Airbnb CEO Brian Chesky to Launch a New AI Lab Airbnb CEO Brian Chesky is moving from behind-the-scenes AI power broker to direct competitor, backing a new artificial intelligence lab while continuing to run the home-sharing giant. The initiative underscores a growing rift between what big-tech “frontier labs” are building and what some industry leaders say they actually need.

Early relationships and rising frustration

Chesky’s move comes after nearly two decades in the AI orbit. He first met OpenAI CEO Sam Altman through Y Combinator in 2006 and later advised him on managing OpenAI’s hypergrowth, even helping broker Altman’s return after his brief ouster in 2023. But despite that closeness, Chesky has grown dissatisfied with existing large language model offerings.

By 2023–2024, Chesky was already publicly saying Airbnb hadn’t signed an LLM partnership because “existing products weren’t quite ready” for what he wanted to build. His core argument: travel and commerce require rich, visual and interactive experiences, not just text-based chatbots.

Airbnb’s current AI play

Airbnb has not stood still on AI. The company hired former Meta generative AI leader Ahmad Al-Dahle as CTO in January 2026, rebuilt its app around an LLM for conversational search, automated 40% of customer support via bots, and added AI-generated listing descriptions and review summaries, with a voice assistant planned later in 2026.

Still, Chesky has concluded that buying from frontier labs is “not enough” and that Airbnb must help build at the “model layer, not just the application layer.”

The new lab and the competitive landscape

Bloomberg first reported that Chesky plans to back a new AI lab focused on user interaction and design, a move that “marks Chesky as one of many Silicon Valley machers who are unsatisfied with the models coming out of the frontier labs.” He will remain Airbnb CEO and will not lead the lab directly.

This places Chesky in implicit competition with his former mentee’s OpenAI, as well as newer efforts like Brett Adcock’s Hark and Mira Murati’s Thinking Machines Lab, which similarly argue that the next defensible edge lies in how humans interact with AI, not just how smart the models are.

Continue reading https://foxvector.com/stories/019ea83b-13d9-123b-71a2-21bbd244eccd