"YC startups are now expected to grow 10–20% every single week."

On October 20, 2025, Garry Tan joined Andrew Warner on the Mixergy podcast for a 50-minute conversation about how AI is reshaping the Y Combinator playbook, the remarkable revenue velocity of recent batches, and the unconventional tactics today’s founders are using to win markets.

AI is rewiring the YC playbook

Tan opened by describing how AI has shifted from a niche feature to the organizing principle for nearly every company in the YC pipeline. “If you arrive at YC without an agent or workflow powered by AI, you’re already behind,” he said. The conversation touched on founders who now spin up custom copilots for onboarding, customer support, and even engineering QA before the batch begins. Tan emphasized that the winners leverage AI not as a press-release buzzword but as a way to collapse time-to-product: “Founders are launching in weeks what used to take teams months.”

10–20% weekly revenue growth is the new floor

When Warner asked about metrics, Tan didn’t hesitate: “Our best teams are putting up 10 to 20 percent revenue growth every week. Two years ago we were thrilled with two to four percent.” He traced the shift to a tighter feedback loop made possible by AI-first tooling—automated sales playbooks, self-learning onboarding flows, and product analytics that tell founders what to ship next. According to Tan, those who cannot stay on that trajectory quickly self-select out of the batch: “The pace is brutal, but it reflects what the market rewards right now.”

The founder playbook for 2025

Tan outlined a handful of tactics he now recommends to every incoming founder:

• Start with distribution. “The best teams show up having already built a waitlist or community,” he noted. By the time Demo Day approaches, they are refining pricing, not searching for customers.
• Automate the boring parts. Tan highlighted companies that built headless AI assistants to handle billing issues, compliance questionnaires, and procurement redlines. “That frees the human team to focus on learning from customers instead of chasing paperwork.”
• Be ruthless with experiments. He pointed to founders running daily growth retros and shipping at least one measurable improvement every 24 hours. “It’s not about velocity for its own sake,” Tan said. “It’s about compounding learning faster than your competitors.”

Advice for the next generation of YC founders

Reflecting on his own journey from engineer to investor to YC’s leader, Tan urged aspiring founders to combine optimism with pragmatism. “AI gives you leverage we’ve never had before, but it doesn’t remove the obligation to understand your customer better than anyone else,” he said. His parting advice: “Find the problem that keeps you up at night, use AI to remove the friction around solving it, and hold yourself to the growth curve that today’s YC companies are proving is possible.”