster than any in the accelerator’s history, with real revenue to show for it. According to Tan, the winter 2025 batch is growing at a pace of 10% per week across the board.

“It’s not just the top one or two startups — the entire group is growing 10% week over week,” said Tan, himself a Y Combinator alum. “That’s never happened before at this stage.”

Tan credits the surge to breakthroughs in AI. Developers are now able to automate repetitive work and generate new code using large language models. He referred to this trend as “vibe coding,” where AI models take the lead in writing software — in some cases, even building entire applications.

As AI takes on more of the development burden, many startups have been able to operate with smaller teams. Tan said that roughly a quarter of companies in the current YC batch wrote 95% of their code using AI.

“That might sound intimidating, but it means founders don’t need 50 or 100 engineers,” Tan said. “Startups can reach $10 million in revenue with fewer than 10 people. That reduces the need to raise large amounts of capital, and funds stretch much further.”

Tan noted that the previous “growth-at-all-costs” mentality common during the low-interest-rate era has given way to a renewed focus on profitability — a shift that’s also been seen at major tech firms like Google, Meta, and Amazon, which have gone through waves of layoffs and slowed hiring.

While this has created uncertainty for engineers, Tan sees opportunity.

“It’s easier than ever to build a startup,” he said, pointing out that today’s top engineers don’t have to prove themselves by joining a big tech company. “There’s a lot of anxiety among young software engineers, but maybe the one who couldn’t get hired at Meta or Google ends up building a business making $10 million or even $100 million a year with a small team — that’s a powerful shift.”

Around 80% of the companies that presented at demo day were focused on AI, with some in robotics and semiconductors. Tan said these startups are reaching commercial validation faster than previous generations.