Booking Holdings

IThe business

Booking Holdings sells nothing it owns. It lists places to stay, flights, rental cars, tours and restaurant tables for other people, and it takes a cut when a traveler books. It runs five brands: Booking.com, Priceline, Agoda, KAYAK and OpenTable.² Booking.com is the one that matters. At the end of 2025 it listed about 4.4 million properties, 500,000 of them hotels, motels and resorts and 3.9 million of them homes and apartments.² In 2025 travelers booked 1,235 million room nights through the group, 68 million flight tickets and 88 million rental car days.²

The cut is called revenue, and it was 14.5% of gross bookings in 2025.² Gross bookings are the total value of travel booked, net of cancellations: $186.1B in 2025, up 12.4%.² Revenue was $26.9B, operating income $8.8B and net income $5.4B.² Revenue is booked when the traveler checks in, not when they reserve, so a cancelled trip does not reverse revenue.² Marketing, mostly paid search with Google, is booked when the reservation is made.² That timing difference is one reason cash flow and earnings can differ in a quarter.

Two ways it gets paid

Under the older agency model the traveler pays the hotel on arrival and the hotel pays Booking a commission. Under the merchant model Booking takes the traveler’s money at booking, pays the hotel later and keeps its margin. The group moved quickly from the first to the second. Merchant bookings were 63% of the total in 2024, 70% in 2025 and 73% in the second quarter of 2026.²⁵ In 2025 merchant gross bookings grew 24.8% to $130.0B while agency fell 8.7% to $56.1B.²

The shift matters to an owner for three reasons. It gives Booking the payment, which is its own revenue line: payment rebates and fees. Management says the extra revenue from facilitating payments exceeded the extra cost in 2025.² It also brings the customer’s money onto Booking’s balance sheet for weeks, which I come back to under Financials, because it is a form of float and not profit. And it brings chargebacks, fraud and customer-service cost. Sales and other expenses rose $381M in 2025 on merchant transaction costs alone.²

Where the volume comes from

The group does not report bookings by region in the way a retailer would, so I use what it gives. Revenue is attributed by the brand’s home, not the traveler’s: $2.6B, or 9.6%, was US and $24.3B came from outside the US in 2025, of which $21.7B sits in a Netherlands entity C (calculated).² In the fourth quarter of 2025 Asia and the US each grew room nights at a low double-digit rate and Europe and the rest of the world at high single digits.⁹ The 10-K names Asia and the US as the areas where it is putting the most brand and marketing effort.²

Alternative accommodations, meaning homes and apartments, were about 36% of Booking.com room nights in 2025, up from 35%, and grew about 10%.²⁹ They compete directly with Airbnb and Vrbo, and the 10-K says they may carry lower margins.² Flights are the other growth line. Airline tickets rose 36.6% to 68 million in 2025, but only 3.7% in the second quarter of 2026, against 28.5% in the first.²³ Rental car days fell 6.5% in the second quarter.³

The last four quarters

Room nights grew 9.0% in the fourth quarter of 2025, 5.9% in the first quarter of 2026 and 5.3% in the second.³ Gross bookings grew 15.8%, 15.2% and 9.0% over the same quarters, but the first two were inflated by a weaker dollar: at constant currency they were 11%, 8% and 8%.³ Management links the slowdown to the Middle East conflict, which cut long-haul travel and raised airfares, while domestic and short-haul trips stayed healthy.⁵³ For the third quarter it guided room nights to grow 3% to 5% and gross bookings 4% to 6%.³ That would be the slowest room-night growth in any quarter in my file, which starts in the second quarter of 2024, and it is the first thing I will look at on the next report.

The balance of the year is not in the numbers I have. Booking reported the third quarter of 2025 on 28 October, but nothing for the third quarter of 2026 had been filed by the latest close.¹⁵ So everything below stops at the second quarter and the August guide.

IIThe moat

A travel site is worth something only if both sides keep coming back to it: hotels to list, travelers to book. Booking’s case is that it has both on a scale nobody else matches outside a handful of rivals. The evidence has to come from the filings, not from the brand.

Supply: who is on the platform

The most useful single fact is who the suppliers are. The top ten global hotel chains make up only a low double-digit share of Booking.com room nights, and almost 90% of room nights come from independent hotels, alternative accommodations and smaller chains.⁹ An independent hotel with no sales team and no loyalty program rents Booking’s reach, its payments, its 40-plus languages and its guest messaging.² A chain, by contrast, can push guests to its own app. Where Booking’s supply is fragmented, no single supplier can leave without losing little, and the supplier gets little leverage. The 10-K also warns that consolidation among suppliers can lead to lower commission rates, so the argument is not free of risk.²

Demand: how travelers arrive

The second fact is how much demand arrives without paying for it. The direct share of room nights was a mid-fifties percentage over the last four quarters, similar to a year earlier, and higher once affiliate partners are excluded.³² Travelers on the mobile app, where the vast majority book direct, were a high-fifties share of room nights over the twelve months to June 2026, up from a mid-fifties share a year earlier.⁵ Genius, the loyalty program, covers Level 2 and 3 members who accounted for a high-fifties share of Booking.com room nights in 2025, up from the mid-fifties in 2024, and they book direct at a meaningfully higher rate.⁹

These shares are given in bands, not numbers, which limits what I can say. But repeat direct demand is what makes a platform cheaper to run, and the marketing line shows it: 4.4% of gross bookings in 2025 against 5.5% in 2016.²⁶ It rose to 4.7% in the second quarter of 2026, from 4.6%.³ I read the one-tenth of a point as noise and the three-quarter trend as the thing to watch.

Scale in payments and data

Merchant payments are a $130B a year flow and a source of customer data and fraud learning that a small rival does not have.² That is a real, if modest, edge. The 10-K says the group runs on its own technology and increasingly on public cloud.² Customer-service cost per booking fell about 10% in 2025 as AI handled more contacts, according to the fourth-quarter call.⁹ These are cost advantages, and they compound only while volume grows.

What does not protect it

Three things do not. First, search. Performance marketing is mostly paid search on Google, which the 10-K names, and the same filing says Google links travel search to Maps and to its Gemini assistant.² Booking pays the toll and competes with the toll collector. Second, price. For many travelers price is the main reason they pick a site, and some rivals will lose money on a booking to win share.² Third, switching: a traveler can open Airbnb, Expedia, Agoda or a hotel’s own site in a few seconds, and the 10-K says barriers to entry are low.² AI makes that lower. Booking’s own risk factor says AI assistants may reduce visits to dedicated travel sites and that AI agents could turn into full-service booking platforms.²

So I call the franchise real and untested. It is real in supply and in repeat demand. It is untested because every earlier threat, from Google to Airbnb to Expedia, was a site a traveler had to choose. An assistant that books for them on a prompt may not give Booking a name to choose.

IIIThe competitive advantage

The advantage over online rivals shows up in the margin. The latest quarter of each of the three public peers I can read is below. They are not measured the same way, so I show them as reported.

Where Booking wins

It wins on cost to reach a customer. Booking spent 4.7% of gross bookings on marketing in the June quarter and made 36.0% adjusted EBITDA on revenue.³ It wins on breadth of supply, with 4.4 million properties.² And it wins on cash. Free cash flow was 49.6% of revenue in the June quarter and 33.8% for 2025.³⁴ Expedia’s was 29.6% of revenue in the same quarter C (calculated).⁷

Where it is losing ground

It is losing ground on growth. Expedia grew bookings 12% against Booking’s 9%, and Airbnb 16%.⁷⁸³ Part of that is mix: Expedia’s business-to-business arm grew bookings 21%.⁷ Some of it is the Middle East, which hits Booking’s long-haul and cross-border mix harder than a domestic-heavy competitor.⁵ I do not know how much. But Airbnb’s letter says it is doing better than its largest peers, and for one quarter the numbers back it.⁸

It is also losing ground in one business the company bought. In September 2025 KAYAK, the meta-search brand, was written down $457M, $180M of goodwill and $277M of trade names, because its forecast cash flows fell on expected higher customer acquisition costs.² That is the same pressure the whole company faces, in a smaller unit that has no direct traffic to lean on.

The honest summary

Booking has the best economics in its industry and, this year, not the best growth. The market is paying for the second. Which is the real signal depends on whether the Middle East effect fades, which I expect to show by the first half of 2027, and on whether bookings that start inside an AI assistant, a number the filings do not give, become a large share.

IVManagement

The people are Glenn Fogel, chief executive throughout the five years in the proxy’s pay table, and Ewout Steenbergen, chief financial officer since March 2024.¹⁰ The more useful question is what the pay plan rewards, how much of their own money is in, and what they did with capital over a full cycle.

How they are paid

Fogel’s 2025 pay in the summary table was $35.4M, down from $44.8M in 2024 and $46.7M in 2023.¹⁰ His base salary is $1.25M. His bonus target is 200% of that, and the 2025 plan paid 1.62 times target, or $4.04M, because revenue grew 7% against a 6% target and the company’s own “Compensation EBITDA” grew 13% against 6%.¹⁰ The rest is stock: 40% restricted shares that vest over time and 60% performance shares. The performance shares pay on revenue and Compensation EBITDA summed over three years, move up or down 25% on the stock’s return against a broad group of travel companies, and are capped at target if the stock’s absolute return is negative over the period.¹⁰

So what does that pay them to do? To grow revenue and a cash profit measure, and to keep the share price from falling. It does not pay them for return on capital or for earnings per share, and it does not set a limit on what they spend to grow. That is not unusual, and the share-price governor is a good feature. But look at the bar. The 2025 performance shares need revenue over 2025 to 2027 of $81.7B for target and $85.3B for the maximum.¹⁰ Revenue was $26.9B in 2025 and I expect about $29.0B in 2026 E (estimate), so 2027 revenue of $25.8B earns target and $29.4B earns the maximum, which is below what 2026 will report C (calculated).²³ The plan, as I read it, is nearly cleared with a year to go. The proxy says the committee set these goals after considering budget and consensus, and growth has since run ahead of them, so I don’t read it as bad faith. It does mean the 2027 payout depends little on 2027.

Their own money

Fogel owns 537,900 shares, of which 345,500 are in a grantor retained annuity trust of which he is trustee, worth about $80M at today’s price C (calculated).¹⁰ The ownership rule asks for six times salary, or $7.5M, so he is far above it. The general counsel, Peter Millones, owns 452,300 shares.¹⁰ All directors and executive officers together own 1,312,775 shares, under 0.2% of the company. Vanguard owns 9.4% and BlackRock 8.2%.¹⁰ I reviewed ownership, not sales, and did not go through the Form 4s, so I can’t say whether executives have been selling.

What they did with capital over a cycle

This is the part that matters most, because nearly all the cash went to repurchases. From 2016 to 2025 Booking spent $48.5B repurchasing about 505M shares after adjusting for the split, an average of about $96 a share C (calculated).^2131112146 At today’s price those shares are worth $75B, so the programme is ahead by about 54% C (calculated). The diluted share count fell 35%, from 1,252M to 816M C (calculated).²⁶ In the first half of 2026 it spent another $7.8B.³

The record has two halves. The money spent in 2018 to 2021, at about $75, has roughly doubled, and the money spent in 2022 and 2023, at about $96, is up 54%. The money spent in 2024, 2025 and the first half of 2026, about $21B, went in at averages of about $154, $198 and $176, and is now under water, by a quarter in the case of 2025. A buyer of its own stock at any price is not a disciplined one, and management has not tied its repurchases to a stated value. Since 2020 it has also shown it can stop: in 2021 it spent only $163M, borrowed $4.1B the year before, and the share count barely moved.¹¹ That was the right call when cash flow dropped to $85M, and it is the main reason I do not mark management down for the higher prices.

The money came partly from debt. Principal outstanding was $20.3B at 30 June 2026, up from $18.9B at year-end, after $3.0B of new notes in May, and $19.1B of it is in euros.⁵ Booking pays a dividend too, $0.42 a share a quarter after the split, or about 1.1% of today’s price C (calculated).³

Acquisitions

The acquisitions record is weaker. OpenTable lost $941M of goodwill in 2016, KAYAK and OpenTable another $1,062M in 2020, and KAYAK $457M in 2025: $2.5B of write-downs on businesses bought in the 2010s C (calculated).⁶¹¹² The 2021 deal for Getaroom cost about $1.2B net of cash.¹¹ In November 2021 it agreed to buy Etraveli, a flight booking supplier, and in September 2023 the European Commission prohibited it, at a $90M termination fee.² Booking now relies on a commercial arrangement with Etraveli for flights.¹³ I am not holding the failed deal against management, since the regulator moved the goalposts, but flights are where its stated growth plans are, and it does not own the plumbing.

Cost discipline

The Transformation Program began in the fourth quarter of 2024 to take out cost and fund reinvestment. Management first expected $400M to $450M of run-rate savings, raised that to $500M to $550M in 2025 and raised it again in August 2026 to about $650M by the end of 2027.²³ Headcount was about 24,300 at the end of 2025, flat on a year, while revenue grew 13%.² Management said on the February call it would reinvest about $700M above baseline in 2026 in AI, the Connected Trip, advertising, payments and the US and Asia.⁹ That is the right use of savings if the reinvestment works, and a reason the margin may not rise as much as the savings suggest.

My read. They run the business well and allocate capital less carefully. Pay is tied to growth, not to returns. The repurchases are large enough that the timing matters, and the best prices were paid during the years the stock was cheapest and the cash flow weakest. Skin in the game is real. I would give more credit if I saw a stated rule that links repurchases to value.

VFinancials

Ten years, calendar, in US dollars. I use the figures as filed and mark the ones I calculated. The first thing to see is that 2020 broke the series, and the second is that the margin has since passed its old peak.

The margin

Operating margin was 35.8% in 2017, 36.8% in 2018 and 35.5% in 2019, then 32.8% in 2025 after the KAYAK write-down and other items.²¹² Adjusted EBITDA margin, which the company defines to exclude one-off items but not stock pay, was 36.9% in 2025 against 35.0% in 2024, and 36.0% in the June quarter.⁴³ The gain is cost, not take rate. Revenue was 14.5% of gross bookings in 2025, below the 15.6% of 2019. Marketing was 4.4% against 5.2%, so Booking now keeps a bigger part of a smaller cut.²¹² Personnel fell from 14.1% of revenue in 2024 to 12.6% in 2025 with a $176M pension release inside it.²

The question for the next ten years is whether the lower marketing rate is permanent. I think part of it is: the direct share is up, the app share is up, and the loyalty program is bigger. Part of it is the cycle. Marketing ran above 5.5% of bookings in 2016 and 2017 when Booking was growing room nights 20% a year and spending to do it. It would not take a collapse in the platform to send it back. It would take a change in what Google, or an assistant, charges for the traveler’s click.

What an owner can take out

Net income is a poor guide to Booking’s cash, and the reason is partly accounting and partly the float. In 2025 net income was $5.4B, down 8%, while adjusted EBITDA rose 20%.⁴ The gap came from a $1.4B unrealized loss on euro debt, a $457M impairment and a $0.36B charge for convertible notes, none of which moved cash.² In the second quarter of 2026 net income rose 118% for the opposite reason: a $195M euro-debt gain and no convertible charge.³ So I ignore GAAP growth in both directions and use cash.

Reported cash flow overstates what is free. Over the twelve months to June 2026 Booking made $9.86B of operating cash flow and spent $0.32B on equipment and software, which the company counts as $9.54B of free cash flow.⁵²³ But stock pay was $0.60B, a real cost paid in shares that the repurchases then have to offset. That leaves $8.94B. Inside it is $0.84B from the working-capital float that I take out below.

The result: about $8.1B of owner cash, against $9.5B of free cash flow and $7.2B of net income. That is 15% below the free cash flow headline C (calculated). At today’s market value of $111.6B it is a 7.3% yield, or 8.0% if I leave the float in C (calculated).

The float

When a traveler pays Booking in advance and the hotel is paid weeks later, Booking holds the money in between. That balance, “deferred merchant bookings,” was $5.3B at year-end 2025 and $10.1B at 30 June 2026, because bookings made in spring are mostly stayed in summer.³ As merchant mix rose, the balance grew faster than bookings, and operating cash flow got a lift. The net inflow from receivables, prepaid items, deferred bookings and other items was $1.8B in 2022, $1.7B in 2023, $0.4B in 2024 and $0.5B in 2025 C (calculated).¹³² In 2020, when travel stopped, the deferred-bookings line alone took $2.3B out of cash flow, and total operating cash flow was $85M.¹¹

That is why I price the float carefully. As the merchant share nears three-quarters, the lift fades. In my base I let it contribute 2.5% of each year’s added bookings, about what 2024 and 2025 showed, and in the bear 1.5% E (estimate). If bookings fell, it would go the other way, fast.

Cash that is not Booking’s

The balance sheet needs the same care. At 30 June 2026 Booking had $17.2B of cash against $20.2B of debt, so the “net debt” was $3.0B.³ That is misleading. Most of the cash is the traveler’s and the hotel’s: $5.1B of payables, $3.8B of accruals and $10.1B of deferred bookings sit against $5.8B of receivables and other current assets, so about $13.2B of the cash is owed to someone else C (calculated).³ At year-end, a quieter date, the same sum gives about $9.8B owed and about $7.4B of cash that is Booking’s own, against $18.7B of debt, or about $11.3B of net debt that is really owed, about 1.4 times owner cash C (calculated).² I treat the debt as a cost that the owner earnings already bear, since interest is deducted before them. I do not count the cash as spare.

The debt itself is long-dated and mostly in euros: $19.1B of $20.3B of principal, with $2.0B due within twelve months and a $2B credit line unused.⁵ The euro debt is a partial hedge for euro earnings. It also produces the swings in net income, such as a $1.4B loss in 2025 and a $528M gain in the first half of 2026.²³ Stockholders’ equity is a deficit of $10.8B, because $62.1B of repurchased stock is deducted.³ I treat that as an accounting result of buying back shares, not a sign of distress. Interest and dividend income was $921M in 2025, about 11% of owner cash before tax, so lower rates matter: each point of yield on $17B is about $170M a year C (calculated).²

Tax

The effective tax rate was 20.9% in 2025, helped by Dutch rules that tax income from qualifying innovative activities at 9% instead of 25.8%.² The 10-K lists the loss of that benefit as a risk, and a new international minimum tax and US changes are being assessed.² At 21% my base is close to what the company pays. If the rate rose to 25%, owner cash would fall by about 5% C (calculated). I put 23% in the bear case.

The price over ten years

The stock closed at $148.13 on 30 October 2026. That is 36.3% below its record close of $232.64 on 7 July 2025, and 28% below the $207.02 of 5 August, the day after the second-quarter report.¹ I could not find a filing that explains the fall since August. The press ties it to the Expedia and Meta AI assistant news of 23 September, to an upheld EU ruling that blocked the Etraveli deal, and to a cut in one bank’s 2027 estimates.¹⁶ Those are press reports, not filings, and I treat them as colour. In February the stock fell 9.3% on 3 February, a day the press tied to a sector-wide scare over AI agents that I could not confirm in a filing, and 6.1% on 19 February, the day after the fourth-quarter report and the 2026 guide C (calculated).¹⁴ The pattern is that the share price reacts to AI news within a day and the reported numbers have not yet moved.

VIValuation

First, how this goes badly wrong

Before the numbers I asked what would most likely ruin this business or this investment, and worked back. Three things could. None needs a surprise.

The traveler is reached somewhere else, and Booking pays more to be there. Booking’s cost of finding a customer is mostly what it pays Google, plus what it pays affiliates and meta-search. The 10-K says AI answers may satisfy a traveler without sending them to Booking, and that terms, placement and data access on those platforms could raise its cost or cut its traffic.² KAYAK already shows the pattern: it was written down on higher expected acquisition cost.² If Booking’s marketing returned to the 5.5% of bookings it spent in 2016 and 2017, with no change in revenue per booking, margin would fall about seven points, because one point of bookings is about seven points of revenue C (calculated).⁶ That is my bear. A tougher version, in which the assistant owns the booking and Booking becomes a supplier of inventory at a lower take rate, is my stress case.

A regulator or a tax authority takes a slice. Spain’s competition authority fined Booking.com and restricted its parity and ranking terms, and Booking has accrued $485M and is appealing.² Switzerland’s price office ordered a cut in its average commission, now suspended on appeal.² Greece and Hungary opened new cases in 2025, hotels in Germany, the Netherlands and elsewhere are pursuing parity claims, and the EU has named it a gatekeeper under the Digital Markets Act.² The tax side is the Dutch 9% rate on innovation income, which the 10-K says had a significant beneficial impact on the tax rate in 2025.² Each is survivable alone. Together they lower the take rate or raise the tax rate for a decade.

Borrowed repurchases meet a shock. Booking has about $20.3B of debt and spent $7.8B on its own shares in the first half of 2026, at about $176.⁵ Its cash is mostly other people’s. In 2020 operating cash flow was $85M and the float moved against it.¹¹ A repeat would not break Booking, which has long-dated debt and an undrawn $2B line, but it would force a choice between the dividend, the repurchases and the debt, at a share price that would not be kind.⁵

All three are versions of one question: is Booking a toll road for travel or a toll booth that a new road can bypass? They set my bear case. I keep a fourth outside the model because I can’t size it: if assistants become the main place trips begin, the old customer relationship may be worth much less than the cash flows I project.

What the price implies

At $148.13 the market values Booking at $111.6B on 753M diluted shares, the 751.4M outstanding on 27 July plus 2M of dilution C (calculated).⁵¹ That is 15.5 times the trailing adjusted earnings of $9.58 a share and 16.5 times GAAP earnings of $9.00 C (calculated).⁴³ The owner yield is 7.3% without the float and 8.0% with it, against the 4.94% ten-year Treasury C (calculated).¹⁷ The nearest in the last ten years was the end of 2022, at 7.2%.

What does that pay for? With margin flat at 37.2% and revenue at 14.3% of bookings, the price needs gross bookings to grow about 1.4% a year for ten years, about a seventh of the 10% constant-currency growth of 2025 and below the 3% I use after year ten.² If the margin drifts down to 34%, it needs 2.6%. Or on my base growth, the price works only if the margin falls to roughly 23% by 2036. I take that to mean the market is pricing either almost no growth or a deep fall in margin, a much worse picture than the 7% bookings growth and 39% margin I use.

Reverse DCF: the inputs

Solve the model for today’s price instead of for value.

What I can’t judge

Two inputs decide this valuation, and I can’t know either. One is how much of the travel decision moves into an AI assistant, and who gets paid when it does. Management says it expects to meet whatever demand moves to general-purpose assistants, and it names OpenAI, Google, Microsoft and Amazon as companies it works with.⁹ Morgan Stanley is reported to call AI a new acquisition channel rather than a threat; the filings call it a risk.¹⁸² Both can be true for a while. I have no data that would decide it, and nor does anyone else yet. The other is the normal marketing rate. 4.4% of bookings is the lowest in my ten years, and I do not know if it is a floor or a cycle low. So here is the range and not one number. Read my $223 with the grid beside it.

My base value at other growth rates and margins

Value per share. Across: the 2036 adjusted EBITDA margin, reached in a straight adjustment from my base path. Down: points of gross bookings growth a year added to my base.

The discount rate

I use 10.0%. The ten-year Treasury yields 4.94%, and Damodaran’s implied equity risk premium was 4.20% on 1 October.¹⁷¹⁹ My ten-year beta on month-end returns against the S&P 500 is 1.17, which gives 9.9%. The five-year beta of 1.16 gives 9.8% and the three-year beta of 0.77 gives 8.2% C (calculated).¹²⁰ I round the ten-year figure to 10.0%. It is a little under the 10.5% on my Broadcom page and above the 9.0% on my Visa page, because travel demand follows the economy and a card network’s volume follows it less.

My base case at other rates

Value per share; discount rate down the side, growth after 2036 across the top.

Three cases

The model is simple on purpose. I build gross bookings from the 2026 estimate of $202.6B, which is the first half as reported, plus the third-quarter guide and my own assumption for the fourth. I apply a revenue share of bookings and an adjusted EBITDA margin, which already deducts stock pay. Then I take off cash tax on profit after depreciation and net interest, capital spending at 1.1% of revenue, net interest of $0.3B in the base, and I add the float at 2.5% of each year’s added bookings. What is left is owner cash after interest, so I discount it at the cost of equity and do not adjust for debt or cash. The last two months of 2026 are added as a stub. Per-share values divide by 753M shares.

My base 2027 revenue of $31.4B is 8% above my 2026 estimate and my adjusted EBITDA of $11.7B is 9% above. That is in line with the 8% revenue ambition management states and a bit under the 15% earnings-per-share ambition once the share count shrinks 3% a year.⁹ So I don’t differ from management in the near years. I differ from the market in the later ones.

Bear. Gross bookings grow 4.5% falling to 3%. Revenue per booking slides from 14.4% to 13.2% as the assistant or the platform takes a larger share, and marketing rises back to levels like 2016 and 2017, so adjusted EBITDA margin falls from 36.9% to 32% by 2030 and 31.5% after. Tax rises to 23%. Revenue reaches $38B in 2036. This is a good business with a bad decade, and it is worth $135.

Base. Bookings grow 7.5% fading to 4%. Revenue per booking drifts from 14.4% to 14.2%, flights and payments roughly balancing. Margin adds about two points to 39%, as the Transformation Program saves $650M and mix pays for part of the reinvestment. Revenue reaches $52B and owner cash $15.6B.

Bull. Bookings grow 9.5% fading to 5%, flights and attractions work, and the direct share rises. Margin reaches 42%. Revenue reaches $61B.

Where would $268B of 2030 bookings come from? From $203B, that is 7.2% a year for four years, slightly below the 8% a year the company says it aims for at constant currency. I see it as room nights at 4% to 5%, price at about 2% and flights and attractions adding about 1 point E (estimate). If room nights stay at 3%, as the third-quarter guide implies, the base needs more from price and flights, and it moves toward the −2 row of the grid.

What it is worth, against the price

The price sits 34% below my base, and that is the plain result. But why might the market be right? Because it can see what I can. Room nights slowed from 9% to 5% in two quarters and are guided to 3% to 5%. Peers are growing faster. KAYAK was written down on higher acquisition cost. Press reports say a new Meta assistant has a deal with Expedia and that an EU court upheld the block on Booking’s one big deal.¹⁶ And the reported figures lag the thing they are meant to measure. Stocks don’t fall 36% on nothing, and the market may be saying the toll road is about to get a bypass. If so, my bookings growth is too high and my margin is too high, and the grid shows what that costs. The $135 bear is a fair picture of that world, and the $90 stress case is a harder one. The price sits just above the bear and far above the stress case.

Other uses of the same dollar

The ten-year Treasury pays 4.94%. Against that, my base here returns 13.5% a year on its cash flows and my bear 9.3% E (estimate).¹⁷ That gap is large, and I treat that as a warning to check my work again, not as good news. The closest comparison on this site is Uber, another marketplace with a travel overlap, an AI-bypass debate and a $216B annual booking figure. There the price is 24% below my base and needs 6.2% growth.²¹ Booking’s gap to base is wider. Alphabet, which owns the search box Booking pays and the AI assistant it competes with, trades 43% above my base there.²² Visa, a steadier toll, is 18% above.²³ Those pages use their own prices, rates and methods, so a gap of a few points is noise. Booking’s gap is not. It is what the market charges for the assistant question, and it is wider than the one it charges Uber.

What past lows looked like

The stock has fallen more than 30% from a high three times in ten years: 48% from March 2018 to March 2020, 40% from February to October 2022, and 36% now C (calculated).¹ At the end of 2022, at $81, the owner yield was 7.2%, and the stock closed at $142 a year later and $199 a year after that C (calculated). At the 2020 low of $46.09 the year’s cash flow ended at $85M.¹¹ I don’t take comfort in those. The 2022 case was a cycle fear. This one is a structure fear. It resembles 2022 in yield and nothing else.

The margin of safety

• Price$148.13Close, Q4 2026 [1]
• Base value$223My estimate E (estimate)
• Margin of safety below$135set from my bear case C (calculated)

I set it at $135, the value of my bear case. At that price even my bad case earns the 10% I ask for, and the base case would return 14.5% a year E (estimate). I chose the bear, not a discount to base, because the danger here is a change in how travel is booked, and a discount to base would not protect against that. At $148.13 the price is 10% above $135, so there is a margin against my base and none against my bear. It does not cover the stress case either: bookings growth of 1% to 2%, revenue per booking falling to 12%, margin to 27%, which gives $90. The stock last closed at or below $135 on 7 August 2024, and the close of $147.94 on 28 October 2026 was its lowest since August 2024.¹

Two different questions

How good is the business, and how good is the price? They are separate. As a business, Booking is excellent on most tests I have: a 36.9% adjusted EBITDA margin, $8.5B of owner cash, a share count down 35% in nine years and cash that mostly matches earnings once the float is counted. Its flaw is a question and not a fault: it has to keep the traveler, and a new kind of intermediary is arriving. As for the price, it was poor at $232.64, when the base gave 9.7% a year and the bear 6.9%. At $148.13 it looks better on my numbers, and the market’s doubt is the reason. So this looks like a very good business at a price that depends on the answer to one question, which is a different thing from a good business at a bad price. It would be a mistake to treat the price as proof the answer is known.

Where this could be in 5 and 10 years

Value in a given year is the page’s own model at that date: the present value, at 10%, of the owner cash that comes after it, per today’s 753M shares and in that year’s dollars. Owner cash is after interest, tax, stock pay and capital spending. Year 5 is October 2031 and year 10 is October 2036.

The base value goes from $223 to $271 to $317, about 4% a year and then 3%, well under the 10% I discount at, because most of the return comes out as cash: $164 a share over ten years, more than today’s whole price. The 16.3% is higher than the 13.5% above because that one values the 2036 shares at the return itself, and this one at 10%.

For the base to be right at five years, bookings need to grow about 7% a year to $285B in 2031, revenue per booking stay near 14.2% and the margin reach 38.5%. At ten years it needs $364B of bookings and a 39% margin, and the 2036 value is all terminal value, so it rests on 3% growth after that. The bear differs on take rate and marketing: revenue per booking slides to 13.2%, margin to 31.5%, and growth fades to 3%. The bull has bookings growing 9.5% and fading to 5%, with a 42% margin. Capital needed is small, 1.1% of revenue, so it moves little. The discount rate moves more: at 12% the base today is $173.

The most likely way the base is wrong is growth in the next two years. Room nights are guided to 3% to 5% and the base needs bookings up 7.5%. Two points less every year puts today’s base at $191 and the 2031 and 2036 values at $226 and $261.

These are estimates of value, not forecasts of the share price. The price can sit above or below value for years.

VIIThe case for the price

What most people miss

Three things cut in Booking’s favour and get little attention. Marketing fell from 5.5% of gross bookings in 2016 to 4.4% in 2025, through the years Google added AI to search.⁶² Second, supply is not concentrated: the ten biggest hotel chains are a low double-digit share of Booking.com’s room nights, and almost 90% come from independents, homes and small chains.⁹ Third, $14.5B of repurchase authorization remained at 30 June, 13% of the company at today’s price C (calculated).³

The bull case

In plain words: Booking is the cheapest way for millions of independent owners to fill a room, it is the best funded business in travel, and AI will send it more customers, not fewer. On the February call management said it intends to meet whatever demand moves from search to general-purpose assistants, aims to grow bookings and revenue at least 8% and adjusted earnings per share at least 15% a year, and did better than that at constant currency in 2025.⁹ The sell-side version comes from Morgan Stanley, which press reports say started coverage with a $230 price target in September, calling AI a new acquisition channel and not a threat. I have seen only the secondary report.¹⁸

One large investor has been adding. Dodge & Cox reported about 24.9M shares after adjusting for the split at 31 March and 25,536,950, worth $4.55B, at 30 June, about 3% of the company.²⁴²⁵ That says nothing about why, and the data are four months old. Of 19 other well-known managers I checked, including Pershing Square, Fundsmith, Akre, Coatue, Tiger Global and ARK, 18 showed no position in the latest filing and Baillie Gifford held 37,987 shares, about $6.8M.²⁶²⁷

Against reality

Much of the case is in the numbers already. My base has bookings growing 7.5% in 2027 and 2028, revenue per booking near 14.3% and margin reaching 38.3% by 2030, which gives owner cash of $9.1B in 2027 and $11.4B in 2030 E (estimate). So if Booking keeps its direct share, holds marketing near 4.8% of bookings and meets its 8% ambition, my $223 or better is possible. The price needs far less: about 1.4% bookings growth at today’s margin.

The weak spot is current growth. Room nights slowed from 9.0% to 5.3% in two quarters and are guided to 3% to 5%. Expedia grew bookings 12% and Airbnb 16% in the June quarter against Booking’s 9%.³⁷⁸ The Middle East explains some of it. The case also rests on a hope with no number in the filings: that assistants send travelers to Booking at no more cost than Google does.²

Multiply it out. My base needs five things: bookings growth of 7% or more through 2028, 55%; revenue per booking near 14.2%, 70%; marketing under about 4.8% of bookings, 65%; no regulatory or tax hit above about three points of take rate, 80%; float and cash conversion holding, 85%. These are my guesses, and they give 17% E (estimate), perhaps 25% with overlap. The price needs growth of about 1.5%, revenue per booking above 13.5% and a margin above 34%: 85%, 80% and 70%, or about 48% E (estimate). So the price can ask for less than my base without my base being likely.

What a 10x would take. Ten times $148.13 is $1,481, about $1.12 trillion of market value on today’s shares. If Booking retires 3% of its shares a year, the count falls to about 556M by 2036, and the same price is $0.82 trillion. At 20 times earnings, above today’s 15.5, that needs about $41B of net income. At a 25% net margin, about what it made over the last twelve months, that is $164B of revenue, or 5.7 times my 2026 estimate, and gross bookings of roughly $1.1 trillion. It means 19% revenue growth a year for ten years, against 11% from 2016 to 2025 including the pandemic, plus a higher multiple.²⁶ Without the share retirement the revenue needed is $223B. As far as I know only a handful of companies have risen tenfold from a value above $100B in ten years, each on a surge in demand for something new, and I have not audited that list. For Booking, a tenfold rise from here is close to impossible, and I would not build a case on it.

What they have to hit

Six numbers a reader can check. The thresholds are my base case, and they agree with the signposts below.

Where I come down

The marketing line moved me: it fell while Google added AI to search. The supplier mix moved me too. Management’s view that AI will help didn’t, because it has no number. A year with bookings growth above 7% and marketing at or under 4.8% would raise my values most. This page shares one idea with Adobe, Intuit and GoDaddy, that AI will bypass the product people use now, so it is not an independent view; if I am wrong on that idea I am wrong on all four.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest. Note that Booking, Adobe, Intuit and GoDaddy came to me through the same screen for the same reason, a fall tied to fear of AI. They are one bet on one idea, not four.

• A growth gap with peers. Expedia and Airbnb grew bookings 12% and 16% in the June quarter against Booking’s 9%.⁷⁸³ If that persists for a year, Booking is losing share and my base is too high. Part may be the Middle East.
• Flights and attractions. They are lower margin, and flight tickets slowed to 3.7% growth in the June quarter from 28.5%.³ The 10-K says non-accommodation growth can lower the overall margin.²
• Alternative accommodations. About 36% of Booking.com room nights compete with Airbnb and Vrbo and may carry lower margins.²
• Regulation and parity. Spain, Switzerland, France, Greece and Hungary have cases open, hotel class actions are being prepared in several countries, and the EU names Booking a gatekeeper.² The Spanish fine of $485M is accrued but under appeal.²
• Tax. The Dutch 9% innovation rate and the 2018 US transition liability are both live. A 25% rate instead of 21% would take about 5% off owner cash C (calculated).²
• Currency and debt. Earnings are mostly in euros, pounds and other currencies and are reported in dollars; $19.1B of the debt is in euros. A weak euro lowers reported earnings and the debt together, but the earnings first.⁵²
• Geopolitics. Management cites the Middle East conflict for slower long-haul travel.⁵ I don’t forecast it, and a longer conflict lowers my 2027 numbers.
• Float reversal. If bookings fall, the deferred balance unwinds. It took $2.3B out of operating cash flow in 2020.¹¹
• Reporting changes. The company reports mix shares in bands such as “mid-fifties,” so a slide of a couple of points could go unseen for a while.³