Booking Holdings

Shares of leading online travel agency companies Airbnb, Booking, and Expedia have declined about 10% during the last few trading sessions on fears that wide-moat Meta's AI agent, Muse, will disintermediate the platforms.

Why it matters: Our experience using Muse shows that the agent used Booking and Airbnb to search for travel accommodations. This supports our view that AI agents will depend on the supply, trust, and conversion that these platforms present. For example, when we asked Muse to find an accommodation during specific dates in a US city for three adults it immediately went to Booking's platform. When we then asked it to just search for vacation rentals in that city during those dates it shifted its search to Airbnb's network. We also found Muse's responses to be slow and circuitous, requiring additional prompts and time to narrow down the choices. We expect Google and other companies to have their own agents that should also use these leading online travel company platforms.

The bottom line: We are maintaining our fair value estimates of $185 per share for wide-moat Airbnb, $217 for wide-moat Booking, and $262 narrow-moat Expedia. We think investors have overestimated AI agents' impact on the relevance of these companies. We see Airbnb as best positioned. We believe Airbnb's alternative accommodation business is particularly protected from the AI threat, as most of its 5.5 million hosts are individuals with no websites for AI search engines to locate. We also see Booking as insulated, with about 10% of its accommodation room nights from brand chains that could see more direct bookings from AI use in the future. Meanwhile, we estimate Expedia's exposure to brand chains is about 20%-25% of its accommodation room nights.

After reviewing second-quarter results, we have increased our fair value estimate to $217 per share from $211 due to the time value of money. Our fair value estimate implies a 2027 enterprise value/adjusted EBITDA multiple of 16 times. The key drivers of our financial model are agency and merchant booking growth and online and offline advertising expenses.

Booking's second-quarter revenue increased 8%, ahead of its 4%-6% guidance, driven by a 5% lift in room nights. Adjusted EBITDA increased 9%, aided by decreased labor expense. The high-single digit revenue growth target for 2026 was maintained. We model 2026 revenue growth of 9%. We expect revenue growth to average a high-single-digit percentage through 2030, as we believe AI will further enhance the company's already strong competitive positioning.

Booking announced plans to extract $650 million in costs by 2027, or about 4% of 2025 operating expenses, which it will reinvest in its network. We see the cost initiative, driven by reductions in the company’s workforce and real estate portfolio, and via procurement and technology development efficiencies, as achievable. We don’t expect these cuts to sacrifice booking growth; in fact, they should elevate demand on the company's platform, aided by the reinvestment of some savings in flight, vacation rental, experience, payment, and AI offerings. Additionally, we believe Booking will leverage its marketing and customer service expenses, driven by revenue scale and aided by AI. We model operating margin will reach 38% in the next 10 years from 33% in 2025.

We model total, agency, and merchant booking growth to average 7%, 1%, and 9%, respectively, over 2026-35 as the company benefits from reinvestment of planned cost savings, secular investment in AI, onshoring manufacturing, and infrastructure in the US that should spark economic growth, further demand recovery in Asia-Pacific, and an ongoing increase in middle-income households (around 1 billion have been added to the global population during the last 10 years, according to Oxford Economics). Our forecast is aided by online alternative accommodation industry bookings growing by about 10% during 2026-30. We see Booking’s online vacation rental room night share in the 30s.

We rate Booking’s moat as wide, as its successful expansion into payments, flights, and alternative accommodations has strengthened its already stout platform position. Additionally, we are increasingly comfortable that generative AI/large language models will not hinder Booking’s prominent standing in the online travel industry.

In our view, Booking’s platform includes the industry’s most comprehensive array of travel content, cementing the supply side of its network effect advantage. To illustrate, as of April 2026, Booking had 4.5 million accommodation properties worldwide in over 40 languages, comprising 500,000 hotels, motels, and resorts (over 23 million rooms) and 4 million homes, apartments, and other unique places to stay (9 million listings). This compares with narrow-moat Expedia’s 3 million accommodation properties, of which more than 2 million are vacation listings, and wide-moat Airbnb’s roughly 9 million total listings, of which more than 90% are vacation rentals, with boutique hotels representing the rest. Importantly, Booking’s accommodation supply includes a leading presence in international markets, which have more boutique hotels than branded chain hotels. In Europe and Asia, independent hotels represent about 60%-65% and 50% of each market, respectively, versus only 30% in the US. The relative fragmentation of the international hotel market makes it more time-consuming and costly to aggregate and manage. Because these boutique hotels are too small to effectively market on their own, their dependence on online travel agencies for marketing and distribution is greater and stickier, in our view. Meanwhile, Booking also offers topnotch nonaccommodation content, including an industry-leading dining platform in North America through its OpenTable brand, flight content, car rental inventory, payment options, and experiences through various third-party relationships.

Much like traditional Google searches, we think agentic AI engines will draw on the trusted content from Booking's aggregated platforms, which the company has accumulated over decades by forming, servicing, and retaining relationships with suppliers. In our view, the risk that AI agents bypass Booking’s network and go directly to partners is low, as it would require AI agents to either be a merchant of record or rely on the capabilities of the online travel company’s suppliers. Google said in November that it does not aspire for its AI engine to be a merchant of record, which is in line with its traditional search model, and it is likely OpenAI will look to monetize its heavy investment through an advertising model and not seek to expend the capital needed to be a merchant. Specifically, being a merchant of record requires developing a payment network to facilitate transactions, as well as servicing supplier relationships and complying with regulations, which would take time for others to produce. Booking has developed a network of more than 100 local payment methods across dozens of currencies, which the firm's boutique and rental partners (90% of nights) often lack, making it challenging for AI engines to reach these suppliers directly. Specifically, we believe Booking's alternative accommodation business (36% of nights) is particularly protected from the AI threat, as many of its 9 million listings are individual hosts with no websites for AI searches to locate. Instead, AI search discovery will still go through trusted, aggregated networks like Booking. Additionally, Booking’s loyalty members (in the high 50s in total nights) create switching costs. Further, Booking is not sitting idle amid the AI revolution; it is leveraging its data and pristine financial position to train large language models for specific use cases across its platform. Finally, the emergence of agentic AI’s ability to simplify the search process could increase the size of the online travel industry, expanding Booking’s addressable market and helping to offset any potential disintermediation from AI agents.

Booking’s industry-leading travel supply and AI position are driving strong user metrics, buoying the demand side of its network advantage. This global presence is shown in Booking's platform, which reserved over 1.2 billion room nights in 2025 versus 415 million at Expedia and 533 million stays at Airbnb.

The market beneath Booking, Expedia, and Airbnb is highly fragmented, making it extremely challenging for any smaller competitor or new entrant to gain customer traffic or supplier scale. In the online travel agency industry, a successful travel network not only involves attracting and retaining personnel and managing supplier and traveler relationships, but also requires significant advertising expenses to drive and sustain platform traffic. In 2025, Booking spent $8.2 billion, or 30% of sales, on marketing, compared with $7.4 billion, or 50% of sales, at Expedia and $2.6 billion, or 21% of revenue, at Airbnb.

As the traffic scale grows, understanding of consumer behavior increases, as does the ability to enhance the platform experience, thereby improving conversion. One way to gauge how a travel platform is resonating with travelers is by tracking the percentage of traffic that comes directly to the network, as this signals that users are aware of and value the customer experience. In the second quarter of 2026, a mid-60s percentage of Booking's business-to-consumer traffic was direct.

Emerging markets should see strong online travel booking growth over the next 10 years, given middle-income households' increased online usage, and Booking is well positioned.

Mobile application usage is increasing rapidly, and Booking has a dominant global position, which aids the more than 50% of room nights that come from direct traffic.

Booking is strengthening its network effect through organic initiatives and in fast-growing markets like experiences, vacation rentals, flights, and payments, resulting in a fully connected trip.

Google's continued emphasis on placing its paid ads and metasearch platform ahead of free organic search links could place marketing cost pressure on Booking Holdings.

Consumers could seek large-scale AI language models to help plan their travel, which could increase marketing expenses for Booking.

Booking gets a large percentage of bookings from Europe (around 50% of its room night bookers), which could be nearing a more mature growth phase relative to emerging markets.