ABNB (Trans): FY guide raised on both top- and bottom-line; Q3 margin slightly down

Dolphin Research Trans: ABNB FY26 Q2 earnings call notes

I. Key takeaways

1. Shareholder returns: Repurchased $1.1bn of common stock in Q2, underpinned by a strong balance sheet and solid cash generation. Returning capital remains a core pillar of capital allocation.

2. Q3 guidance

Revenue of $4.69bn–$4.77bn (+15%–17% YoY), including an approx. 300bps FX tailwind after hedges.GBV to grow mid-teens YoY, driven by low double-digit growth in nights and seats booked and a modest ADR uplift from mix and pricing.Adj. EBITDA to grow YoY, but OPM to be slightly below Q3 2025 due to timing of investments this year.

No material impact assumed from the Middle East conflict.

3. FY26 guidance raised

Revenue growth raised from low-to-mid-teens to at least mid-teens, reflecting accelerating nights and seats booked.Adj. EBITDA margin raised from 35% to at least 35.5%.

Full-year implied take rate expected roughly flat vs. 2025, due to RNPL timing differences between booking and stay, and higher customer incentives for new initiatives in 2026.Ex-incentives, take rate would have modestly trended up, driven by monetization initiatives and product roadmap execution.

4. Q2 headline metrics

Top line: revenue $3.6bn (+17% YoY), above the high end of guidance; GBV $27.2bn (+16% YoY).Volume/price: nights and seats booked +10% YoY, accelerating vs. Q1; ADR +5% YoY, +4% ex-FX, with notable strength in N. America and Europe.

Regional nights and seats: high-single-digit growth in N. America and Europe (Europe steadily recovered from Q1 Middle East-related drag); ~+20% in LatAm; high-teens in APAC.

Profit: net income $816mn; Adj. EBITDA $1.3bn with margin of 35%, up >100bps YoY, driven by strong revenue growth and efficiency in ops support and product development, partially offset by sustained S&M investment.

Net income YoY growth also reflected a $77mn tax benefit recognized this quarter, tied to recently issued tax guidance impacting prior periods.Cash flow: Q2 FCF $1.3bn; LTM FCF $4.8bn with a 37% FCF margin. Ex-RNPL (which shifts guest payments from booking to near check-in), unearned fees would have grown YoY in Q2.

5. Monetization policy updates

RNPL (Reserve Now, Pay Later): over 20% of Q2 GBV booked via this flexible payment option; expanded to more booking types in Jul given strong results. Management believes it drives more bookings, longer booking windows, and contributes to higher ADR.

RNPL long-term value: beyond the immediate lift in nights, it helps hosts lock calendars earlier and aligns payment preferences with guests, building a longer-term competitive edge.

Single service fee: migration of API-connected hosts (mainly pro property managers) is progressing; now rolling out to most remaining hosts and targeting completion by year-end. Roughly half of active live listings now on the single-fee model.

II. Call details

2.1 Management remarks

1. Growth quality and user mix

App-booked nights rose 23% YoY, accounting for 64% of total nights vs. 59% a year ago.First-time bookers accelerated to +11%, a four-year high, led by Gen Z.

Net nights in expansion markets are growing at roughly 2x core markets; importantly, core markets are also re-accelerating, with the U.S., France, the U.K., and Australia all faster in Q2.While Middle East conflict continues, underlying global demand remains strong, and the impact has been smaller than expected.

2. AI-native rebuild, R&D efficiency, and this quarter's drivers

The company rebuilt itself as AI-native; for key projects, concept-to-launch timelines have shortened by up to 60%.Feature and improvement releases are up nearly 80% vs. the same six-month period last year, with faster build, test, and iteration cycles.

Management stressed there is no single product, partnership, or lever explaining the quarter. Strong execution, a world-class team, and AI-accelerated innovation together drove outperformance — 'no silver bullet'.These changes are embedded in how the biz. is built and run, supporting confidence that momentum will carry into coming quarters and underpinning the full-year raise.

3. Guest-side product (Project Y)

Project Y is the innovation blueprint, compounding hundreds of incremental improvements over time. AI is accelerating search, sign-up, checkout, and payments, reducing friction across the funnel to convert more traffic into bookings and becoming a major growth driver.

Improvements in search and discovery help guests more easily find and book the right homes, hotels, services, or experiences, materially lifting conversion.AI-generated listing highlights and AI-summarized review highlights are live; AI Home Comparison will launch later this year.

Simplified sign-up/login; clearer pricing and cancellation policies at checkout.RNPL continues to expand to more listings and is more prominent in the booking flow.

4. Host-side tools

With the right pricing, insights, and tools, hosts are more likely to succeed, and AI is improving all three.It helps hosts set more competitive prices to win bookings and provides more actionable insights to improve listings and income potential.

New AI tools for new hosts are rolling out to speed ramp-up and clarify pricing and revenue opportunities.

5. AI support and cost efficiency

The AI assistant supports 50+ languages. Roughly 45% of AI-initiated inquiries are resolved without human agents, with markedly shorter resolution times.

AI voice support will begin rolling out later this year to extend the experience to phone.Customer support cost per booking fell ~16% YoY in Q2, partly due to the AI assistant; costs should keep declining as resolution rates rise and voice launches.

6. Services / Experiences

In May, services expanded to groceries, car rentals, airport transfers, and luggage storage; resort passes were introduced to access select top hotel amenities by the day.

Early days, but booking momentum is strong. Each new service is launching faster than the last — groceries took 8–9 months, airport transfers ~6 weeks.

Experiences: added thousands of supply in peak-demand categories in Q2, with supply up nearly 80% YoY. Bookings are accelerating both YoY and QoQ, though the overall base remains small.

7. Hotels

Added thousands of boutique and independent hotels across top global destinations.Featured hotels offer price-match guarantees and up to 50% credits redeemable on future bookings.

About 35% of guests who first book a hotel on the platform return to book homes, indicating hotels bring in new users who also adopt homes, strengthening the core home biz.Hotels still account for only a single-digit percentage of nights, but hotel night growth is ~3x that of homes, implying significant runway.

8. Major events strategy

Born from event-driven stays, the company believes major events simultaneously stimulate supply and demand and strengthen the brand, enabling cities to host more visitors without building new hotels.

As an official World Cup partner, it hosted millions of stays during the tournament, many first-timers. Over 150k listings went live in host cities, creating new local economic opportunities.

This playbook is extending to the Olympics, Tour de France, Art Basel, Lollapalooza, Spain's La Liga, and the recent NASCAR.While event-driven bookings can be temporary, the brand awareness, trust, and new hosts acquired can deliver long-term benefits post-event.

2.2 Q&A

Q: How is the hotel push progressing? In cities with ample home supply (not just regulated markets), is conversion improving?

A: Hotels are tracking well ahead of already high expectations. The strategy is simple: some trips fit homes better, but many suit hotels — last-minute, short stays, business travel, and places where home supply is constrained.The goal is to build the best hotel-booking product on the internet, and we believe we have achieved that.

The second question is supply. Initially we thought convincing hotels to list on a home-first platform would be hard, but outreach flipped the script — many hotels now proactively want to list on Airbnb.Beyond a strong product, hotels value our massive traffic (one of the most visited travel sites globally), younger user base, higher U.S. mix vs. some peers, and very favorable commission rates.

We are seeing strength in both constrained and unconstrained markets. Not everyone arrives intent on booking a home — some prefer only homes, some only hotels, most are open to both.Adding either home or hotel inventory increases conversion and bookings. And ~35% of first-time hotel bookers return to book homes, meaning hotels add new users who often then choose homes, reinforcing both sides. We will step on the gas across all markets, not just constrained ones.

Q: For adjacent services, is the endgame attach to home bookings, or a full-service OTA that can plan a two-week Scotland trip?

A: Think of three phases. Initially we were synonymous with home stays, like Kleenex/Xerox becoming a verb.Phase one (just passed) is largely about attaching to users who still come with home intent — add a service or an experience, or convert them to a hotel when a suitable home isn’t available.

We are now entering the next phase: becoming a one-stop entry point for travel, where users can fulfill all travel needs end-to-end on Airbnb. The subsequent phases extend from travel to living, and then to other ways people connect on the platform.Over the next year we will focus on the one-stop travel phase, but our ambition goes beyond travel.

The singular reason is AI is an existential risk for everyone. We brought in CTO Ahmad Al-Dahle (formerly led Meta’s Llama), moving from mid-pack to a leader in AI among non-frontier labs and non-hyperscalers.We consider ourselves one of the most AI-native companies in Silicon Valley, opening doors to new travel adjacencies and, over time, opportunities beyond travel.

Timing: within the next 12–18 months, users will be able to start with a broad travel intent — 'I want to take my family somewhere in Europe; can you recommend and plan the entire trip?' — and Airbnb will deliver. That is the next stage, not the endpoint.

Q: With a car-rental partner acquired by a competitor, does M&A become a necessity?

A: You are referring to CarTrawler’s acquisition by Expedia. We still view them as a strong partner for car rentals and are confident in the relationship, expecting mutual value and continued collaboration.This does not change our M&A posture, which remains opportunistic.

We are entering many new verticals with ample M&A opportunities, supported by substantial cash on hand and ongoing cash generation, and entrepreneurs willing to join and take stock.But we prefer organic first, with M&A second.

Q: Where are hotels driving clear conversion gains or share, and any timing/scale color?

A: We started in regulated markets where unmet demand was most obvious and have expanded supply acquisition to the top 20 cities. We are highly selective, adding inventory that is incremental rather than directly competing with homes, ensuring quality that consumers will love.

Scale-wise, hotels are still a single-digit percentage of nights, a relatively small segment. But as disclosed, growth is ~3x homes after recent efforts.We have big ambitions to scale over the next few years, in both growth and contribution to the overall biz.

(CEO) We first tackled the truly hard business — global homes — which required working with governments worldwide and enabling a non-habitual behavior.Hotels are not easy, but they make our business easier: hotel partners are eager, and much of our tech and tooling for homes applies to hotels. It is a natural extension, and user experience handles side-by-side curation well via strong, AI-driven personalization.

Q: How is AI search testing going, and when will it become core?

A: Testing began this month with a very small traffic slice and will scale through the year based on results. Early signals are excellent, but users come only a few times a year and expect a location-input search box, so habit-shifting will take months.

Initial launch keeps the current core search as default with a toggle on top to try AI search. We expect strong conversion among those who opt in, but won’t force it.This mirrors the full-price display rollout — start as a toggle, collect positive feedback, then educate more users through this year and into next.

AI search spans 3–4 core capabilities: natural-language input; natural-language responses instead of '300 results'; AI-generated, conversational titles with a more visual feel; and on listing pages, real-time, personalized highlights with the ability to ask follow-up questions.The entire journey will be AI-driven with a chat-like feel but less verbose given travel’s visual nature, yielding high personalization and materially higher conversion. (No comment on ads prerequisites.)

Q: Will the service-fee change roll out to all hosts or mainly pros?

A: We launched the single service fee for API-connected hosts (mostly PMs) last Sep. We have since tested broadly across host cohorts and begun migrating the remaining hosts, targeting full coverage by year-end.This enables simpler pricing guidance and generally exerts downward pressure on prices — adding guest value and keeping us price-competitive vs. other platforms.

Q: As hotel supply scales across cities, what are existing hosts saying?

A: Core hosts have offered limited feedback; they mostly care whether their bookings are rising, and they are. Our message is a rising tide lifts all boats.Like Amazon learned that selling DVDs and CDs helped books, each new category can reinforce the existing ones. More demand lets us better match guests to the right inventory, and hotels are proving incremental to host bookings.

Q: Is personalized trip-building localized by market, or do you need a global planner?

A: Supply-side efforts are city-by-city — hotels, services, experiences roll out by city. But the software is largely global, sometimes tested in one country but not city-level.Trip planning will be a global capability with global testing.

Q: How do you balance incremental margin vs. reinvestment long term?

A: No specific 2027+ guide, but history shows strong underlying economics on both EBITDA and FCF. We went through an investment cycle last year that helped accelerate growth this year, allowing faster growth, heavy reinvestment, and margin expansion simultaneously.We will keep pursuing that playbook — invest where we see opportunity — with an apparent floor to margin stability given the track record.

Q: How does host pricing guidance determine optimal prices?

A: Pricing is one of our largest single growth levers, materially larger than RNPL. We began as a value alternative to hotels, so affordability and value are core; while we don’t set prices, showing hosts how better pricing increases earnings is highly valuable.

We are building a new AI-driven pricing model ingesting hotel prices, Airbnb prices, upcoming city events, and booking-window patterns, now rolling out to hosts. We also make price updates easier with one-tap adoption and timely prompts when events come to town.We expect this to be a major growth driver and the single biggest lever to accelerate the core.

We are undertaking a large host-side revamp of the app. Pricing and calendar are two sides of the same coin, and most hosts should not have a single static price but dynamic, daily pricing like hotels.Hotels have mature revenue management teams, but we believe AI can do even better, and we aim for a very strong model that hotels may want to use in the future.

Q: Experiences supply +80% YoY with faster booking growth, but is it material to total nights/seats?

A: Experiences are growing fast off a small base — a long-term business. Near-term growth is driven by homes, given scale, with pricing, Project Y, conversion, and more supply as key levers.Next horizon is intl expansion and hotels, representing multi-billion-dollar revenue opportunities. Services and experiences are longer-dated and smaller but with significant potential.

We measure experiences by single-market PMF rather than global growth — win 1 market, then 10, then thousands.We are in the dozens-of-markets stage with high growth; industrializing to thousands will unlock rapid scale, but that is not this year’s story.

Q: How does going AI-native affect product costs, and what ops changes mitigate impact?

A: The impact is limited. Over nearly four years since ChatGPT, most real revenue has been enterprise-side, and many consumer monetization paths remain unclear for others.

For us, no major capex is needed — we are not buying GPUs at scale. Inference costs are negligible relative to our unit economics and AOV; small conversion lifts outweigh token and inference costs by a wide margin.We stay disciplined under CTO Ahmad — focusing on product throughput, design quality, and shipped output, not token-maxing vanity.

(CFO) Tactically, the updated guide does assume a meaningful increase in AI spend for the year, while still expanding margins. Offsets are already materializing: customer support cost per booking down ~16% YoY, largely from AI, and slower headcount growth as existing teams deliver more, improving efficiency over time.

Q: What is driving the +11% first-time bookers, and are NPS-type metrics improving?

A: The acceleration reflects compounding from the product roadmap, not any single factor. Expansion markets like India and Brazil are fueling higher first-time growth, and core markets, including the U.S., have also re-accelerated in recent quarters.

RNPL is a clear driver — a highly attractive payment option, especially for users already aware of Airbnb but not yet transacting, by reducing upfront credit-card outlays for future trips.Cumulatively, improvements in sign-up/login, merchandising, and search help new users find the right home and book with confidence. (No comment on NPS.)

Q: How did innovation specifically drive the core acceleration this quarter?

A: We highlighted ~12–15 items culled from hundreds. Examples: a much more personalized homepage with higher engagement and direct bookings; smarter, more personalized search results; and a livelier map showing nearby restaurants, landmarks, and transit.

A seemingly small but high-impact change was a complete rebuild of login/sign-up across iOS/Android/Web. We had massive traffic and tens of thousands of daily login failures from friction; fixes drove large conversion gains with more flexible login options.Payments/checkout saw RNPL expansion across more countries/listings and higher prominence, more flexible cancellation, a redesigned checkout, and interest-free installments. On the host side, single service fee improved price competitiveness, and dashboards were significantly upgraded.

Management reiterated there is no single factor and urged investors to review the 12–15 items and three step-by-step visuals in the shareholder letter. Momentum is building with a very full pipeline, and contrary to fears of running out of low-hanging fruit, Project Y is uncovering bigger opportunities.The result is sustainable growth rather than a one-off boost.

Q: Long-term pricing strategy, especially with ADR now rising faster than hotels?

A: At a high level, we equip hosts to stay price-competitive to deliver value to guests while maximizing host income. Often that means encouraging lower prices; sometimes it is preventing underpricing.

Over the past two years, despite a push for affordability, nominal ADR kept rising. A key driver is 'bedroom nights' growth — larger homes remain disproportionately popular, as families and groups favor multi-bedroom stays, lifting ADR.This suggests a sustainable component of ADR growth reflecting incremental value delivered, not just price inflation, which is important for long-term pricing strategy.

Q: Among non-lodging services, which newer categories show the strongest early traction and unit economics, and how will you pace investment?

A: Car rentals are by far the largest given the asset scale, with a positive surprise of longer booking durations — even longer than lodging on average. We initially expected shorter durations but are seeing longer use cases and will scale rentals globally.

Luggage storage is an unexpected hit. Not every service is sexy, but each can deliver value; some, like car rentals and luggage storage, have clearer revenue impact, and all make users more likely to book lodging with Airbnb.

We have dozens of services in the pipeline. Some will be big, like groceries, with intl expansion and delivery integrations via partners — we will not build delivery ourselves but integrate it tightly into the app with offers and discounts.We expect significant expansion with limited cost because most are partner-led, where fulfillment costs sit with partners and Airbnb provides demand, keeping our costs low. First-party host services will be city-led, reflecting local preferences.

Q: With AI-led product innovation and hotel scale-up, will you pursue more formal B2B opportunities?

A: Using a tech analogy, Apple’s consumer-first model won over BlackBerry’s enterprise-first. We believe consumers will pull enterprises, so B2B matters but starts with consumer preference, then we make it easy for employers to adopt.We are unlikely to pursue white label given the strength of our brand.

<End of text>

Risk disclosure and statement:Dolphin Research Disclaimer and General Disclosure