Booking Holdings Inc. (NASDAQ: BKNG) — Execution Held; the Bargain Did Not

Prepared by: Independent investment research Report date: 2026-08-12 Security: Booking Holdings Inc. common stock (NASDAQ: BKNG) Sector / sub-sector: Consumer Discretionary — Online Travel Agencies / Travel Platforms Price / size (as of 2026-08-11): $212.87 / share · market capitalization approximately $159.9B · enterprise value approximately $162.9B · 751.4M shares outstanding Fiscal year: December. Currency: USD. Headquarters: Norwalk, Connecticut.

⚡ Claude’s Take

This block is the author’s independent opinion, provided for general information. It is not investment advice. Everything below it carries no recommendation and no price target, by design.

Verdict: HOLD / own it, but do not chase it; add on a pullback toward $180–195. Estimated fair value is approximately $215–245, versus $212.87 today. Tag: “Execution held; the bargain did not.”

The business did what a high-quality platform should do under stress. In Q2, room nights increased 5%, gross bookings 9%, revenue 8%, adjusted EBITDA 9%, free cash flow 16%, and adjusted EPS 15%. Adjusted EBITDA margin widened 40 basis points to 36.0%. Booking.com grew listed properties 9% to 4.7 million, app mix moved into the high-fifties, total direct room-night mix held in the mid-fifties, and Connected Trip transactions grew low double digits. Management also raised the Transformation Program’s annual run-rate savings target from $550M to $650M. There is still no numerical evidence that an AI assistant is taking meaningful transaction share: management says combined paid and unpaid LLM traffic is significantly below 1% of room nights.

But Q2 did not satisfy the June bull case’s key test. The prior report required second-half room-night growth to reaccelerate toward high single digits or low double digits as the Middle East shock normalized. Instead, Q2 room nights grew 5.3%, management guided Q3 to 3–5%, and it still assumes indirect conflict effects through Q3. Alternative-accommodation room nights grew only 4%, flights decelerated to 3.7%, car-rental days fell 6.5%, marketing grew 11% against 9% gross-bookings growth, and the filing now says SEO traffic should decline in the short-to-medium term. The moat remains intact, but the operating evidence supports a slower algorithm than the clean “8-8-15” shorthand.

The quote has also changed the decision. Since the June baseline near $162, BKNG has appreciated roughly 31% and now trades at about 17.1x trailing free cash flow, 16.3x trailing EBITDA, and roughly 20x expected 2026 adjusted EPS. AZI places its own-history P/E/P/S composite at the 62nd percentile, versus roughly the 39th percentile in June. At this price, investors are no longer being paid handsomely to discover whether AI, Google, and European regulation weaken the demand-acquisition layer. They are paying a fair price for a great asset and underwriting continued high-single-digit gross-bookings growth, modest margin expansion, and persistent buyback accretion.

Conviction: medium-high on business quality, medium on the stock. The next fact that would make me more constructive is room-night growth returning above 7% while marketing stays at or below roughly 4.6% of gross bookings and direct mix rises. The fact that would make me materially more cautious is two quarters of sub-5% room-night growth paired with a falling take-rate proxy or rising paid-marketing intensity outside a documented macro shock. AI traffic becoming material while Booking’s direct mix falls would be the clearest structural bear signal.

Changes since 2026-06-10

• Call changed from BUY to HOLD. The business is somewhat stronger, but the stock has moved into the prior report’s $210–240 fair-value range.
• The temporary-shock thesis is only partially confirmed. Middle East booker demand largely normalized, but airline capacity, ticket-price, and long-haul effects persist through Q3; guided room-night growth is 3–5%.
• The moat evidence held. Total direct room-night mix remained in the mid-fifties; the higher B2C-only measure excluding B2B remained in the mid-sixties. App and Genius Level 2/3 room-night mix each reached the high-fifties, supply expanded, and LLM-sourced room nights remain immaterial.
• The Google risk became more concrete. Organic-search room nights remain small, yet the filing explicitly expects SEO traffic to decline and marketing intensity rose modestly.
• Capital return accelerated. First-half repurchases were $7.76B, reducing July 27 shares outstanding to 751.4M; the balance sheet remains conservatively net-levered, but buyback returns are less compelling at $213 than at $162.
• Regulatory and operational tails widened. The quarter added an FTC staff complaint recommendation involving Priceline and disclosed an April guest-data incident, alongside the existing DMA, Spain, Switzerland, and mass-claim exposures.

Stock Price Action — Five-Year Event Map

Text timeline: pandemic recovery and European-war shock → rate/inflation de-rating → 2023 earnings recovery → 2024 guidance resets → 2025 peak → AI/Middle-East drawdown → Q2-2026 rebound.

BKNG completed a 25-for-1 split in April 2026. Pre-split traded closes below are divided by 25; return calculations use split-adjusted closes. Event attribution is an interpretation checked against contemporaneous filings, news, travel peers, and SPY—not proof from price alone.

The five-year split-safe high was $232.64 on July 7, 2025. The trailing-52-week low/high were $154.13/$228.83; at $212.87, BKNG is 7.0% below the 52-week high and 8.5% below the five-year high. The stock is 9.97% above its 21-day EMA and 15.64% above both the 50- and 200-day EMAs. Returns are +19.3% over one month, +32.9% over three months, +24.9% over six months, and approximately flat over twelve months. The 50-day EMA has only just crossed above the 200-day by two cents, too small to call a durable signal.

FactorsToday estimates market beta around 1.03, negative Momentum exposure (−0.206), and no active Quality loading under its sparse model; annualized stock-specific volatility is 26.0% and the model explains 47.3% of variance. Five-year realized volatility is about 33% with a 39.5% maximum drawdown. The strong tape is therefore a sharp recovery with substantial idiosyncratic risk, not an unambiguous crowded momentum regime.

1. Executive Summary

Booking Holdings is the world’s largest online travel intermediary by gross bookings and the strongest scaled OTA outside China. Its economic engine is simple but unusually powerful: aggregate fragmented hotel and alternative-accommodation supply, convert high-intent global travel demand, collect an approximately mid-teens revenue share of gross bookings, and recycle a capital-light cash stream into product, marketing, and share repurchases. Booking.com is the profit center; Priceline, Agoda, KAYAK, OpenTable, and Rentalcars.com broaden geography and vertical coverage.

The June quarter was resilient, not explosive. Gross bookings were $51.0B, up 9% reported and approximately 8% in constant currency. Revenue was $7.352B, up 8%; adjusted EBITDA was $2.648B, up 9%; and free cash flow was $3.643B, up 16%. Room nights grew 5.3% to 325M. Constant-currency ADR increased about 2%, so gross-bookings growth still had genuine unit support rather than relying only on price. Reported operating margin was 34.0% and adjusted EBITDA margin was 36.0%.

The quality signals were mostly positive. Domestic room nights grew high single digits; U.S. room nights grew high single digits and Booking’s direct channel grew again. Mobile-app mix increased to the high-fifties, and Genius Level 2/3 represented a high-fifties share of Booking.com room nights. Connected Trip transactions grew low double digits, more than twice the rate of Booking.com transactions. Booking.com listed 4.7M properties, including more than 4.1M alternative accommodations and more than 500,000 hotels, motels, and resorts. Merchant gross bookings reached 73% of total, improving payments control and traveler flexibility, though also increasing processing and chargeback expense.

The weaker signals matter. Q3 room-night growth guidance is only 3–5%. The company lowered its full-year gross-bookings outlook from the pre-conflict plan, mainly because flight-ticket expectations came down, even as it retained high-single-digit reported gross-bookings, revenue, and adjusted EBITDA growth and low-to-mid-teens adjusted EPS growth. Marketing expense rose 11%, faster than gross bookings, as traffic mix shifted and management pursued incremental paid demand. The simple quarterly revenue/gross-bookings proxy fell to 14.42% from 14.56%, although timing, mix, cancellations, and merchant accounting make it an imperfect take-rate measure. Alternative accommodation failed to gain room-night mix, remaining around 37%.

Under the Greenwald framework, Booking has a real moat, but not because travelers cannot switch apps. Consumer switching costs are low. The advantage is the combination of supply captivity, global scale, accumulated conversion data, payment infrastructure, and enough repeat/direct demand to buy traffic more efficiently than smaller rivals. The 4.7M-property supply set is difficult to reproduce with equal availability, reviews, service, local payments, and conversion. Scale lowers the unit cost of experimentation and customer support; supplier participation improves consumer choice; demand then attracts more supply. The moat is widest in European accommodation and narrowest in metasearch, flights, and U.S. alternative accommodation.

The capital-cycle read is favorable for the leader but not for every participant. AI lowers the cost of creating a travel-planning interface, inviting new discovery-layer entry. It does not cheaply create contracted live supply, payment acceptance, fraud control, customer service, or global fulfillment. Capital is therefore abundant at the visible front end and more constrained in the transaction substrate. Booking can be a beneficiary if agents route transactions to its inventory; it loses economics if agents commoditize Booking into a wholesale feed. Today, management’s LLM traffic data and partnerships with OpenAI and Google support the former possibility, but the volumes are too small to prove it.

At $212.87, the equity is no longer priced for obvious impairment. Market capitalization is approximately $159.9B using the 751.4M shares reported outstanding on July 27; enterprise value is approximately $162.9B after $17.2B cash and $20.2B debt. Trailing free cash flow is approximately $9.54B, producing a 6.0% equity free-cash-flow yield. The stock is around the 62nd percentile of its own P/E/P/S history and has rallied materially above its 21-, 50-, and 200-day exponential averages. The valuation can work if the core algorithm persists, but the margin of safety is modest.

The thesis now rests on four measurable questions: whether room-night growth recovers above the Q3 trough; whether direct, app, and Genius mix continue to offset SEO and paid-traffic pressure; whether revenue capture holds despite DMA remedies and supplier bargaining; and whether repurchases retire 4–5% of shares annually without pushing gross leverage or average purchase price beyond sensible levels. Business quality remains high. Expected return from today’s price is much more ordinary.

2. Business Overview

Booking Holdings operates a collection of travel and restaurant platforms, but economically it is one dominant accommodation marketplace surrounded by adjacent distribution assets.

Booking.com — the core franchise. Booking.com provides accommodations, flights, rental cars, attractions, taxis, insurance, and payments across a global traveler and partner network. It is especially strong in Europe’s fragmented independent-hotel market, where no single supplier can replicate the breadth of inventory or efficiently reach global demand alone. At June 30, 2026, it offered approximately 4.7M properties, up from 4.3M a year earlier. More than 4.1M were alternative accommodations and more than 500,000 were hotels, motels, and resorts. Booking.com remains the overwhelming source of group economics even though segment profit is not separately reported.

Agoda and Priceline — geographic and merchant complements. Agoda has particular relevance in Asia and provides accommodation, flights, activities, and ground transport. Priceline is a U.S.-focused discount and packaging brand and the launchpad for Penny, an increasingly agentic travel assistant. Management is consolidating the three brands’ business-to-business operations into one unit led by Agoda CEO Omri Morgenshtern. The logic is sound: combine the strongest inventory, tools, and partner capabilities instead of presenting fragmented products to banks, airlines, travel-management companies, and smaller affiliates. The unit-economics caution is equally important. Management explicitly says B2B growth must be incremental because wholesale distribution can cannibalize richer B2C transactions.

KAYAK and OpenTable — useful but not central. KAYAK is a metasearch business exposed to both Google’s travel products and generative-search substitution; the $457M 2025 impairment is an economic warning, not merely an accounting adjustment. OpenTable provides restaurant discovery, reservation, and management software. Both extend traveler engagement, but neither changes the conclusion that Booking.com’s accommodation marketplace carries the valuation.

The company monetizes through three reported revenue lines:

The merchant transition is structurally significant. Merchant gross bookings increased 14.5% to $37.0B while agency gross bookings declined 3.3% to $14.0B; merchant mix reached 73%, up four points. Merchant processing lets Booking offer more currencies, payment timing, and flexibility, supports the Connected Trip, and creates card-rebate and float economics. It also raises processing, fraud, chargeback, and service costs and makes reported revenue grow differently from gross bookings. Over the trailing twelve months, the filing says incremental payment-facilitation revenue exceeded associated incremental variable expense. That confirms positive economics, not a free margin lunch.

The operating model is capital-light but marketing-intensive. Booking spends primarily when a traveler is near purchase, using search engines, affiliates/B2B, metasearch, social media, and brand media. Q2 marketing was $2.371B, 32.2% of revenue and 4.65% of gross bookings. The variable acquisition model protects cash in a demand shock, but Google remains both a critical distributor and a potential rent extractor. Direct app and repeat traffic are therefore not vanity metrics: they determine how much gross profit Booking keeps.

The Connected Trip is the strategic attempt to increase traveler frequency and lifetime value. A customer who books accommodation, a flight, an attraction, ground transport, and insurance in one itinerary creates more transactions, more first-party data, more service opportunities, and more reasons to return through the app. Connected Trip transactions grew low double digits in Q2, more than twice Booking.com’s total transaction growth. Attractions grew double digits and flight tickets grew 3.7%. The current economics are mixed because non-accommodation services carry lower margins; the long-term payoff must come from retention, cross-sell, and lower acquisition cost rather than near-term mix alone.

Verdict. This remains an unusually attractive platform model: fragmented supply, global demand, capital-light fulfillment, high cash conversion, and multiple routes to increase lifetime value. The business is not a software annuity—travel is discretionary and each transaction must be won—but its breadth, conversion, and payment capability create recurring marketplace economics at scale.

3. Industry Dynamics

The relevant industry is not all travel. It is online accommodation intermediation by geography and customer segment, with adjacent flight, car, attraction, payment, and B2B distribution layers. Accommodation supply is fragmented; demand discovery is concentrated; fulfillment is operationally complex. That combination creates attractive economics for scaled intermediaries and persistent tension over who captures the rent.

Value chain and bargaining power. Hotels and alternative accommodations create the inventory. Global chains can invest in direct distribution and loyalty; independent properties need international traffic, translation, payment acceptance, fraud management, and round-the-clock support. Booking, Expedia, Airbnb, and regional platforms connect that supply to travelers. Google, metasearch, social media, and AI assistants sit above the platforms as discovery gateways. The profit pool follows whichever layer controls high-intent traffic, checkout, servicing, and repeat data.

Supplier bargaining power is asymmetric. A Marriott or airline can negotiate, drive loyalty traffic direct, and multi-home. A small European hotel can list on multiple platforms but cannot reproduce Booking’s global demand or conversion stack. Consumer bargaining power is high because comparison is easy, although saved preferences, reviews, Genius discounts, and app convenience create habit. Google has high bargaining power over paid acquisition. The OTA’s structural job is to keep enough demand direct and enough supply fragmented that neither side captures the entire spread.

Competitive map. Booking is strongest in European and cross-border hotels, Expedia in U.S. travel and embedded B2B distribution, Airbnb in branded alternative accommodations, and Trip.com in China and Asian corridors. Q2 unit growth was not favorable to Booking: Airbnb nights and seats grew 10%, Expedia room nights 6%, and Booking room nights 5.3%. These definitions and geographies differ, so the comparison is a warning rather than a precise share calculation. Booking’s economic lead remains clear: its 36.0% Q2 adjusted EBITDA margin compares with Expedia’s 25.9% and Airbnb’s roughly 35%.

Airbnb is the cleanest consumer-discovery competitor. Its app produced 64% of nights, app nights grew 23%, and hotel nights are growing roughly three times faster than homes, albeit from a single-digit share of volume. This threatens Booking at the edge where differentiated brand demand meets hotels. Airbnb still lacks Booking’s global hotel depth, agency/merchant flexibility, and B2B distribution, but its lower Google dependence is strategically valuable.

Expedia is the closest full-stack OTA rival and the benchmark for B2B. Its Q2 bookings grew 12%, revenue 14%, EBITDA 23%, and room nights 6%; B2B achieved a twentieth consecutive quarter of double-digit growth. Expedia also bought conversational planner Layla and is expanding Google and ChatGPT participation. Booking’s decision to consolidate three B2B operations is partly a response to that momentum. Booking retains superior margins and supply depth, but it cannot treat wholesale distribution as an uncontested adjacency.

Trip.com is the regulatory analogue. China’s competition and algorithm remedies show how a government can change marketplace conduct and monetization, rather than merely impose a fine. Its local moat is durable but monetization is constrained. That is relevant to Europe’s DMA and national commission cases: a platform can retain volume while losing some economics.

Marathon capital-cycle view. This is a late, contested cycle, not a classic hotel-capacity overbuild on Booking’s balance sheet. High OTA returns are attracting capital into the demand interface: Google AI Mode, general-purpose assistants, Airbnb hotels and services, Expedia’s B2B and AI acquisitions, and loyalty subsidies across the industry. AI makes it inexpensive to build an attractive itinerary interface. It does not make it inexpensive to contract 4.7M properties, maintain live availability, localize content and payments, manage fraud and chargebacks, provide service, and comply with dozens of regulators.

The supply response is therefore concentrated where Booking is weakest—discovery—and constrained where it is strongest—transaction infrastructure. That predicts pressure on acquisition rents before it predicts disappearance of OTAs. Regulation accelerates the same cycle by removing parity restrictions and widening business-user data access. The most likely mean-reversion mechanism is a higher rent paid to traffic or distribution partners and a lower share of the traveler relationship, not a new entrant recreating Booking.com from scratch.

Structural growth and cyclicality. Travel demand grows with global income and the continued shift from offline to online booking, but it remains discretionary and shock-prone. Pandemics, wars, airline capacity, foreign exchange, and consumer confidence can interrupt the trend. Booking bears little owned-asset operating leverage, so variable marketing and minimal capex protect cash; it still experiences booking cancellations, partner failures, and working-capital movements. The business is secularly supported but cyclically exposed.

Verdict. Industry structure is favorable for scaled full-stack intermediaries and unfavorable at the demand gateway. Booking owns scarce transaction and supply infrastructure, but Google, AI agents, branded rivals, and regulators contest who owns discovery. The correct industry assumption is durable high returns with gradual pressure—not winner-take-all expansion and not imminent commoditization.

4. Competitive Position & Moat

Booking has a durable competitive advantage, but describing it as a wide consumer-switching-cost moat overstates the case. Under Greenwald’s tests, the advantage is economies of scale combined with partial supplier and traveler captivity.

Scale economics—the primary barrier. A new competitor can build a travel website or AI itinerary tool. It cannot quickly reproduce 4.7M properties with live rates and availability, hundreds of millions of reviews and stays, localized content, fraud controls, merchant payments, cross-border tax capability, partner tools, and global service. Booking spreads those fixed and semi-fixed investments across $51B of quarterly gross bookings. The 36% adjusted EBITDA margin, positive incremental payment economics, and ability to absorb more AI/cloud investment while expanding margin are the observable return on that scale.

Scale alone does not prevent entry; it can become a commodity if customers are perfectly mobile. Booking couples it with several forms of soft captivity:

• Traveler habit and loyalty. B2C direct mix held in the mid-sixties and direct room nights grew. Higher-tier Genius customers exceeded 30% of active customers and generated a high-fifties share of room nights. Saved payment details, prior-trip records, reviews, discounts, and service history reduce friction, but a traveler can still compare elsewhere at negligible cost.
• App engagement. App room-night mix rose to the high-fifties. Most app nights are direct, and repeat behavior is favorable. The app reduces exposure to search, though mobile bookings can have shorter stays and lower ADRs.
• Supplier dependence. Independent properties receive incremental global demand and outsource localization, payments, fraud, and service. They still multi-home and increasingly can offer better direct terms under the DMA, so this is reliance rather than exclusive lock-in.
• Connected Trip behavior. Multi-vertical transactions grew low double digits and users reportedly return more frequently. This can deepen captivity if a complete itinerary is easier to manage in one account, but the disclosed transaction mix is still only low double digits and unit economics are not broken out.

The resulting loop is powerful: more supply improves choice and conversion; higher conversion attracts traveler demand; greater volume improves marketing, service, payment, and experimentation economics; better tools and demand attract more supply. Both sides multi-home, so this is a scale-and-liquidity flywheel rather than a hard winner-take-all network effect.

Profitability test: pass. Commodity intermediaries do not normally sustain a mid-thirties EBITDA margin while investing and competing globally. Booking’s margin expanded in Q2 despite paid-marketing pressure. Revenue capture remained approximately 14.4%, only around ten basis points below the prior year on a measure affected by timing and mix. No current financial fingerprint says the moat has broken.

Share-stability test: amber. Public data do not yield a clean global share series, but Booking’s 5.3% room-night growth trailed Airbnb’s 10% nights/seats and Expedia’s 6% room nights. Alternative-accommodation nights grew 4% and mix stayed at 37%. Those facts do not prove share loss—the companies serve different corridors and products—but they show Booking was not the unit-growth leader. The moat is delivering returns more clearly than incremental share.

Where the moat is strongest. European independent accommodation, cross-border travel, global payment/service capability, and high-intent fulfillment. The fragmentation and localization burden are highest there, making Booking most valuable to both sides.

Where it is weakest. Top-of-funnel inspiration, U.S. alternative accommodations, metasearch, flights, and generic trip planning. KAYAK’s 2025 impairment is the clearest evidence that a discovery asset can deteriorate even while Booking.com’s transaction moat remains healthy.

AI pressure test. Management directly attributes some SEO pressure to Google’s AI Overview, and the filing expects SEO traffic to decline in the short-to-medium term. SEO room nights remain a small share; direct mix did not fall; and combined paid and unpaid LLM traffic is significantly below 1% of room nights. Booking says third-party data give it the majority of organic LLM travel referrals and is testing paid OpenAI traffic on a cost-per-click basis. It is also an early partner in Google’s limited agentic booking product.

That is an early-warning configuration, not current disintermediation. AI can improve Booking’s ranking, partner messaging, coding, and customer service—the latter’s cost per booking is falling double digits. It can also become a new toll collector. If an assistant owns identity, comparison, and transaction routing, Booking may be one interchangeable inventory API. If Booking remains merchant and servicer of record, agents could expand efficient demand and strengthen the scaled fulfillment layer. Volumes are too small to choose confidently between those outcomes.

Regulatory pressure test. Booking has been subject to DMA obligations since November 2024, including allowing hotels and car rentals to offer better prices or terms elsewhere and providing business users more data access. The European Commission says these changes are producing lower prices and better conditions for users. That intentionally reduces supplier captivity. The Q2 filing shows no broad take-rate break, but national remedies now include a suspended Swiss order to reduce hotel commissions for three years after a final adverse judgment, the Spanish ranking/parity case, and a French Agoda conduct order. Regulation is an economic risk, not only a fine.

Greenwald verdict: durable, but narrower at the discovery edge. Scale plus partial captivity should preserve above-normal returns. The moat does not protect every acquisition channel or guarantee share gains. Its health should be judged through direct/app/Genius mix, marketing per gross-bookings dollar, stable revenue capture, and contribution economics—not property count alone.

5. Growth History & Forward Opportunities

Growth is still organic, broad, and cash-generative, but Q2 exposed the difference between the long-run algorithm and the next few quarters.

The long-run algorithm. Management’s shorthand is approximately 8% gross-bookings growth, 8% revenue growth, and 15% adjusted EPS growth. The bridge is plausible: unit growth plus ADR and ancillary mix drive gross bookings; a roughly stable revenue-capture rate converts that to revenue; modest fixed-cost leverage lifts adjusted EBITDA slightly faster; and buybacks reduce diluted shares by roughly 4–6% annually. The first half broadly matched the last step—adjusted EPS increased 14%—even though room-night growth slowed.

The arithmetic also exposes the dependency. Approximately one-third to one-half of the gap between high-single-digit operating growth and mid-teens per-share growth can come from shrinking the denominator. That is real value creation when the stock is repurchased below intrinsic value. It is less powerful when shares are bought at a full multiple, and it cannot rescue a structurally declining cash stream. The correct KPI stack is therefore room nights, gross bookings, revenue capture, adjusted EBITDA, free cash flow, and net share count—not EPS alone.

Q2 operating scorecard. Units grew, but every travel vertical did not participate equally:

Regional performance argues against a broad franchise break. Domestic room nights grew high single digits and international nights were slightly positive. Europe grew mid single digits, with intra-Europe travel high single digits. Asia grew mid single digits, with domestic Asia low double digits. The U.S. grew high single digits, and Rest of World improved to mid single digits after declining in Q1. This pattern is more consistent with conflict-related route disruption and uneven international demand than with a universal loss of consumer relevance.

Near-term guidance is the restraint. For Q3, management expects room nights to grow 3–5% and gross bookings, revenue, and adjusted EBITDA to grow 4–6% reported. Constant-currency gross bookings and revenue should grow 5–7%. Full-year reported gross bookings, revenue, and adjusted EBITDA remain guided to high-single-digit growth, while adjusted EPS is expected to grow low-to-mid teens. Management assumes travel conditions remain broadly stable and indirect Middle East effects—higher ticket prices, reduced capacity on certain routes, and softer long-haul demand—persist through Q3.

The prior June report anticipated a cleaner second-half reacceleration once the direct conflict impact passed. That forecast was too optimistic. Middle East-originating traveler demand largely normalized and the Q2 cancellation spike eased, but network effects through air capacity and price last longer. The thesis is delayed rather than disproved because the U.S. and domestic regional trends remain healthy. A Q4 room-night recovery above 7% would support that distinction; another sub-5% quarter without a documented shock would not.

The credible long-duration runways:

• Direct and loyalty. B2C direct room-night mix has remained in the mid-sixties, app mix reached the high-fifties, and Genius Level 2/3 mix also reached the high-fifties. More direct repeat demand lowers acquisition cost and improves personalization.
• Alternative accommodations. The inventory base is enormous and grew to more than 4.1M properties, but Q2 room-night growth of 4% shows that inventory alone does not guarantee share gains—particularly in the U.S., where Airbnb’s brand is stronger.
• Connected Trip / cross-sell. Attractions grew double digits, flights and ground transport enlarge the itinerary graph, and unified payments can reduce friction. Lower-margin vertical mix is acceptable only if it produces higher frequency and cheaper repeat acquisition.
• Payments and fintech. Merchant mix at 73% creates positive incremental variable economics, flexibility for travelers and partners, and float. It also increases fraud, processing, and working-capital complexity.
• B2B distribution. Consolidating Booking, Agoda, and Priceline’s B2B units can improve product quality and win large bank, airline, and travel-management partners. Management’s focus on incrementality is necessary because B2B has lower unit economics than B2C.
• AI-enabled service and conversion. Customer-service cost per booking continues to fall double digits while satisfaction remains high. AI expense is only a low-single-digit share of technology spend, and management says technology cost per merge request is falling. These are current cost benefits; consumer-agent upside remains optionality.

Verdict. Booking can still compound gross bookings and revenue at high single digits over a cycle, with per-share growth reaching low-to-mid teens through modest leverage and repurchases. The near-term evidence does not justify treating 8-8-15 as a quarterly entitlement. The underwriting should use mid-single-digit room-night growth as the current baseline and demand proof before restoring a faster one.

6. Financial Quality

The financial model combines exceptional cash generation with noisy GAAP net income and a working-capital balance sheet that looks stranger than the underlying risk.

Filed Q2 scorecard:

The first half generated $6.751B of free cash flow, up 7%. Trailing free cash flow is approximately $9.539B, calculated as FY2025 free cash flow of about $9.087B plus the $452M first-half year-over-year increase. At the current equity value, that is a 6.0% trailing free-cash-flow yield. This is the cleanest valuation anchor because net income is distorted by non-operating currency and equity marks while free cash flow contains predictable seasonal merchant-booking movements that wash out across a full year.

Margin quality. Marketing was $2.371B, up 11% and 32.2% of revenue. Sales and other expense increased 5%, personnel was nearly flat, G&A increased 9%, and IT increased 20%. Total operating expense increased 7%, slower than revenue. Adjusted fixed operating expense increased 6%, driven by cloud, software licenses, currency, and strategic investment. The Transformation Program produced approximately $250M of savings in 2025 and enabled a $550M annual run rate by year-end; the target is now $650M by the end of 2027. Most of the incremental $100M should arrive in 2027.

The mix trend has two offsets. Direct/app/loyalty growth should reduce acquisition cost over time, while merchant payments, lower-margin flight and attraction services, and AI/cloud investment raise variable and technology cost. Q2 shows both: personnel leverage and transformation savings widened EBITDA margin 40 basis points, but marketing intensity rose roughly seven basis points of gross bookings and IT grew faster than revenue. Management is still producing leverage, just not from every line.

Revenue capture. Dividing revenue by gross bookings gives 14.42% in Q2 versus 14.56% a year earlier. That fourteen-basis-point decline is worth monitoring but is not a clean commission-rate disclosure. Revenue timing follows traveler check-in, gross bookings are net of cancellations, merchant revenue includes payment economics, and vertical/geographic mix changes. March cancellations reduced Q2 revenue relative to bookings even though Q2 cancellation rates normalized. A persistent multi-quarter decline, especially alongside DMA remedies or higher supplier direct-booking mix, would be evidence of pressure; one quarter is not.

Balance sheet and float. At June 30, cash was $17.214B, short-term debt $2.000B, and long-term debt $18.180B, for $2.966B of net debt. That is less than 0.3x trailing adjusted EBITDA. Current liabilities included $10.121B of deferred merchant bookings, up seasonally from $5.270B at year-end. This traveler/supplier float is operational funding rather than conventional debt, but it creates refund, cyber, and liquidity obligations. Total liabilities exceeded assets, producing a $10.783B stockholders’ deficit. Negative equity is primarily the result of cumulative repurchases recorded in treasury stock, not operating insolvency; price/book and ROE are therefore meaningless.

Gross debt matters more than net debt in a severe travel shock because much of the cash supports traveler balances and international operations. Even on that harsher lens, interest coverage and cash generation are strong. First-half operating cash flow of $6.934B covered $553M of interest expense more than twelve times before capex. The company issued $2.975B of long-term debt and repaid $1.000B in the half, extending the balance sheet while funding unusually large repurchases.

Accounting quality. Q2 GAAP net income benefited from $159M of other income, while Q2-2025 suffered $962M of other expense. Adjusted net income removes currency remeasurement, equity-security marks, transformation cost, and other items. Not every adjustment is economically weightless—the recurring cost of transforming an organization and investment losses deserve scrutiny—but operating income, adjusted EBITDA, and full-year free cash flow triangulate to a business whose core earning power is improving. Stock-based compensation was $281M in the first half, and net share count still declined sharply; dilution is controlled.

Verdict. Financial quality remains elite: roughly 36% adjusted EBITDA margin, negligible capex, strong cash conversion, and low net leverage. The main watch items are paid-marketing intensity, the quality of transformation adjustments, gross debt used alongside buybacks, and whether merchant/vertical mix can preserve revenue capture.

7. Capital Allocation & Management

Capital allocation remains unusually shareholder-oriented, but the first half crossed an important line: cash returned exceeded cash generated, and gross debt filled the gap.

Repurchases. Booking spent $7.758B of cash on repurchases in the first half. The treasury-stock accounting basis was $7.814B including excise tax and included approximately 44M shares. In Q2 alone, it repurchased 21.99M open-market shares for $3.792B on the treasury basis; excluding excise tax, the earnings release reports $3.7B. The weighted monthly prices were $178.49 in April, $162.76 in May, and $170.81 in June. Those purchases were well below today’s $212.87 quote and have already created per-share value.

Shares outstanding fell from 783.2M at the proxy’s March 16 record date to 751.4M on July 27, a 4.1% reduction in roughly four months. Q2 diluted weighted-average shares fell 5.5% year over year. Adjusted net income grew 8%, but adjusted EPS grew 15%; the denominator supplied almost half of the per-share growth differential. The remaining authorization was $14.5B at quarter-end, enough to retire roughly 9% of today’s equity value.

Dividend. First-half dividends paid were $664M. The board declared a $0.42 split-adjusted quarterly dividend for September, implying an annualized yield of approximately 0.8% at the current share price. The dividend is a modest, growing baseline; repurchases remain the primary capital-return lever.

Funding and leverage. Repurchases plus dividends totaled $8.422B in the first half, equal to 124.8% of $6.751B free cash flow. Booking simultaneously issued $2.975B of long-term debt and repaid $1.000B. Gross debt rose $1.444B from year-end to $20.180B. Net debt is still only $2.966B, and the business can comfortably service it. The issue is not solvency. It is whether borrowing to repurchase stock remains value-creating as the price moves from a discounted to a normal multiple.

Under Marathon’s framework, aggressive buybacks near peak returns can signal a mature capital cycle: the incumbent has fewer high-return internal uses and distributes or leverages cash rather than expanding capacity. Booking’s case is more favorable because much reinvestment runs through marketing, personnel, cloud, and product expense rather than capex, and management is deliberately funding AI and Connected Trip. Still, a buyback above free cash flow should clear a higher valuation hurdle than one funded entirely from surplus cash.

M&A and reinvestment. Management has avoided large-scale empire building since KAYAK and OpenTable. KAYAK’s impairment reinforces that restraint. It is instead consolidating internal B2B capabilities, investing in payments and Connected Trip, and using AI for service and engineering productivity. This is rational so long as internal spending is not starved to preserve an EPS algorithm. The raised $650M savings target is meant to create reinvestment capacity; investors should watch whether technology and product outcomes—not merely adjusted-margin exclusions—materialize.

Management incentives and governance. Ten of eleven director nominees are independent. Long-term awards generally use 60% performance units and 40% restricted units. Performance units use three-year revenue and “Compensation EBITDA,” modified by relative total shareholder return; an absolute-TSR governor caps vesting at target when shareholder return is nonpositive. Compensation EBITDA expenses stock compensation and treats capex as an expense, both sensible, but no explicit free-cash-flow, ROIC, or per-share-value metric exists. CEO Glenn Fogel’s 2025 compensation was $35.4M, including $26.2M of stock awards; roughly $3.9M of tax equalization/gross-up is a governance blemish. Executive ownership was below 1%, typical for a professional-management company but not founder alignment.

Insiders. Since the prior report, disclosed open-market-coded activity consisted of three Rule 10b5-1 sales totaling $1.423M and no purchases. Across the complete trailing 24-month parsed Form 4 corpus, there were no open-market purchases; nearly all sales rows were plan-based. CFO Ewout Steenbergen and CHRO Paulo Pisano also adopted plans for future sales. Direction is negative, but plan prevalence and modest recent dollar amounts make the signal low-value rather than alarming.

Verdict. Management has an excellent record of converting cash into per-share growth and avoiding destructive acquisitions. First-half repurchases were well timed relative to today’s price. The future return is more sensitive to price discipline: capital return above free cash flow, partly funded with debt, is acceptable at a discount and increasingly debatable at full value.

8. Changes and Headwinds — Last Two Years

The last two years moved Booking deeper into payments and cross-vertical fulfillment while simultaneously increasing pressure at the discovery and regulatory edges.

Merchant transformation. Merchant gross bookings rose from roughly two-thirds of total to 73% in Q2-2026. This moves Booking from commission agent toward merchant infrastructure: it collects from travelers, remits partners, manages more currencies and payment timing, and earns card/payment economics. It also increases chargeback, fraud, service, and processing exposure. The filing’s statement that incremental payment revenue exceeded incremental variable expense over the trailing year is the key proof that the transition creates contribution, not merely gross accounting revenue.

Connected Trip moved from concept toward measurable use. Flights, attractions, cars, taxis, insurance, and payments now create a low-double-digit share of Booking.com transactions in multi-vertical combinations. Q2 Connected Trip transactions grew low double digits, more than twice total Booking.com transactions. The missing disclosure is incremental retention and contribution margin. Without it, cross-sell is strategically promising but cannot yet receive standalone valuation credit.

The transformation program deepened. Organizational changes begun in late 2024 produced about $250M of 2025 savings and a $550M year-end run rate. The target is now $650M by end-2027, with most incremental savings in 2027. Procurement is a material source. Personnel growth was only 0.5% in Q2, but IT rose 20% as cloud and software spending increased. This is an efficiency-and-reinvestment program, not simply austerity.

AI became both product infrastructure and distribution risk. Priceline announced a more fully agentic Penny experience, Booking.com is testing AI discovery, and the group uses AI in engineering, customer service, partner messaging, and voice workflows. Management reports positive ROI and double-digit service cost reduction, but does not disclose reproducible contribution economics. Meanwhile Google AI Overview is already pressuring organic search, and Google AI Mode is expanding travel planning and future booking completion. The debate is no longer whether AI touches travel; it is who owns attribution and checkout.

B2B is becoming a coordinated platform. Booking, Agoda, and Priceline historically ran separate B2B units. Combining them should improve inventory and tools for banks, airlines, travel-management companies, and affiliates. The strategic pressure is Expedia’s long record of double-digit B2B growth. The economic guardrail is incrementality: B2B distribution can grow quickly while cannibalizing higher-margin B2C demand.

Capital return accelerated and leverage rose. First-half buybacks and dividends exceeded free cash flow by $1.67B; net debt issuance was $1.98B. This is not balance-sheet stress, given approximately $3.0B net debt and $9.5B trailing FCF, but it changes the allocation question from “how much surplus cash exists?” to “at what price should the company borrow to retire equity?”

Regulatory exposure broadened. DMA anti-parity and data-access rules are now operational. Spain’s parity/ranking liability was $472M; a Swiss commission-reduction order is suspended on appeal; Agoda faces an October French conduct deadline; and hotel/consumer collective claims continue. In July, FTC staff told Priceline it intended to recommend a complaint involving disclosures, fees, support, and billing. The April guest-data incident adds privacy and reputation risk. These are varied exposures, but several can change business conduct and economics rather than simply consume cash.

25-for-1 split. The April 2026 split did not change value, but improved accessibility and requires care when comparing historical prices, shares, and insider transactions. All per-share figures in this report are split-adjusted unless stated otherwise.

Verdict. Strategic change has mostly strengthened the fulfillment platform—merchant payments, cross-sell, AI productivity, and B2B coordination. The negative changes sit at the boundary: more expensive traffic, more assertive regulators, slower unit growth, and a capital-return pace above internally generated cash.

9. Risk Analysis

Tail-risk interaction matters more than any single item. AI could shift demand to Google or assistants, requiring more paid traffic at the same time DMA helps hotels offer better direct prices. That would reduce conversion and revenue capture while raising acquisition cost—the precise two-sided squeeze that could impair the moat and FCF. Conversely, if AI sends high-intent demand while Booking remains merchant and servicer, customer-service automation could lower cost and strengthen returns. Monitoring each risk independently misses this interaction.

The balance sheet reduces survival risk but does not eliminate valuation risk. A platform can remain profitable and solvent while its terminal multiple compresses. At today’s near-normal historical valuation, a modest reduction in long-run FCF growth or terminal margin can produce a meaningful drawdown even without an earnings collapse. Five-year history includes a 39.5% maximum drawdown and several company-specific 9–10% earnings-day moves.

Key falsification dashboard:

1. Total direct mix falls by two points or more while marketing/GB rises above 5% for multiple quarters.
2. Revenue/gross bookings declines more than roughly 50 basis points on a trailing basis without a timing/mix reconciliation.
3. Room-night growth remains below 5% after Middle East air capacity normalizes while Airbnb/Expedia sustain high-single-digit units.
4. Alternative-accommodation mix falls and Airbnb hotel nights become material.
5. Gross debt rises above the stated through-cycle framework while repurchases continue above FCF at a premium valuation.
6. A DMA or national remedy forces broad commission reductions rather than local process changes.

Verdict. No single risk currently breaks the franchise. The combined AI–Google–regulation risk attacks the load-bearing demand-acquisition layer and deserves a real valuation discount. Geopolitics and cycle remain the nearer-term earnings risks; capital allocation is the self-inflicted risk investors can observe most quickly.

10. Valuation Discussion — Embedded Expectations

This section contains no recommendation or price target. It asks what the current quote requires from the operating model.

Live bridge. At $212.87 and 751.3805M filed shares outstanding, equity value is $159.946B. Adding $20.180B debt and subtracting $17.214B cash gives enterprise value of $162.912B excluding $514M of finance leases. Using trailing results through June 2026:

These are trailing, live-price multiples. They differ from period-end aggregators that retain a stale $178.24 share price. GAAP P/E is affected by currency and equity marks; EV/FCF and EV/EBITDA are more useful, though FCF includes seasonal merchant-booking working capital.

Own history. AZI’s valuation index places the current P/E at the 57th percentile, P/S at the 66th, and the two-component composite at the 62nd percentile of available history. Negative book equity makes P/B unavailable. In June the composite was near the 39th percentile. The stock has therefore moved from a below-median skepticism price to a somewhat above-median price while near-term room-night expectations declined.

Embedded expectations. A 17.1x FCF multiple and 6.0% FCF yield can support a satisfactory return if FCF compounds high single digits, share count continues falling, and the terminal multiple remains in the mid-teens. The quote does not require the most optimistic Connected Trip or AI outcome. It does require the franchise to remain durable: a flat FCF stream that eventually receives a 12x multiple would not be protected by a year or two of buybacks.

A simple reverse-DCF intuition is useful. Starting from $9.54B trailing FCF, five years of 8% growth produces approximately $14.0B before considering share-count effects. A mid-teens terminal multiple on that cash stream supports the present enterprise value plus a moderate return along the way. Five years of 3% growth produces only about $11.1B; at a low-teens multiple, current value becomes difficult to defend. The market is thus underwriting something closer to the long-run high-single-digit algorithm than to permanent low-single-digit impairment.

Scenario structure without a target:

Peer context. Booking deserves a premium to Expedia because its margin is roughly ten points higher, its accommodation supply is broader, and cash conversion is stronger. Airbnb deserves some premium for brand/direct traffic and faster unit growth, but its hotel expansion also demonstrates that the competitive boundary is moving. Cross-company multiples are distorted by differing merchant/agency revenue recognition, stock compensation, and balance sheets; they should validate, not dictate, intrinsic value.

Buyback effect. At the first-half average price of roughly $173, $7.4–7.8B of repurchases retired more than 4% of the share base in several months. At $213, the same dollars retire about 19% fewer shares. If operating FCF grows 8% and net share count falls 4%, per-share FCF can grow roughly 12%; if FCF is flat, the buyback still creates about 4% per-share growth. But funding the latter with debt does not create enterprise value and only creates equity value if repurchases occur below intrinsic value.

Verdict. The current quote embeds a durable high-quality compounder, not a broken OTA. It is reasonable if high-single-digit operating growth and disciplined repurchases persist, generous if AI/regulation reduce the terminal economics, and attractive only if room-night reacceleration or Connected Trip produces upside beyond the disclosed algorithm.

11. Variant Perception

Consensus framing. Booking is widely regarded as the highest-quality conventional OTA: global scale, European hotel leadership, mid-thirties margins, and aggressive capital return. The counter-consensus risk is equally familiar: Google owns discovery, generative agents may own planning, regulators are weakening parity, and room-night growth is slowing. After the rebound, the market appears to have moved from “structurally impaired” toward “quality compounder with known risks.”

Strongest bull case. Q2 shows the moat working through a geopolitical shock. Domestic and U.S. demand grew high single digits, direct volume increased, app and Genius engagement rose, supply expanded, payments generated positive incremental economics, and margin widened. LLM volume is immaterial, while Booking is an early transaction partner for both OpenAI and Google. AI already reduces service cost. Once airline capacity normalizes, room nights can return toward high single digits; transformation savings, merchant payments, Connected Trip, and a shrinking share base can sustain low-to-mid-teens per-share cash growth. The current valuation then represents a fair entry to a rare capital-light compounder.

Strongest bear case. The June bull test is failing in real time. Room nights are growing 5%, Q3 guides 3–5%, alternative accommodation 4%, flights 4%, and cars are declining. Marketing grows faster than bookings while Google AI Overview reduces SEO. Airbnb and Expedia grew units faster, and regulators explicitly want suppliers to offer better direct terms. Booking is using debt to buy stock above free cash flow, which flatters EPS precisely as organic units slow. LLM traffic is tiny because the channel is early, not because it is harmless. At 17x FCF and above-median own-history valuation, the market now discounts too little terminal risk.

Variant view. The most useful middle position is that the business is neither breaking nor fully vindicated. The market was too pessimistic near $154–162, and the Q2 result correctly reversed some of that. It may now be too willing to treat the remaining slowdown as entirely geopolitical. The differentiated view is to assign no current AI revenue impairment, yet demand a lower multiple because the discovery layer is demonstrably changing and total direct mix is stable rather than rising.

The four assumptions that matter most are falsifiable:

1. Underlying unit growth: does room-night growth recover above 7% once Middle East airline capacity normalizes?
2. Acquisition economics: can total direct mix rise from the mid-fifties while marketing remains below roughly 4.7% of gross bookings?
3. Revenue capture: does the trailing revenue/gross-bookings proxy remain around the mid-14% range despite parity and commission remedies?
4. Per-share compounding quality: can net shares fall 4–5% annually while capital return returns to at or below free cash flow and gross leverage stays controlled?

What would falsify the bull. Two or three quarters of sub-5% room-night growth independent of a travel shock, accompanied by direct-mix decline, marketing/GB expansion, or a falling revenue-capture proxy. A broad European commission remedy or material direct-hotel advantage visible in conversion would also do it.

What would falsify the bear. Room nights reaccelerate above 7%, direct/app/Genius share rises, and AI-originated demand grows while Booking remains merchant and servicer at equal or better contribution economics. Continued EBITDA-margin expansion alongside those outcomes would show AI is enlarging—not taxing—the platform.

Verdict. The operating debate is balanced; the valuation debate is less so after a 38% recovery from May. Evidence favors franchise durability, but no longer offers a large discount for taking that view.

12. Fact vs. Interpretation Table

13. Open Questions

1. What share of room nights and contribution profit comes from the B2C direct, B2B, paid search, SEO, metasearch, and emerging AI channels? The total/B2C denominator gap makes disclosure increasingly important.
2. How do Connected Trip cohorts differ in repeat frequency, gross profit per traveler, service cost, and cancellation rate from accommodation-only cohorts?
3. What are merchant-payment incremental revenue, processing/chargeback cost, working-capital requirement, and loss rates by region?
4. How much of Q3’s 3–5% room-night guide is quantified Middle East air-capacity/fare effect versus underlying international demand?
5. Why did flight-ticket growth slow from 28.5% in Q1 to 3.7% in Q2, and what portion is market demand, air capacity, or deliberate unit-economics discipline?
6. Are Airbnb’s faster hotel nights or Expedia’s B2B growth taking transactions Booking otherwise would have won?
7. What direct-booking price differential has appeared since DMA anti-parity implementation, and how has it affected Booking conversion, supplier-funded discounts, and commissions?
8. What percentage of AI-originated bookings leave Booking as merchant and servicer of record? How do customer-acquisition and repeat economics compare with paid search?
9. What formal remedies could result from the Priceline FTC matter, Swiss commission order, French Agoda order, and Spanish appeal? Which could scale beyond one jurisdiction or brand?
10. Will annual repurchases normalize to free cash flow now that the stock has re-rated, or will gross leverage continue rising toward the through-cycle target?
11. How much of the extra $100M transformation target will drop to margin versus fund AI, B2B, loyalty, and Connected Trip?
12. What was the scope and customer impact of the April guest-data incident, and what remediation or regulatory cost should investors expect?

14. What Must Be True

For the operating thesis to work from today’s valuation:

• Room-night growth must recover from Q3’s 3–5% range toward at least 6–8% over the next four quarters without relying primarily on ADR.
• Revenue capture must remain around the mid-14% range after normalizing booking/travel timing and mix.
• Total direct mix must at least hold in the mid-fifties, while app and higher-tier Genius shares keep rising; a falling total mix would make the loyalty story economically weak.
• Marketing expense cannot persistently outgrow gross bookings. A temporary 4.7% ratio is manageable; a structural move above 5% would challenge incremental margins.
• Merchant-payment revenue must continue exceeding incremental processing, chargeback, fraud, and service cost.
• AI agents must send incremental high-intent demand or leave Booking in control of checkout and servicing. Becoming an undifferentiated inventory feed would not support current economics.
• Regulatory remedies must remain local or behavioral rather than force broad commission reductions across core Europe.
• Free cash flow must compound at least mid-to-high single digits, and net share reduction should be funded primarily from cash generation rather than serial leverage.

For the bear case to become the better explanation:

• Unit growth stays below 5% after geopolitical effects fade and trails peers on comparable corridors;
• direct mix declines while paid acquisition intensity rises;
• the trailing revenue/gross-bookings proxy falls materially;
• alternative-accommodation mix contracts despite supply expansion;
• or national/DMA remedies create visible supplier-direct price advantage and commission compression.

The cleanest twelve-month scoreboard is therefore: room nights, alternative-accommodation nights and mix, total direct mix, app mix, Genius L2/L3 share, marketing/gross bookings, revenue/gross bookings, merchant mix and loss expense, adjusted EBITDA margin, FCF, and net shares outstanding.

15. Source Appendix

Primary evidence is weighted toward SEC filings, company materials, regulator releases, and identity-verified earnings calls. Current price comes from AZI; factor exposures come from FactorsToday and are treated as model output rather than company fact. ROIC’s displayed price-sensitive multiples were rejected because they used a stale price; only its reconciled trailing financial denominators and verified transcript were used.

1. Booking Holdings, Q2 2026 Form 10-Q, filed 2026-08-04: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000037/bkng-20260630.htm
2. Booking Holdings, Q2 2026 earnings release, filed 2026-08-04: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000036/q2-26bkngearningsrelease.htm
3. Booking Holdings, FY2025 Form 10-K, filed 2026-02-19: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000009/bkng-20251231.htm
4. Booking Holdings, 2026 proxy statement, filed 2026-04-21: https://www.sec.gov/Archives/edgar/data/1075531/000130817926000310/bkng-20260420.htm
5. Booking Holdings Q2-2026 earnings call, 2026-08-04, ROIC transcript; issuer identity and reported figures reconciled to the SEC release.
6. European Commission, Booking.com DMA compliance obligations, 2024-11-14: https://digital-markets-act.ec.europa.eu/booking-must-comply-all-relevant-obligations-under-digital-markets-act-2024-11-14_en
7. European Commission, first DMA review Q&A, 2026: https://digital-markets-act.ec.europa.eu/about-dma/dma-review-qa_en
8. European Commission, 2025 DMA implementation report announcement, 2026-05-22: https://digital-markets-act.ec.europa.eu/commission-publishes-2025-report-digital-markets-act-implementation-2026-05-22_en
9. Google, agentic travel planning and booking partners, 2025-11-17: https://blog.google/products-and-platforms/products/search/agentic-plans-booking-travel-canvas-ai-mode/
10. Google, AI travel tools update, 2026-04-17: https://blog.google/products-and-platforms/products/search/summer-travel-tips-google-search-ai/
11. Google, AI Mode scale update, 2026-05-19: https://blog.google/products-and-platforms/products/search/ai-mode-us-insights/
12. Airbnb Q2 2026 Form 10-Q, filed 2026-08-06: https://www.sec.gov/Archives/edgar/data/1559720/000155972026000027/abnb-20260630.htm
13. Expedia Q2 2026 Form 10-Q, filed 2026-08-06: https://www.sec.gov/Archives/edgar/data/1324424/000132442426000053/expe-20260630.htm
14. AZI Trading, BKNG price/EMA history through 2026-08-11: https://azitrading.com/controls/download-data.php?t=BKNG
15. FactorsToday, BKNG loadings: https://www.factorstoday.com/api/stock-loadings/BKNG
16. FactorsToday, methodology: https://www.factorstoday.com/about

APPENDIX A — Standard Diligence Questionnaire

Booking Holdings Inc. (NASDAQ: BKNG) · August 12, 2026

General

What is the one-sentence thesis? Booking is the highest-quality scaled OTA, protected by supply breadth and transaction infrastructure rather than hard consumer lock-in; its cash-compounding engine remains intact, but the current price already assumes durable high-single-digit growth.

What changed since the prior report? Q2 beat guidance and preserved margin, yet room-night reacceleration remained below the prior test and Q3 guidance is 3–5%. The stock rose from roughly $162 to $212.87, eliminating most of the prior valuation discount. SEO pressure, an FTC staff recommendation against Priceline, a guest-data incident, and more debt-supported buybacks are new risks.

What is the most important variant fact? LLM referrals remain significantly below 1% of room nights and total direct mix is stable, so AI has not yet disintermediated current bookings. The variant judgment is that this does not justify ignoring terminal risk at an above-median valuation.

Cyclicality and Earnings Nature

How cyclical is demand? Leisure travel is discretionary and exposed to macro, war, health events, foreign exchange, airline capacity, and ticket prices. Booking’s asset-light model and variable performance marketing make cash flow less operationally leveraged than hotel or airline earnings, but bookings and cancellations move quickly.

What portion is recurring? There is little contractual recurring consumer revenue. Recurrence comes from repeated travel behavior, supplier dependence, app/Genius engagement, and marketplace scale. That makes the stream resilient in aggregate but transactional at the unit level.

What are the best leading indicators? Room nights, cancellation rates, ADR, airline capacity, total direct mix, app and Genius share, marketing/gross bookings, and gross-bookings growth. Revenue lags booking activity until check-in.

Is reported earnings quality good? Operating income and free cash flow are high quality; GAAP net income is noisy because equity marks, euro-debt remeasurement, tax items, and impairments flow below or through earnings. Q2 adjusted net income grew 8% versus 118% GAAP growth. Use operating income, adjusted EBITDA, and trailing FCF together.

Business Quality and Competitive Moat

What is the moat? Greenwald economies of scale plus partial traveler and supplier captivity. Booking’s 4.7M-property supply, reviews, conversion data, localization, payments, fraud control, service, and global marketing are costly to replicate at equivalent liquidity.

What is not the moat? Consumers face negligible monetary switching cost, suppliers multi-home, and basic travel search or itinerary generation is easy to reproduce. Property count alone is not defensible without availability, conversion, and fulfillment.

Where is the moat strongest and weakest? Strongest in European independent accommodations and cross-border fulfillment. Weakest in top-of-funnel discovery, metasearch, flights, U.S. alternative accommodations, and generic itinerary planning.

Does current evidence show erosion? Not in profitability: adjusted EBITDA margin widened to 36.0%, revenue capture is roughly stable, and direct volume grew. Amber signals are marketing growing faster than bookings, explicit SEO decline, and slower unit growth than Airbnb and Expedia.

Is there a network effect? A scale-and-liquidity flywheel exists: more supply improves choice and conversion; more demand attracts supply; greater volume funds better tools and service. Because both sides multi-home, it is not hard winner-take-all lock-in.

Financial Condition and Balance Sheet

What is trailing cash generation? Derived TTM CFO is $9.859B, capex $320M, and FCF $9.539B on $28.241B revenue. The 33.8% FCF margin should be viewed over a full year because deferred merchant bookings create seasonality.

How leveraged is the company? Cash was $17.214B and debt $20.180B, for $2.966B net debt, below 0.3x trailing EBITDA. Gross debt and traveler float deserve monitoring in a shock, but current coverage is strong.

Why is book equity negative? Cumulative treasury-stock purchases exceed retained earnings additions. The $10.783B deficit is an accounting consequence of capital return, not distress. ROE and price/book are invalid valuation tools.

What hidden liability matters? $10.121B of deferred merchant bookings represents traveler/supplier obligations. Regulatory liabilities include $472M for Spain. Cyber refunds, chargebacks, and partner failures are operational tail risks.

Capital Allocation and Management

What is the allocation record? Strong: large share-count reduction, limited large M&A, modest capex, and a growing dividend. KAYAK’s impairment is the cautionary acquisition outcome.

What changed in 2026? H1 repurchases of $7.758B plus $664M dividends equaled 125% of FCF. Debt issuance net of repayment was $1.975B. The balance sheet can support it, but value creation now depends more on repurchase price.

Are incentives aligned? Performance awards emphasize three-year revenue and Compensation EBITDA with a relative-TSR modifier and absolute-TSR governor. The measure expenses stock compensation and capex, a positive. It lacks explicit ROIC, FCF, and per-share-value metrics. Insider ownership is below 1%.

What do insiders signal? Recent activity is three 10b5-1 sales totaling $1.423M and no purchases. Direction is negative but low-signal because transactions are planned and modest relative to compensation and holdings.

Valuation and Market Data

What is the live bridge? At $212.87 and 751.3805M shares, market cap is $159.946B. Enterprise value is $162.912B excluding finance leases. Trailing multiples are 16.29x EV/EBITDA, 17.08x EV/FCF, and 22.19x GAAP earnings.

Where is valuation versus history? AZI’s P/E/P/S composite is at the 62nd percentile. Negative book value prevents a P/B input. This is moderately above median, not an extreme premium.

What does the price require? Approximately high-single-digit FCF growth, stable mid-teens revenue capture, modest margin leverage, and continued share shrink. A permanent low-single-digit growth outcome plus a low-teens FCF multiple would not support the current enterprise value.

What does momentum say? The stock is roughly 16% above both 50- and 200-day EMAs after a 38% rebound from May, but the twelve-month return is flat and factor Momentum loading remains negative. The setup is strong recovery, not unambiguous crowding.

Risks and Downside

What is the existential risk? AI or Google controls travel discovery and transaction routing, turning Booking into an interchangeable inventory feed while charging for access to travelers. That can combine with DMA-driven supplier-direct pricing to squeeze both conversion and take rate.

What is the nearer-term risk? Q3 room-night growth of 3–5%, persistent Middle East air-capacity effects, and any broader consumer slowdown. Earnings-day history shows high sensitivity to small guidance changes.

What regulatory outcome is most damaging? A broad commission reduction or ranking/loyalty restriction across core Europe, not a one-time fine. The suspended Swiss order is a local example of a direct monetization remedy.

What is the self-inflicted risk? Continuing debt-supported repurchases at a full valuation while organic units slow. This can increase EPS without increasing enterprise value.

Recent News and Events

What was the latest reported event? Q2 results on August 4, 2026. The stock rose 6.6% the next session versus a flat market, validating a company-specific positive surprise.

What did guidance say? Q3 room nights +3–5%; gross bookings, revenue, and adjusted EBITDA +4–6% reported. Full-year gross bookings, revenue, and adjusted EBITDA high single digit; adjusted EPS low-to-mid teens.

What new legal facts emerged? FTC staff intends to recommend a complaint involving Priceline disclosures, fees, support, and billing. Booking also disclosed an April guest-booking-data incident and ongoing Swiss, French, Spanish, Greek, Hungarian, and collective-claim matters.

Monitoring Checklist

Quarterly: room nights by region; alternative-accommodation growth/mix; flight and car units; cancellation rate; gross bookings and ADR; revenue/gross bookings; merchant mix; marketing/gross bookings; total and B2C direct mix; app/Genius mix; adjusted EBITDA margin; FCF; gross debt; shares outstanding; average buyback price; formal AI referral economics; and regulatory remedies.

APPENDIX B — Source Appendix

Booking Holdings Inc. (NASDAQ: BKNG) · August 12, 2026

Primary — Booking Holdings and SEC

1. Q2-2026 Form 10-Q, filed August 4, 2026 — financial statements, KPIs, balance sheet, legal matters, direct/app mix, property count, merchant economics, SEO, capital allocation, controls: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000037/bkng-20260630.htm
2. Q2-2026 Form 8-K, filed August 4, 2026: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000036/bkng-20260804.htm
3. Q2-2026 earnings release, Exhibit 99.1 — results, guidance, non-GAAP reconciliations, quarterly statistical history: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000036/q2-26bkngearningsrelease.htm
4. FY2025 Form 10-K — multi-year financials, risk factors, cash-flow bridge, KAYAK impairment, capital allocation: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000009/bkng-20251231.htm
5. 2026 proxy statement — board independence, ownership, incentives, compensation, governance: https://www.sec.gov/Archives/edgar/data/1075531/000130817926000310/bkng-20260420.htm
6. Q1-2026 earnings release, April 28, 2026 — prior guidance and Middle East baseline: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000024/q1-26bkngearningsrelease.htm
7. Q4-2025 earnings release, February 18, 2026 — original annual algorithm and outlook: https://www.sec.gov/Archives/edgar/data/1075531/000107553126000008/q4-25bkngearningsrelease.htm
8. Booking Holdings Q2-2026 earnings call, August 4, 2026, retrieved through ROIC. Identity verified against the operator greeting, named executives, and SEC-reconciled figures. Management-only claims are labeled as hypotheses in the report.

Primary — Insider Transactions

9. Vanessa Wittman Form 4, filed July 29, 2026: https://www.sec.gov/Archives/edgar/data/1075531/000122926526000020/form4-07292026_040701.xml
10. Robert Mylod Form 4, filed July 30, 2026: https://www.sec.gov/Archives/edgar/data/1075531/000121460726000012/form4-07302026_040701.xml
11. Robert Mylod Form 4, filed August 6, 2026: https://www.sec.gov/Archives/edgar/data/1075531/000121460726000014/form4-08062026_040802.xml

The independent 60-month census reconciled to 473 SEC filings, including 251 Form 4s. Raw ownership XML parsed for 244 of 251; seven older accessions were blocked by SEC rate limits. None falls within the trailing 24-month window, so current insider direction is complete. This is disclosed as a corpus limitation, not silently imputed.

Primary — Regulation and Competition

12. European Commission, Booking.com DMA obligations effective November 14, 2024: https://digital-markets-act.ec.europa.eu/booking-must-comply-all-relevant-obligations-under-digital-markets-act-2024-11-14_en
13. European Commission, first DMA review Q&A — anti-parity and business-user effects: https://digital-markets-act.ec.europa.eu/about-dma/dma-review-qa_en
14. European Commission, 2025 DMA implementation report announcement, May 22, 2026: https://digital-markets-act.ec.europa.eu/commission-publishes-2025-report-digital-markets-act-implementation-2026-05-22_en

National matters—the Priceline FTC staff recommendation, Swiss commission order, Spanish CNMC case, French Agoda order, Greek and Hungarian investigations, collective claims, and April data incident—are sourced to the Q2 10-Q above.

Primary — AI and Distribution

15. Google, “New ways to plan travel with AI in Search,” November 17, 2025 — travel partners and future booking completion: https://blog.google/products-and-platforms/products/search/agentic-plans-booking-travel-canvas-ai-mode/
16. Google, “7 ways to travel smarter this summer,” April 17, 2026 — AI Mode travel planning and hotel price tracking: https://blog.google/products-and-platforms/products/search/summer-travel-tips-google-search-ai/
17. Google, AI Mode scale update, May 19, 2026: https://blog.google/products-and-platforms/products/search/ai-mode-us-insights/
18. Booking Holdings media room — Priceline Penny and Booking Holdings Ads product releases: https://www.bookingholdings.com/media-room/

Primary — Peers

19. Airbnb Q2-2026 Form 10-Q, filed August 6, 2026: https://www.sec.gov/Archives/edgar/data/1559720/000155972026000027/abnb-20260630.htm
20. Expedia Q2-2026 Form 10-Q, filed August 6, 2026: https://www.sec.gov/Archives/edgar/data/1324424/000132442426000053/expe-20260630.htm

Market, Factor, and Event Data

22. AZI Trading, BKNG price and EMA history through August 11, 2026: https://azitrading.com/controls/download-data.php?t=BKNG
23. AZI peer histories used for event checks: https://azitrading.com/controls/download-data.php?t=EXPE ; https://azitrading.com/controls/download-data.php?t=ABNB ; https://azitrading.com/controls/download-data.php?t=TRIP ; https://azitrading.com/controls/download-data.php?t=TCOM ; https://azitrading.com/controls/download-data.php?t=SPY
24. FactorsToday, BKNG factor loadings: https://www.factorstoday.com/api/stock-loadings/BKNG
25. FactorsToday, stock-specific volatility: https://www.factorstoday.com/api/stock-specific-vol/BKNG
26. FactorsToday methodology: https://www.factorstoday.com/about

Analytical Frameworks

30. Bruce Greenwald and Judd Kahn, Competition Demystified — barriers to entry, captivity, economies of scale, and local market definition.
31. Marathon Asset Management / Edward Chancellor, Capital Returns — supply-side industry analysis, management response, and capital-cycle mean reversion.

Data Limitations

• ROIC’s transcript-list endpoint returned unrelated issuers and was rejected for enumeration. The selected latest call was accepted only after identity and financial reconciliation.
• ROIC’s displayed market cap and multiples used a stale $178.24 price. All live price-sensitive values were rebuilt from the $212.87 AZI close, the filed share count, and filed cash/debt.
• Peer unit definitions and geographic mixes differ. Relative unit growth is a directional warning, not a calculated global share series.
• LLM referral, Connected Trip, and AI productivity economics are management assertions where no auditable channel or cohort disclosure exists.
• Pre-split price levels were divided by 25; return series used adjusted prices to avoid split distortion.