Brian Chesky: Airbnb revenue growth accelerated from 10% to 18%, 60% of code now written by AI

Airbnb: Revenue growth re-accelerates, AI adoption doubles peers

Airbnb's revenue growth has jumped from 10% to 18% year-over-year in the most recent quarter, the first meaningful re-acceleration since the pandemic. For a marketplace doing roughly $100 billion in gross booking value annually, Chesky frames that as genuinely hard to pull off — marketplaces at scale tend to compound downward, not up.

The driver, he argues, wasn't a product overhaul or a marketing push. It was a deliberate effort to rebuild startup intensity inside a large company. Chesky started with a small team he called Project Hawaii, focused narrowly on conversion rate and the guest journey, working at the pace of Airbnb's earliest days rather than as a scaled organization. He reviewed work weekly or daily, removed management layers, and gradually spread that operating model across other teams. The financial results, he notes, lagged the internal change by roughly two years — a reminder that in large companies, the numbers are a trailing signal.

AI adoption

60% of Airbnb's code is now written by AI, which Chesky says is roughly twice the rate of comparable peers. The operational impact is already visible: cost per customer service ticket is down 10%, and 40% of customers who contact Airbnb have their issue resolved entirely by AI without human intervention.

Airbnb's new CTO is Ahmed, who previously led the Llama models at Meta. His observation on joining was that the guest app represents only about 20% of what Airbnb actually is — the bulk of the operation sits in payments, customer service, dispute resolution, a $3 billion host guarantee program, and the host app serving 5.5 million hosts across more than 100 countries.

“We've accelerated growth for the first time since the pandemic. We grew 10% last year in revenue. And this quarter, we announced that revenue was 18%... 60% of our code is now written by AI, which is twice our benchmark of our competitors and peers. 40% of people who contact Airbnb, the AI solves the problem for them. And we brought it through the entire journey, so everything is really accelerating.”

The chatbot interface argument

Chesky's sharpest claim is that the chatbot is the wrong interface for travel and e-commerce. His case is structural: chatbots are text-forward, don't support direct manipulation, make comparison shopping difficult, and are single-player by default. When OpenAI launched third-party apps, Airbnb's stock fell roughly 7% on the news — Chesky thought the idea was sound but argues it needed a richer SDK to work, similar to how Apple's App Store succeeded because it gave developers a genuinely expressive interface rather than routing everything through iMessage.

The interface he expects to win is agentic but visually rich — capable of conversation, comparison, maps, photos, and video. New image and video generation models make that feasible in ways they weren't 18 months ago, and he expects real-time AI-generated interfaces within one to two years.

He's careful to separate that from dismissing AI's current value to Airbnb. Referral traffic from chatbots is already converting at higher rates than Google, so today's models are additive rather than disruptive. Disrupting the travel journey end-to-end would require something much more immersive.

Where AI investment is missing

Almost every AI startup, in Chesky's read, is targeting enterprise or developer tooling. Of a recent YC batch of ~175 companies, only 16 were consumer. He attributes this partly to a mismatch in team composition — most new AI labs are built entirely by researchers, with no product designers or consumer-focused thinkers — and partly to the fact that enterprise is simply easier to monetize right now.

He expects a consumer AI wave within the next two years, accelerated by image and video generation models that finally allow something beyond text-forward interfaces. The reference point he uses is instructive: he thinks this moment for consumer AI resembles the early web design era, when the most talented designers largely avoided the internet and left a gap that product managers filled. Designers who sit out AI, he argues, face the same risk.

Growth vectors beyond core STR

Chesky identifies three adjacent markets he's actively pursuing.

Hotels — roughly eight to nine times the size of Airbnb's current gross bookings — are the largest. About half the world's hotels are independent boutiques rather than chain properties, and Chesky argues they're being forced toward Marriott and Hilton franchising primarily for distribution and loyalty programs. Airbnb sees an opening as an alternative distribution channel on lower commissions.

Services — what he describes as an "Amazon for services" covering roughly 80 verticals — is a second bet. No single vertical may be large, but the aggregate is.

Longer stays (30-plus days) is the third, driven by remote work and nomadic living patterns.

On the core business, he sees a path to $200 billion in gross bookings over time, roughly double current levels, based on the simple math that for every Airbnb guest today, eight or nine people still stay in a hotel — and converting even one more per Airbnb booking would get there.