Airbnb Says Nearly 60% of Its Engineer-Authored Code Is Now AI-Written. F&B Should Care

Software engineer at a desk with multiple monitors showing code editors and a Slack window open in a corner.

On the Q1 call Brian Chesky disclosed that nearly 60% of code Airbnb engineers produce is AI-written — twice the industry average — and 40%+ of guest issues now resolve without a human. Hospitality-tech founders should read this as the cost-curve benchmark.

I was on the Airbnb Q1 call this morning with a cold flat white and a fresh notebook, ready to mark up the usual nights-and-seats line. Then Brian Chesky dropped a number I had to write down twice. “Nearly 60% of the code our engineers produce is now written by AI,” he said, “which we estimate is about twice the industry average.” I underlined it. Then I circled the second number that should be making every hospitality-tech founder reading this very, very awake: more than 40% of guest support issues are now being resolved by Airbnb’s AI Assistant without a human agent in the loop. That figure was around 33% in Q4 2025. One quarter. Seven points.

Here is my contrarian take, up top, before the bullets start: this is not an AI marketing line. This is Airbnb publishing a cost curve. It is the new benchmark a hospitality-tech founder should beat — or, if you cannot beat it, at least be honest with your board about how far away you are. Cost-per-booking is down roughly 10% year-over-year. Adjusted EBITDA was $519M on $2.68B revenue, a 24% EBITDA jump on 18% revenue growth. That is operating leverage that increasingly looks model-led, not headcount-led.

The 60% number, decoded

Let’s be careful with the Chesky stat, because the temptation is to read “60% AI-written” as “60% of engineering work is automated.” It is not. It is the share of produced code that flows through an AI tool — Copilot-style completions, agentic refactors, generated tests — before a human ships it. The honest read is that AI is now load-bearing across the inner loop, not a side experiment. Twice the industry average, per Chesky, is the part founders should chew on. If Airbnb’s internal benchmark is 60% and the mean is roughly 30%, the platforms you are competing with for engineering productivity are pulling away.

My read: the relevant follow-up question is not “are you using Copilot?” It is “what percent of your shipped code is AI-touched, and what is your trailing four-quarter trend?” If you cannot answer that on your next board call, you are behind Airbnb. Boards will start asking, because Airbnb just made the question legible.

The financial picture underneath the engineering line: revenue $2.68B (+18%), Adj EBITDA $519M (+24%), net income $160M, GBV $29.2B (+19%), nights and seats 156.2M (+9%). CFO Ellie Mertz raised the FY26 outlook to low-to-mid teens revenue growth with an Adjusted EBITDA margin of at least 35%. That is a guide-up and a margin floor in the same breath.

Why the AI Assistant cost-savings line matters more to F&B than the engineering one

Engineers love the 60% number. F&B founders should care more about the 40%+ guest-resolution number. Here is why: deflection on guest contacts is the closest analog you have to host-side and operator-side support — the late dietary request, the “where’s my reservation,” the after-9pm refund question. Airbnb’s contact center used to be a famously expensive operation, and one full quarter of additional AI deflection (from ~33% to 40%+) is what is driving the cost-per-booking line down 10% Y/Y. That is not a moonshot lab project. That is a quarter of work.

If you are building voice or chat in restaurant tech, the cost-curve benchmark is now publicly disclosed: get to 40% containment, on a contact mix that includes refunds and emotional escalations, in twelve months. We dug into where this curve bends for restaurants in the voice agent maturity piece, and I’d argue Airbnb’s numbers validate the upper end of that ramp. OpenTable’s AI moves we covered last month read very differently against this backdrop — the gap between “AI features shipped” and “AI deflection measured” is the gap that matters now.

My read: the operators I trust will not be asked “do you have AI?” in their next vendor RFP. They will be asked “what’s your deflection rate, and how do you measure it?” Have an answer.

What’s coming on May 20

Chesky teased — and the Q1 Shareholder Letter confirms — that the company’s Summer Release is scheduled for May 20. Anticipated: a deeper hotels push and another turn of the Experiences flywheel, which is the most underrated piece of the Airbnb story for anyone selling into hospitality. Per the letter, verbatim: “nearly a quarter of guests who are new to Airbnb and book an experience go on to book a stay or a service” and “roughly one in three experience bookers book a stay within 90 days.”

That is a cross-sell rate that any restaurant group running a tasting menu, a chef’s table, or a ticketed event series should be staring at. Experiences are the wedge. Stays are the LTV. If a third of experience-only bookers convert to a stay inside 90 days, the F&B equivalent — a one-off ticketed dinner converting to a recurring chef’s table or a multi-night package — is sitting on the table, untouched, at most hospitality groups I talk to.

I will be on the Summer Release stream May 20. Bring me your hot takes; I will run the best ones.

— Hana edits The Pass. Tips: tips@tabletransfers.com.

Featured More

The Voice Agent Maturity Curve

mise

·

12 min read

The Four Margins of a Restaurant

mise

·

14 min read

The AI Premium in Hospitality M&A: Broker Story or Real Number?

the bottom line

·

9 min read

What the DoorDash/SevenRooms Deal Actually Buys

the bottom line

·

11 min read

Browse all 494 posts

Related posts

Toast Quietly Renamed Sous Chef. The Pilot Was the Point.

the pass

·

6 min read

Toast Quietly Renamed Sous Chef. The Pilot Was the Point.

Darden +4.2% comps and the boring Bahama Breeze ending

the pass

·

5 min read

Darden +4.2% comps and the boring Bahama Breeze ending

Bahama Breeze is closing. Darden's portfolio thesis just got tighter.

the pass

·

5 min read

Bahama Breeze is closing. Darden's portfolio thesis just got tighter.