Airbnb is adding hotels, grocery delivery and airport rides to become 'an Amazon for services' — while nights booked grow 10%

Every figure in this article is drawn from Airbnb's second-quarter 2026 shareholder letter and earnings call, the company's 2026 Summer Release announcements, or from CNBC, Skift and Fox Business as attributed. Sections marked as analysis are identified as such.

Brian Chesky has stopped describing Airbnb as a place to stay. "I imagine one day we'll have dozens, possibly even hundreds of categories, just like Amazon," he said in May, launching a release that added boutique hotels, car rentals, grocery delivery, airport pickups, luggage storage and resort day passes to an app that until recently did one thing. "I think we can build a little bit, like an Amazon for services, at least for traveling and living."

The results published this month explain the ambition rather better than the ambition explains itself.

A record quarter with one slow number in it

Revenue rose 17 per cent to $3.6 billion, a record. Gross booking value rose 16 per cent to $27.2 billion. Net income of $816 million carried a 23 per cent margin against 21 per cent a year earlier, adjusted EBITDA of $1.3 billion a 35 per cent margin, and free cash flow over the trailing twelve months reached $4.8 billion — a 37 per cent margin, which is a software company's number attached to a travel company's business. Airbnb bought back $1.1 billion of stock in the quarter.

Then the number underneath all of it: nights and seats booked grew 10 per cent to 148.3 million. Average daily rate rose 5 per cent to $184. Put those two together and the 16 per cent booking-value growth is roughly two parts volume to one part price.

Ten per cent is not a bad number in absolute terms — it accelerated from the first quarter, first-time bookers rose 11 per cent, the best in four years, and app bookings grew 23 per cent to reach 64 per cent of the total. But it is the growth rate of a mature marketplace, not of the category-defining disruptor Airbnb was a decade ago. Everything Chesky announced in May follows from that fact.

It is worth being precise about what decelerated and what did not. Revenue grew faster than bookings, and bookings faster than nights, which means Airbnb is extracting more per stay at both ends — a slightly higher take rate and a slightly higher nightly price. That is a perfectly respectable way to grow, and it is also the way a marketplace grows when it has stopped adding rooms and travellers as quickly as it used to.

What Airbnb now sells

The catalogue has expanded faster than most travellers have noticed. Grocery delivery through Instacart in more than 25 American cities. Airport rides through Welcome Pickups in over 160 cities. Luggage storage through Bounce at more than 15,000 locations worldwide. Car rentals in the app. More than 3,000 curated Experiences, including tours attached to the Tower of London, Tokyo Skytree and the Taj Mahal, with Experiences supply up nearly 80 per cent year on year in the quarter.

And, most consequentially for the original business, hotels: thousands of boutique and independent properties across more than twenty destinations. Airbnb says hotel nights are growing roughly three times faster than homes, and that about 35 per cent of first-time hotel bookers come back to book a home.

That last statistic is the strategic argument in a single line. Airbnb is not adding hotels because hotels are a better business than homes. It is adding them because a traveller who cannot find the right home now leaves the app, and a third of the ones who book a hotel instead come back for a home later.

The hotels decision is the one that would have been unthinkable at this company five years ago. Airbnb was founded as the alternative to hotels; its entire early marketing was an argument against them. Listing them is a concession that the app's job has changed from replacing an industry to booking a trip — and that the customer, not the category, is the asset worth defending.

Analysis: an aggregator solving an aggregator's problem

There is a well-worn path here and Booking.com walked it first. A marketplace that owns demand eventually discovers that its growth is capped by the supply of one product category, and starts selling adjacent ones to the audience it already has. The economics are attractive because the customer acquisition is already paid for: every incremental category is sold to someone who opened the app anyway.

Airbnb's version has a specific advantage and a specific risk. The advantage is that its brand carries an unusual amount of trust for a two-sided marketplace, and its cash generation — 37 per cent of revenue — means it can fund a dozen experiments without asking anyone for money. The risk is that the further it travels from homes, the more it competes with Booking, Expedia, Viator, GetYourGuide and, eventually, the airlines' own platforms, none of which are asleep, and several of which have run these categories for twenty years.

There is also an execution question that the shareholder letter does not answer. Services and Experiences remain undisclosed as revenue lines; Airbnb reports them in supply growth and anecdote rather than dollars. Until a number appears, an investor cannot distinguish between a second business being built and a feature list being lengthened — and the company has been relaunching Experiences, in one form or another, since 2016.

The counter-evidence in Airbnb's favour is geographic rather than categorical. Brazil's origin nights accelerated past 30 per cent growth, India's first-time bookers more than doubled, and Japan grew origin nights in the high teens. Those are markets where the core product still has years of straightforward expansion left, and where 10 per cent global growth understates what is happening locally.

The AI line that actually shows up in the accounts

Airbnb makes an unusually concrete claim about artificial intelligence, which is rare enough to be worth checking. Its customer-support assistant now resolves nearly 45 per cent of issues that begin with it, in more than fifty languages, and support cost per booking fell 16 per cent year on year. Chesky separately said internal AI tooling had cut time-to-launch by 60 per cent and increased shipped features by 80 per cent.

The first of those is verifiable in the margin line and the second is not, but the direction is consistent: the reason a 148-million-night marketplace can add six product categories in a year without visibly degrading is that the cost of running each additional one has fallen. Chesky's Amazon comparison is really a claim about operating leverage, not about assortment.

Analysis: the guidance is the tell

Airbnb raised full-year revenue growth guidance to at least the mid-teens and set the adjusted EBITDA margin floor at 35.5 per cent, while warning that third-quarter margin will be *slightly down* year on year on the timing of investment. That sentence is the whole strategy compressed: the company is willing to spend against a decelerating core, but only within a margin structure it has no intention of surrendering.

It is a comfortable position and a genuinely uncomfortable question at the same time. Airbnb has $4.8 billion of annual free cash flow, a brand that turned a noun into a verb, and a founder who intends to sell hundreds of categories to the people already holding the app. It also has a core product growing at ten per cent while its average nightly rate does a fifth of the work — and no amount of luggage storage changes what that number means.

It is worth noting how much better this problem is than the one most consumer platforms have. Airbnb is not defending a shrinking core or bleeding users to a rival; it is compounding at ten per cent on 148 million nights with an industry-leading margin and a buyback running at over a billion dollars a quarter. The question is not survival. It is whether the second act is large enough to matter.

The test will arrive in the ordinary way. If nights growth stays at ten per cent and Services and Experiences become a visible line in the accounts, Chesky's expansion will read as foresight. If the new categories stay too small to disclose separately, the 2026 Summer Release will read as what companies do when the original idea stops compounding.