Airbnb Raises Forecast as AI Spending Expands
Airbnb raised its full-year revenue-growth outlook after reporting $3.61 billion in second-quarter revenue on August 6. The company said it shipped nearly 80% more features in the first half of 2026 and cut development time by as much as 60%; CEO Brian Chesky told CNBC that Airbnb would spend more on AI inference because management sees larger gains in bookings, product delivery, and support costs.
Airbnb raised its full-year revenue-growth outlook after reporting second-quarter revenue of $3.61 billion, up 17% year over year. In its August 6 shareholder letter filed with the U.S. Securities and Exchange Commission, the company said gross booking value reached $27.2 billion, nights and seats booked increased 10% to 148.3 million, and adjusted EBITDA rose 21% to $1.3 billion.
Airbnb now expects full-year revenue growth of at least the mid-teens, up from its earlier low- to mid-teens outlook. It also raised its full-year adjusted EBITDA margin forecast to at least 35.5%. For the third quarter, the company projected revenue of $4.69 billion to $4.77 billion, representing 15% to 17% year-over-year growth.
Management links AI spending to operating measures
The shareholder letter says Airbnb has rebuilt its product organization around an AI-native approach. Management reported that time from concept to delivery fell by as much as 60% on key initiatives and that the company shipped nearly 80% more features and improvements in the first half of 2026 than in the same period a year earlier. These are company-reported operating measures rather than an independent evaluation of AI productivity.
In an August 7 CNBC interview, CEO Brian Chesky said Airbnb would spend more on AI tokens than originally forecast because inference costs were small compared with the revenue and productivity benefits management attributed to the technology. CNBC reported that Airbnb kept headcount roughly flat while increasing AI spending. Quartz separately reported that customer-support cost per booking declined about 16% year over year and that Airbnb's assistant resolved nearly 45% of support issues without a human agent.
The evidence supports a narrower conclusion than the headline share-price reaction might suggest: Airbnb is pairing higher inference spending with workflow measures it can track, including delivery time, feature throughput, support cost, and automated resolution. It does not isolate how much of the quarter's booking or revenue growth came from AI rather than travel demand, product changes, pricing, or other operating factors.
What practitioners can measure
For product and ML teams, the useful part of Airbnb's disclosure is the choice of unit-level and workflow-level metrics. Token spending alone does not show whether an AI deployment is valuable. Release cycle time, support cost per booking, resolution rate, reliability, and margin contribution provide a more concrete basis for judging whether a system is improving the business.
Future quarters will show whether those improvements persist as Airbnb expands AI use and associated inference spending. The company has not disclosed the absolute amount of its AI-token budget, the model providers involved, or a controlled attribution method for separating AI effects from its broader product and operations program.
Key Points
- 1Airbnb reported $3.61 billion in second-quarter revenue and raised its full-year growth and adjusted EBITDA margin outlooks.
- 2Management reported up to 60% shorter development cycles and nearly 80% more shipped features, while CNBC said Airbnb plans to spend more on AI inference.
- 3The disclosures tie AI to measurable workflows, but they do not isolate AI's contribution from travel demand and broader product changes.
Scoring Rationale
Airbnb's earnings disclosure provides concrete, company-reported measures connecting AI use with development speed and support economics. The figures are useful for practitioners evaluating inference spending against workflow outcomes, but the company does not isolate AI's causal contribution to revenue or bookings.
Sources
Primary source and supporting public references used for this report.
Practice with real Hotels & Lodging data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Hotels & Lodging problems
