The Airbnb AI spending surge follows a 15% stock rally after CEO Brian Chesky outlined plans to build an AI-native travel platform.
Airbnb shares jumped 15% on Friday after the company reported one of its strongest growth quarters in years and raised its full-year outlook a turnaround CEO Brian Chesky credits directly to artificial intelligence.
Speaking to CNBC after the earnings release, Chesky revealed that Airbnb plans to spend “a lot more” on AI inference tokens than initially budgeted. He noted that the compute cost “pales in comparison” to the massive gains in revenue and employee productivity the company is seeing in return.
Airbnb has slashed product-development cycles by roughly 60% and is shipping 80% more features year-over-year all while maintaining flat headcount despite increased AI expenditure, according to Chesky.
“AI is easily the best thing to happen to Airbnb,” Chesky stated. “We are transitioning into an AI-native company, and that shift is the primary driver behind our recent performance.”
That conviction marks a sharp shift for Chesky. Just a year ago, he noted that the central debate within the company was whether AI posed an existential threat or a major opportunity for Airbnb.
The pivot is especially striking coming from a CEO who has historically approached product development through the lens of design rather than traditional software engineering.
Chesky studied industrial design at the Rhode Island School of Design before co-founding Airbnb. Years later, he formed close ties with former Apple design chief Jony Ive and OpenAI CEO Sam Altman eventually introducing the two men and sparking the partnership behind their collaborative AI hardware project.
Now, Chesky is applying that same design-first philosophy to drive Airbnb’s internal AI transformation.
Airbnb accelerated its transformation in January by hiring Ahmad Al-Dahle Meta’s former head of generative AI and a key architect of the Llama model family as Chief Technology Officer. Chesky noted that Airbnb had been “maybe middle of the pack in AI” before bringing Al-Dahle on board with an explicit mandate to build an “AI-native” platform.
The strategic pivot is already yielding concrete results: AI tools are driving higher booking conversion, streamlining listing and pricing workflows for hosts, and lowering customer support expenses.
Airbnb is currently piloting AI-powered search using artificial intelligence to generate personalized listing highlights and tailored answers for guests, while helping hosts create and accurately price their properties. The impact is especially stark in customer support, where an AI assistant now handles nearly 45% of guest inquiries entirely without human intervention.
“It’s really across the board: More demand, more supply, cheaper customer service,” Chesky noted.
Internally, Airbnb tracks employee AI adoption by measuring token usage, though Chesky views that as a crude metric, favoring team output instead.
“Across the board, teams are significantly more productive,” Chesky said, noting that efficiency gains originated in engineering and have since expanded to product management, design, marketing, and creative services. “I have so underestimated the impact of AI.”
That surge in productivity is also shaping Airbnb’s hiring strategy. Headcount has remained largely flat year-to-date despite a steep rise in AI investments, and Chesky told investors to expect revenue growth to outpace staffing additions for the foreseeable future.
“Our philosophy isn’t to use AI to replace people, but to amplify what our people can do,” Chesky explained, projecting that revenue per employee will continue its upward trend.
The underlying economics are central to Chesky’s growing optimism around AI. While many consumer tech companies struggle to monetize AI features enough to justify high inference costs, Chesky contends that Airbnb possesses a uniquely advantageous business model.
“The main challenge right now is that many companies feel uncertain about how to actually monetize consumer AI,” Chesky noted.
At Airbnb, inference costs “pale in comparison” to the substantial margins generated on every booking and the additional revenue driven by accelerated product cycles, according to Chesky.
“We are going to spend a lot more on AI tokens this year than originally forecasted,” he acknowledged. “But that’s a net positive—the ROI is clearly proven, and as a result, our top-line revenue is significantly higher.”
Airbnb remains selective about token expenditure, deploying a suite of over a dozen specialized AI models across its organization. While development teams leverage powerful tools like Anthropic’s Claude Code and OpenAI’s Codex, Airbnb restricts access to slower or more expensive frontier models when routine tasks don’t require high-tier capabilities.
Chesky is particularly bullish on open-source models for consumer applications, arguing that while frontier models remain essential for complex reasoning, most everyday consumer tasks simply don’t require high-cost compute.
“Consumers mostly do not need frontier models for most things,” Chesky explained. “It’s a matter of throttling the right job for the right tool.”
While he declined to specify which open-source models Airbnb is currently running in its consumer products, Chesky noted that the precise composition of the company’s AI stack has become a key competitive advantage.
See also: Google Wallet Expands Parental Controls with Direct Digital Allowance Balances for Kids
The pivotal question is whether AI remains an internal utility for Airbnb or fundamentally alters how travelers discover accommodation. As AI agents mature in crafting itineraries and executing complex tasks, ecosystems like OpenAI and Alphabet could theoretically become the primary discovery and booking layer.
However, Chesky remains skeptical that conversational chatbots will disintermediate Airbnb. He notes that travel planning is inherently visual, highly comparative, and collaborative three fundamental areas where text-centric chatbot interfaces fail to deliver a compelling user experience.
“I do not believe the chat interface is the right interface for travel,” Chesky asserted.
