RESOLVING THE LEDGER
This is one outlet's own report from CNBC — the article as it was filed. Other outlets are covering the same event; open the full story to compare every source side by side.
See the full story · 1 sourcesAirbnb shares surged 15% Friday after the company delivered one of its strongest growth quarters in years and raised its full-year outlook — a turnaround CEO Brian Chesky is directly crediting to artificial intelligence .
In an exclusive interview with CNBC after the earnings report, Chesky said Airbnb will spend "a lot more" on AI tokens this year than it originally forecast because the cost of inference "pales in comparison" to the revenue and productivity the company is getting back.
Airbnb is cutting product-development time by roughly 60%, shipping about 80% more features year over year and keeping headcount roughly flat even as AI spending rises sharply, Chesky said.
"I think now it's safe to say AI is the best thing to have happened to Airbnb," he said. "I think we're becoming an AI-native company, and I think that is probably the number one explanation for our results."
That conviction marks a sharp turn for Chesky. A year ago, Chesky said the question inside the company was essentially, "Is AI good for Airbnb or is AI bad for Airbnb?"
The shift is notable for a CEO who has long approached technology more like a designer than a traditional engineer.
Chesky went to the Rhode Island School of Design and studied industrial design before co-founding Airbnb. Years later, he would become close with former Apple design chief Jony Ive and OpenAI CEO Sam Altman — and eventually introduce the two men, helping spark the relationship that led to their work together on AI hardware.
Now, Chesky is trying to bring that same design-first sensibility to Airbnb's own AI overhaul.
The company hired Ahmad Al-Dahle , Meta 's former head of generative AI and a leader of its Llama work, as chief technology officer in January. Chesky said Airbnb had been "maybe middle of the pack in AI" before Al-Dahle arrived with a mandate to help make the company "AI-native."
There are already measurable signs of that shift. Chesky said AI is helping Airbnb attract more bookings, make it easier for hosts to list and price homes, and lower customer-service costs.
The company is piloting AI-powered search, using AI to generate personalized listing highlights and answers for guests, and helping hosts create and price listings. In customer service, 45% of guests who interact with Airbnb's AI agent never need to speak with a human agent.
"It's really across the board: More demand, more supply, cheaper customer service," Chesky said.
Internally, Airbnb is also closely tracking how employees use AI. Chesky said the company looks at individual token usage as one measure of adoption, but considers that a relatively crude metric and focuses more heavily on the output of teams.
"What we're seeing is that across the board, teams are significantly more productive," Chesky said. He said those gains began with engineers and have since spread to product management, design, marketing, and creative services. "I have so underestimated the impact of AI."
That productivity is showing up in Airbnb's hiring plans as well. Headcount is roughly flat year to date, even as spending on AI rises sharply, and Chesky said investors should expect revenue to grow "a lot faster" than staffing for the foreseeable future.
"Our philosophy has been not necessarily to use AI to have fewer people, but to use AI to get more out of the people," Chesky said. He added that revenue per employee should continue to rise.
The economics are central to Chesky's increasingly bullish view of AI. While many consumer companies are still trying to figure out how to generate enough revenue to justify inference costs, he argues Airbnb has an unusually favorable model.
"One of the problems is a lot of people feel like they don't know how to make money on consumer with AI," Chesky said.
At Airbnb, he said, inference costs pale in comparison with "the amount of money we make on every booking" and the additional revenue generated by faster product development.
"We are going to spend a lot more on AI tokens this year than we forecasted," he said. "But that's great because the ROI is there, and therefore our revenue is much higher."
Airbnb is still trying to be selective about where it spends those tokens and the company is using more than a dozen AI models internally, Chesky said. Those include models like Anthropic 's Claude Code and OpenAI's Codex. But it limits access to some slower and more expensive models when their additional capability is not necessary for the task.
Chesky is particularly bullish on open-source models for consumer-facing products and said frontier models remain important for the hardest problems, but most consumer tasks do not require the most expensive systems.
"Consumers mostly do not need frontier models for most things," Chesky said. "It's a matter of throttling the right job for the right tool."
He declined to name the open-source models Airbnb is deploying in consumer products, saying the makeup of the company's technology stack is becoming competitively sensitive.
The bigger question is whether AI ultimately becomes more than a tool inside Airbnb — and instead changes how travelers reach the platform in the first place. As AI agents become better at assembling itineraries and acting on behalf of users, companies like OpenAI and Alphabet could theoretically become the interface through which travelers discover and book accommodations.
But Chesky is skeptical that chatbots will replace Airbnb as the transaction layer.
Travel is visual, difficult to compare through text alone and often planned collaboratively, he said, all areas where the traditional chatbot interface remains weak.
"I do not believe the chat interface is the right interface for travel," Chesky said.
He expects chatbots to become important for inspiration and itinerary building, but said he does not see them becoming major booking platforms "in the coming future."
For now, Chesky said he is focused on using AI to extend Airbnb's own growth runway. He t...
AIPROPX is an independent multi-source news index — we track, compare, and connect coverage from across the web into one place you won't find anywhere else.