Airbnb Reports AI Writes 60% of Its Code

Airbnb CEO Brian Chesky told investors on the company's Q1 2026 earnings call, as reported by Business Insider and TechCrunch, that AI generates nearly 60% of the code Airbnb's engineers produce, and that many design and engineering managers are "going back to coding or using Claude Code." Chesky told investors, per TechCrunch, that AI is especially useful for building tools for Airbnb's API partners: "AI gives huge leverage, where you might have needed a team of 20 engineers before, an engineer can now spin up agents to do a lot of work under supervision." Airbnb reported Q1 revenue up 18% year-over-year to $2.7 billion, net income up 3.9% to $160 million, gross booking value of $29 billion, and nights booked up 9% to 156.2 million, per TechCrunch. The company said its customer-support AI bot now resolves more than 40% of issues without escalating to a human agent, up from about 33% earlier in the year, and Yahoo Finance's earnings recap links these efficiencies to an approximately 10% year-over-year reduction in cost per booking.

TechCrunch notes Airbnb's 60% figure echoes comparable disclosures from Google (Sundar Pichai has said more than 30% of the company's new code is AI-generated), Microsoft (Satya Nadella has said up to 30% of the company's code was AI-written), and Spotify (which has said its best developers have not written a line of code by hand since December). None of these companies have published a shared definition of "AI-written" versus "AI-assisted" code or a reproducible measurement methodology, so cross-company comparisons should be treated as directional rather than precise.

Chesky was notably candid about where AI still falls short: he said, per TechCrunch, that "no one has figured out AI for travel or e-commerce yet," citing chatbot interfaces that rely too heavily on text, lack direct manipulation controls, handle comparison across many options poorly, and are not built for the multiplayer, map-native nature of most bookings. That caveat is a useful counterweight for practitioners evaluating AI-agent interfaces in similarly visual, comparison-heavy domains, and a reminder that a high AI-code-generation share does not automatically translate into solved AI-product experiences.

Watch for how Airbnb and peer companies define and measure "AI-written" code in future disclosures, whether the claimed cost-per-booking and support-automation gains are corroborated in the 10-Q or subsequent investor materials, and whether other large consumer-tech firms adopt similar earnings-call disclosures as a competitive signaling norm.