Booking Holdings
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
The world's largest online travel agency, with a two-sided hotel-and-alternative-accommodation network that independent European properties cannot replace and a Genius/direct-channel mix that is slowly reducing Google dependence — durable, not impregnable, and still more paid-search exposed than Airbnb.
Booking's moat is supply density in global accommodations plus a growing merchant-and-loyalty stack — a two-sided marketplace that is the default hotel search in much of Europe, with Connected Trip and Genius as the compounding layers:
- Supply Density, Especially Independent Hotels: Booking.com, Agoda, Priceline, and Kayak together are the largest OTA platform globally. Independent hotels and alternative accommodations — ~37% of Booking.com room nights — depend on Booking's demand in a way chains do not, because they have no comparable brand-dot-com funnel. That supply density pulls demand; demand pulls more supply. Airbnb is the verb in homes; Booking is the verb in hotels outside North America. The US remains the open flank: high-single-digit room-night growth is real progress, not yet category leadership.
- Genius and the Direct Mix: B2C direct mix has sat in the mid-60% range over the trailing four quarters, and the mobile app is high-50% of room nights — both up year over year even as SEO remains under pressure across consumer internet. Genius Level 2 and 3 members are more than 30% of the active customer base and a high-50% share of room nights; they book further ahead, return more often, and convert more on direct. That is the mechanism that makes the Google tax optional rather than structural — still unfinished, because the residual paid mix is large enough that a search-funnel shock would still hit growth.
- Merchant Rails and Connected Trip: Merchant gross bookings are ~73% of the total, up about four points year over year, and merchant revenue grew 15% in Q2 to $5.13B. The payments platform is what lets Booking stitch flights (+4% tickets despite Middle East capacity cuts), attractions (double digits), and stays into a Connected Trip — already a low-double-digit share of Booking.com transactions, growing more than twice the platform rate. Multi-vertical bookers return more frequently. The bundle is early, not locked; it is the path from a hotel search box to a travel operating system.
Moat Verdict
Booking is a net AI beneficiary on operations — customer-service cost per booking is falling at a double-digit rate via Voice AI, and Penny/inspiration tools are being tested as discovery surfaces — while remaining a net risk on distribution, because a material slice of high-intent traffic still arrives via Google. The AI-resilient moats are the accommodations network and proprietary booking data; the AI-vulnerable ones are the learned search UI and any remaining paid-search dependence. Durability in the AI era tracks whether Genius and Connected Trip keep pulling that residual paid mix onto owned channels.
74.1 resilient · 65.0 vulnerable · 80/20 = 72.2 · = 72
Open a moat to read its note.
Genius, saved trips, and the Booking.com/Priceline/Agoda apps create real habit — app mix is high-50% of room nights — but the core search-and-filter UI is substitutable, and agentic assistants (including Booking's own Penny) lower the switching cost further.
Ranking, Genius scoring, cancellation prediction, and merchant-payment routing are differentiated systems trained on decades of bookings; they are not a unique ontology a competitor cannot eventually replicate with enough demand data.
Hotel and flight inventory is multi-homed across OTAs; Booking does not control a scarce public data source.
Marketplace, payments, and ML talent is broadly available; AI coding tools further reduce any scarcity advantage in ranking and app engineering. The moat does not rest on talent scarcity, so the pillar is not applicable rather than weakened, the Shopify and Airbnb precedent. Previously weakened.
Connected Trip (stays + flights + cars + attractions) is a real bundle at a low-double-digit share of Booking.com transactions, growing more than twice the platform; Genius is the loyalty glue. Early, not locked-in.
Search, booking and pricing data across the largest hotel marketplace feeds ranking and personalisation, but Expedia and Airbnb hold the same class of data at comparable scale and Google sees the searches first. Real and useful, not a dataset rivals cannot replicate — the same bar as Airbnb. Re-rated from strong to intact in the proof-point pass.
Travel and payments regulation is a cost of doing business, not a switching barrier; EU DMA/DSA pressure on Google is a possible tailwind, not a Booking lock-in.
Largest global accommodations marketplace, at more than 1.1 billion room nights a year: independent-hotel supply density attracts demand, and demand attracts more supply. Strongest in Europe; the US is still a share-gain story. Classic two-sided OTA economics, with multi-homing as the cap.
Merchant mix ~73% of gross bookings and a stored-payment / Genius profile make Booking the checkout layer for a growing share of trips, but travel is episodic and travellers still multi-home — not a daily-spend embed like a card network or DashPass.
Genius identity, trip history, and the partner extranet are the de-facto record of many independent hotels' demand and of frequent travellers' stays; they are not the hotel's PMS, so migration risk is real even if inconvenient.
Scale here is the two-sided network itself, already rated under networkEffects; crediting it again as a cost lead would double-count the same evidence.
Booking.com has shifted a growing share of room nights to direct and app channels, reducing rented demand, but a material slice of high-intent traffic still arrives via Google — a strong brand, not yet a default that no longer pays for reach.
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
The world's largest online travel agency, with a two-sided hotel-and-alternative-accommodation network that independent European properties cannot replace and a Genius/direct-channel mix that is slowly reducing Google dependence — durable, not impregnable, and still more paid-search exposed than Airbnb.
Growth Score
Q2 2026 (reported August 4) beat the high end of guidance on room nights, gross bookings, revenue, and adj EBITDA: room nights 325M (+5%), gross bookings $51.0B (+9% / +8% cc), revenue $7.4B (+8% / +7% cc), adj EBITDA $2.6B (+9%) at 36.0% margin. Adj EPS $2.54 (+15%) outran EBITDA because average share count fell 6%. Domestic room nights grew high single digits; international only slightly, as long-haul stayed pressured by Middle East flight capacity and ticket prices. FY26 guide is high-single-digit gross bookings, revenue, and adj EBITDA, with adj EPS growth in the low-to-mid teens; Q3 room nights 3–5% and top-line 4–6%. Transformation run-rate savings were raised to ~$650M by end-2027.
Valuation Score
At ~$148 (September 24, 2026; post the April 2 25-for-1 split) BKNG trades at ~$112B, ~14× Street FY26 EPS of ~$10.45 and ~12× FY27 EPS of ~$12.4 — the lowest multiple in this file's history. The stock fell through the old $155 bear on two new overhangs rather than on the numbers: the EU General Court upheld the block of the €1.63B ETraveli deal, and shares dropped ~5% on each of September 23–24 after Expedia partnered with Meta's Muse AI agent, read as agentic-booking disintermediation risk. Price now sits between the reset $115 bear and the $210 base, below the post-split 52-week low of ~$150 and far under the ~$237 Street mean target.
The Default-Hotel Marketplace Moat
Booking's moat is supply density in global accommodations plus a growing merchant-and-loyalty stack — a two-sided marketplace that is the default hotel search in much of Europe, with Connected Trip and Genius as the compounding layers:
- Supply Density, Especially Independent Hotels: Booking.com, Agoda, Priceline, and Kayak together are the largest OTA platform globally. Independent hotels and alternative accommodations — ~37% of Booking.com room nights — depend on Booking's demand in a way chains do not, because they have no comparable brand-dot-com funnel. That supply density pulls demand; demand pulls more supply. Airbnb is the verb in homes; Booking is the verb in hotels outside North America. The US remains the open flank: high-single-digit room-night growth is real progress, not yet category leadership.
- Genius and the Direct Mix: B2C direct mix has sat in the mid-60% range over the trailing four quarters, and the mobile app is high-50% of room nights — both up year over year even as SEO remains under pressure across consumer internet. Genius Level 2 and 3 members are more than 30% of the active customer base and a high-50% share of room nights; they book further ahead, return more often, and convert more on direct. That is the mechanism that makes the Google tax optional rather than structural — still unfinished, because the residual paid mix is large enough that a search-funnel shock would still hit growth.
- Merchant Rails and Connected Trip: Merchant gross bookings are ~73% of the total, up about four points year over year, and merchant revenue grew 15% in Q2 to $5.13B. The payments platform is what lets Booking stitch flights (+4% tickets despite Middle East capacity cuts), attractions (double digits), and stays into a Connected Trip — already a low-double-digit share of Booking.com transactions, growing more than twice the platform rate. Multi-vertical bookers return more frequently. The bundle is early, not locked; it is the path from a hotel search box to a travel operating system.
Moat Verdict
Booking is a net AI beneficiary on operations — customer-service cost per booking is falling at a double-digit rate via Voice AI, and Penny/inspiration tools are being tested as discovery surfaces — while remaining a net risk on distribution, because a material slice of high-intent traffic still arrives via Google. The AI-resilient moats are the accommodations network and proprietary booking data; the AI-vulnerable ones are the learned search UI and any remaining paid-search dependence. Durability in the AI era tracks whether Genius and Connected Trip keep pulling that residual paid mix onto owned channels.
74.1 resilient · 65.0 vulnerable · 80/20 = 72.2 · = 72
Open a moat to read its note.
Genius, saved trips, and the Booking.com/Priceline/Agoda apps create real habit — app mix is high-50% of room nights — but the core search-and-filter UI is substitutable, and agentic assistants (including Booking's own Penny) lower the switching cost further.
Ranking, Genius scoring, cancellation prediction, and merchant-payment routing are differentiated systems trained on decades of bookings; they are not a unique ontology a competitor cannot eventually replicate with enough demand data.
Hotel and flight inventory is multi-homed across OTAs; Booking does not control a scarce public data source.
Marketplace, payments, and ML talent is broadly available; AI coding tools further reduce any scarcity advantage in ranking and app engineering. The moat does not rest on talent scarcity, so the pillar is not applicable rather than weakened, the Shopify and Airbnb precedent. Previously weakened.
Connected Trip (stays + flights + cars + attractions) is a real bundle at a low-double-digit share of Booking.com transactions, growing more than twice the platform; Genius is the loyalty glue. Early, not locked-in.
Search, booking and pricing data across the largest hotel marketplace feeds ranking and personalisation, but Expedia and Airbnb hold the same class of data at comparable scale and Google sees the searches first. Real and useful, not a dataset rivals cannot replicate — the same bar as Airbnb. Re-rated from strong to intact in the proof-point pass.
Travel and payments regulation is a cost of doing business, not a switching barrier; EU DMA/DSA pressure on Google is a possible tailwind, not a Booking lock-in.
Largest global accommodations marketplace, at more than 1.1 billion room nights a year: independent-hotel supply density attracts demand, and demand attracts more supply. Strongest in Europe; the US is still a share-gain story. Classic two-sided OTA economics, with multi-homing as the cap.
Merchant mix ~73% of gross bookings and a stored-payment / Genius profile make Booking the checkout layer for a growing share of trips, but travel is episodic and travellers still multi-home — not a daily-spend embed like a card network or DashPass.
Genius identity, trip history, and the partner extranet are the de-facto record of many independent hotels' demand and of frequent travellers' stays; they are not the hotel's PMS, so migration risk is real even if inconvenient.
Scale here is the two-sided network itself, already rated under networkEffects; crediting it again as a cost lead would double-count the same evidence.
Booking.com has shifted a growing share of room nights to direct and app channels, reducing rented demand, but a material slice of high-intent traffic still arrives via Google — a strong brand, not yet a default that no longer pays for reach.
Growth Analysis
Growth Drivers
Key Risk
If Google AI Overviews and Gemini trip-planning capture a material share of high-intent hotel queries by end of 2027 and Booking's mid-60% B2C direct mix fails to offset, paid-channel CAC rises and room-night growth stalls below 3% even after the Middle East long-haul drag fades.
Score Derivation
71.4 base + 4 margin − 5 risk = 70
Base 71 (7–11% CAGR, midpoint 9%) + 0 trajectory (merchant/Connected Trip accelerating; room nights decelerating on ME long-haul) + 4 expanding margin (adj EBITDA +40bps, Transformation $650M) − 5 moderate Google/AI-funnel risk = 70
Price Scenarios (12–24 Months)
Valuation Multiples
| Trailing P/E (GAAP) | ~16× |
| Forward P/E (FY26) | ~14× |
| Forward P/E (FY27) | ~12× |
| Price / Sales (FY26) | ~3.8× |
| Price / FCF | ~15× |
The multiple has compressed from ~20× to ~14× FY26 in six weeks without an estimate cut of matching size: the market is pricing a terminal-value question (do AI agents like Meta's Muse become the booking front end and tax Booking's direct traffic?) plus the loss of ETraveli as an inorganic lever. At ~12× FY27 the stock prices a meaningful share of that disintermediation risk; the base case assumes the risk is real but gradual and holds a mid-to-high-teens multiple rather than the old ~20×.
Approximate figures as of September 2026.
Where We Are vs Targets
Loading live price…
AI agents become a meaningful booking front end, Booking loses direct-traffic share and pays up for agent distribution, FY27 growth stalls, and the multiple compresses to ~11× FY26 EPS as the market prices a disintermediated OTA.
- Meta Muse and other agents, starting with the Expedia partnership, capture high-intent trip planning; Booking's direct mix stops rising and marketing as a share of gross bookings climbs
- Room-night growth falls to low single digits as Middle East long-haul stays weak and ETraveli's flight-led Connected Trip lever is gone
- ~11× FY26 Street EPS of ~$10.45 — implying ~$115
Consensus earnings hold — ~$10.45 FY26 and ~$12.4 FY27 on high-single-digit revenue growth and buybacks — AI-agent risk proves gradual, and the multiple partially recovers to ~17× FY27, below the ~20× the stock held before the September sell-off.
- FY26 revenue ~$29B (high single digits) and adj EPS growth low-to-mid teens on share-count reduction
- Booking negotiates agent distribution on terms similar to metasearch, so AI channels add cost at the margin rather than replacing direct traffic
- ~17× FY27 EPS of ~$12.4 — implying ~$210, still below the ~$237 Street mean target
Booking's supply depth and Genius loyalty make it the inventory layer AI agents book through, estimates prove conservative, and the multiple returns to ~21× FY27 as the disintermediation discount unwinds.
- Booking partners with major AI agents on favourable terms and agent-originated bookings show up as incremental room nights
- Connected Trip and merchant mix keep rising despite the ETraveli ruling; US share gains continue
- ~21× FY27 EPS of ~$12.4 — implying ~$260, near the pre-sell-off multiple