The CEO whose company now ships code written mostly by machines thinks the problem with AI isn't the tech—it's that nobody's building anything grandma would actually use.
The Summary
- AI now writes 60% of Airbnb's code, yet CEO Brian Chesky says the real AI crisis is a product problem, not a perception problem
- Silicon Valley's startup class has gone enterprise-heavy: 159 of 175 companies in Y Combinator's latest batch are B2B, not consumer-facing
- Chesky's fix for AI backlash: build things regular people can actually use, like affordable on-demand healthcare, instead of more SaaS tools for corporations
- He credits "founder mode"—running a $110 billion company like a small startup—as the key to winning when agents start shipping most of your code
The Signal
Airbnb is already living in the future most companies are still planning for. Sixty percent of its code comes from AI. Not augmented by AI. Not "AI-assisted." Written by AI. That's not a pilot program or a productivity experiment. That's the new normal at a company with 150 million users and $110 billion in market cap.
But Chesky isn't here to talk about his own wins. He's pointing at everyone else's failure. The AI backlash among Americans isn't about fear of the unknown, he argues. It's about the complete absence of products that matter to actual humans. "I think part of it's a narrative issue that we're not talking about AI correctly," Chesky said on Yahoo Finance's "Power Players" podcast. "But part of it is we need to actually be developing more products that just regular people can use."
"I love AI because AI allows me to have a doctor on demand and I can't have that. I can't afford that."
The numbers back him up. In Y Combinator's most recent batch—where Chesky sits on the board—159 out of 175 startups were enterprise, not consumer. That's 90%. Eighteen years ago, when Chesky started Airbnb, founders wanted to build things people touched every day. Now they're "afraid" to ship consumer products. The money, the talent, the ambition—it's all going into tools for other companies, not people.
Key shifts since Airbnb's founding:
- 2008: Consumer products were the dream. Make something people love.
- 2026: 90% of top accelerator startups are B2B. Enterprise is safe. Consumer is risk.
- Result: AI gets rolled out to optimize corporate workflows while regular people see nothing that improves their lives.
This isn't just a missed opportunity. It's a structural problem. When the most sophisticated AI tools only show up inside Salesforce or Slack or some HR platform, people don't get to feel the upside. They only hear the fear: AI will take your job. AI will replace you. Meanwhile, the people building the AI are optimizing sales pipelines and writing Jira tickets faster. No wonder public sentiment is souring.
Chesky's example—on-demand doctors for people who can't afford traditional care—isn't hypothetical. It's the kind of thing AI could do tomorrow if anyone bothered to build it for humans instead of just for enterprises. Healthcare access. Personalized education. Legal help that doesn't cost $400 an hour. These are consumer problems AI could solve, but Silicon Valley would rather sell another seat license.
The Implication
If Chesky is right, the agent economy has a branding problem that only product can fix. You can't market your way out of "AI is going to take my job" when the only AI products people encounter are corporate efficiency tools. The antidote is consumer value so obvious it changes the conversation.
For founders: the enterprise path is crowded and the consumer path is wide open. If 90% of your batch is going B2B, that's not validation. That's everyone chasing the same safe money while ignoring the market that could actually move public opinion. The companies that win Web4 won't just automate work. They'll deliver things people couldn't get before, at prices that actually matter.