Netflix

IThe business

Netflix sells a monthly subscription to a library of series, films, live events and a few games. It pays for or makes that content up front, and it collects from members in more than 190 countries.¹³ In the US the plans are $8.99 with ads, $19.99 standard and $26.99 premium, after the rise of March 2026.¹⁴ Worldwide, plans run from the equivalent of $1 to $37 a month.³ Advertising is the new line, about $2.7B this year, or 5% of revenue C (calculated).²

That is the whole model. The costs are mostly fixed once a title is made, and each new member pays almost entirely into profit. So the business rewards whoever has the most members. In 2016 revenue was $8.8B, operating income $0.4B and cash flow deeply negative. In 2025 revenue was $45.2B, operating income $13.3B and free cash flow $9.5B.¹⁵¹⁶ For 2026 the company expects $51.1B, up 13%, and a 31.5% margin.²

Where the revenue comes from

The US and Canada brought in $5.6B in the third quarter, or 44% of revenue. Europe, the Middle East and Africa brought in $4.1B, Latin America $1.6B and Asia-Pacific $1.5B.² Growth at constant currency was 11%, 9%, 12% and 15%. So the richest region is still growing faster than Europe, and that came with a US price rise in the mix.² About 57% of revenue is in other currencies, and that makes dollar growth noisy: Latin America grew 21% in dollars in the second quarter and 16% without currency moves.³⁴

Netflix stopped reporting memberships each quarter in 2025. It now reports them only at milestones.¹⁹ The last full count was 301.6M at the end of 2024, and the last milestone was over 325M in the fourth quarter of 2025.²⁰⁵ It says the audience is approaching one billion people.² Management’s own measure of the room left is that it has under 45% of broadband homes and about 5% of TV viewing worldwide.¹³ Those are company estimates, and the second depends on what counts as TV.

Peters described the growth levers plainly on the October call. In the US, the UK and Australia, growth comes mainly from revenue per member and ads. In India, Indonesia and Colombia it comes from new members. Markets such as Mexico and Japan get some of each.⁹ That matters because the richest markets are the most fully sold, so they rely on price.

The hours problem

This is the central fact about Netflix in 2026. Members watched over 97B hours in the first half, up 2%. In the second half of 2025 it was 96B, also up 2%, and in the first half of 2025 over 95B, up 1%.⁴⁵²¹ Revenue grew 15% in the same first half C (calculated).⁴¹³ So Netflix now earns about 25.6 cents for each hour watched, up from 22.8 cents a year earlier C (calculated). That is a 12% rise in the price of an hour.

Per member it looks weaker still. In 2024 the average membership watched about 1.86 hours a day. In 2025, on my rough average of 313M memberships, it was about 1.67 C (calculated).²⁰⁵ That fall of about 10% overstates the problem, because the crackdown on shared passwords added many thin accounts, and “over 325M” is a floor. But the direction is plain.

Management’s answer is that not all hours are equal. Live is the clearest case: it will be just over 5% of content spending in 2026 and about 1% of hours, yet it accounts for six of the top ten sign-up days of the last five years.⁴ Non-English titles are over a third of viewing.² Peters says churn, the share of members who leave, improved in the third quarter on a year earlier, both globally and in the US, and came in better than forecast.⁹ Netflix publishes no churn number, so I have to take that on trust. I find the argument fair for a few years. Over ten, a service people watch less each year has a price problem coming.

Netflix also said in July that its What We Watched report moves from twice a year to once, starting in the first quarter of 2027.⁴ It gave a sensible-sounding reason. Still, a company that stopped reporting members in 2025 is now reporting hours less often, in the year hours became the debate. I would read that as a reason to watch more closely.

Advertising

Netflix long refused to carry ads, and 2025 was only its third year selling them.²²⁵ Ad revenue passed $1.5B in 2025 and will be about $2.7B in 2026, up over 80%.⁵⁹ In October it cut that from about $3B, because some ad products rolled out slower than planned. Peters said the company is “a bit behind our aggressive target for the year.”⁹ The ads plan was over 60% of sign-ups in ad countries in the first quarter, and 15 more countries start in 2027, the first group in March.¹³² Peters also said that the gap between ad-plan and standard revenue per member is narrowing but still there.⁹ That gap is the opportunity. Each member on the $8.99 plan brings in less than one on $19.99 until ad sales fill the difference.

What changed in October

October was a rough month. About 1,000 jobs went, or about 6% of staff, which an analyst on the call called the largest cut since 2022, and it costs about 120 basis points of the fourth-quarter margin.²³²⁹ Within days Bela Bajaria left as chief content officer after a decade, and Peter Rice, formerly of Fox and Disney, was named to replace her.²³⁹ Sarandos was direct about why: “We do need better execution in the U.S.” He added that the US is “just not growing as fast as it could or as fast as we’d like.”⁹ He also said there is no need for a fundamental change of strategy.

I read that as a real admission. The company’s two big answers to the hours question in October were a new content chief and better shows. Both can work. India is its proof: a strong slate lifted hours there 24% and revenue about 18% at constant currency in the third quarter.² But slates take years, and Sarandos himself said that other regions depend on US titles travelling.⁹

IIThe moat

The source of the protection is simple. A film costs the same whether 10M or 300M homes can watch it. Netflix’s amortized content cost was 36% of revenue in 2025. Its cash content spending barely changed in four years: $17.5B in 2021 and $17.7B in 2025, while revenue rose 52% C (calculated).¹⁷¹⁶ Nearly all of the margin rise in that time is this one effect.

Evidence it is widening

Rivals have stopped trying to beat Netflix alone. Warner Bros. Discovery was sold for about $78B in October to Skydance, formerly Paramount Skydance, so HBO Max and Paramount+ now sit under one owner.²⁴ Disney licensed the Ice Age films and other titles to Netflix this year, and Globo, a long holdout in Brazil, signed its largest license ever with Netflix.²⁹ France’s TF1 put its channels inside Netflix, and early data suggest the viewing is incremental.² A service that broadcasters ask to join is a kind of distributor, and distribution is the most durable thing in media.

Margins tell the same story. Netflix made 34.6% in the third quarter, which was helped by timing. Disney’s streaming services made 13% and Warner’s streaming unit 16.6% on adjusted EBITDA, a more generous measure, in their latest reported quarters C (calculated).²⁸²⁵

Evidence it is narrowing

The competitor that matters isn’t a streaming service. In August YouTube had 15.8% of US TV time, a record, and Netflix 7.7%. Streaming was 49.6% in total.⁶ In December 2025 Netflix reached a high of 9.0%, so one month proves little, and September data will come with a change in Nielsen’s method.⁵⁶ But YouTube’s share is about twice Netflix’s, on the big screen, where Netflix used to feel safest. Netflix’s own letter in January noted YouTube leaning into TV, including rights to the Oscars from 2029, and Instagram’s TV app.⁵

That competition has an unusual shape. YouTube doesn’t pay for its content up front. It shares ad revenue with creators, so the cost scales with revenue. Netflix’s advantage of spreading a fixed bill over many homes doesn’t apply against a rival with no fixed bill. Netflix has responded by bringing some of those creators onto its service, with Ms. Rachel and others, and by adding video podcasts and vertical video on phones.⁴¹³ Those moves make sense, and they show which way the pressure runs.

AI video is a second question, and I come back to it. Netflix says about 300 titles have used generative tools in 2026, mostly in post-production, at lower cost.⁴ If cheap tools help Netflix they also help everyone else, and scarcity of good video is what a $20B budget buys.

IIIThe competitive advantage

The advantage over paid rivals is plain from a side-by-side. No two of these lines are measured the same way, so I show them as reported and don’t net them.

Where Netflix wins

It wins on cost per member. Disney spent $2.6B on programming in its June quarter on $5.5B of revenue. Netflix amortized $4.2B in the third quarter on $12.9B R (reported).⁸² So Netflix puts 33 cents of each revenue dollar into content and Disney 47 C (calculated), and Netflix’s absolute budget is still far larger. A rival would have to lose money for years to match it, and most have already tried. Peters said members in some cases pay more than double per hour on a competing service.⁹ That is a company claim, but it fits the price cuts rivals can’t make.

It also wins on knowing what to make for the world. A title from Denmark or Korea can travel because Netflix has the audience, and the third-quarter letter lists record titles for Japan, Denmark, Mexico and the Netherlands.² That took a decade to build.

Where it loses

It loses on advertising, by a wide margin. YouTube’s ads brought in $11.2B in the third quarter alone, up 9%.¹⁰ Netflix’s whole year will be a quarter of that. Disney’s streaming services sold $851M of ads in their June quarter, a run rate above Netflix’s 2026 total C (calculated).⁸ Yet Pershing Square, in a filing I cover below, says Netflix has nearly double the combined subscribers of Disney+ and HBO Max.²² Ad markets reward reach, measurement and self-service buying, which is where Netflix is behind by its own account.⁹

It loses on price of entry. YouTube is free. A viewer deciding what to do on a Tuesday night compares $19.99 with nothing. For now the 2026 price rises went through, with churn lower.⁹ I don’t know how many more will.

And it is weaker in sports, where rights are expensive and rivals have deep pockets. Netflix has picked events, not seasons: NFL games, boxing, the World Baseball Classic in Japan and the 2027 Women’s World Cup.²¹³⁹ That is a sensible way to get sign-ups at low risk. It leaves a whole category of weekly viewing to others.

IVManagement

Netflix has two chief executives. Ted Sarandos, 61, joined in 2000 to run content and has shared the top job since 2020. Greg Peters, 55, joined in 2008 and became co-CEO in January 2023, when Reed Hastings, the co-founder, became executive chairman.¹¹ Hastings gave up his executive role in April 2025 and left the board at the June 2026 meeting.¹¹²⁶ For the first time since 1997 he has no seat at the company.

The board has changed quickly too. Jay Hoag, lead independent director since 2012 and a founding partner of TCV, a venture firm, became chairman in June.²⁷ He is a mixed choice. He failed to win a majority in 2025, which the proxy puts down to missing meetings in 2024, and the board kept him.¹¹ In June he was re-elected with about 93% of votes cast C (calculated).²⁷ Anne Sweeney resigned in July, with no disagreement stated.²⁸

How they are paid, and what that pays them to do

Each co-CEO gets $3.0M of salary, a target bonus of $6.0M, and $31.0M of target stock, half in time-vested units and half tied to return against the S&P 500. Reported 2025 pay was $53.9M for Sarandos and $53.2M for Peters, 255 and 252 times the median employee’s $211,201.¹¹

The bonus is the interesting part. It is 65% constant-currency operating margin and 35% constant-currency revenue, measured over one year. For 2025 it paid 117.57% of target.¹¹ So the largest yearly dial on their cash pay is margin. Nothing in it measures hours, churn or a return on the content budget. A team paid this way will reach for cost control first, and that isn’t a bad thing for owners while costs have room to fall. But margin can rise for years while the product weakens, and the first thing you would see is what we are seeing: strong margins, thin hours. I do not say that is what happened. I say the pay plan wouldn’t have flagged it.

The stock award is better aligned. It pays between nothing and 200% of target by rank against the S&P 500 over three years, with target at the 55th percentile.¹¹ That is a real test, and after a fall of over 50% from the high it will bite. The 2024 tranche paid 200% at the 92nd percentile because the two years to 2025 were excellent.¹¹ Half of the stock award still vests with time alone, and comes in whatever the share price does.

How much of their own money is in

Less than the headline suggests. In April the proxy gave Sarandos 5.72M shares and Peters 2.75M, but 5.38M and 2.57M of those were options.¹¹ The Forms 4 show direct holdings of 178,954 shares for Sarandos and 119,412 for Peters, about $11.1M and $7.4M at the close C (calculated).²⁹³⁰ The options are the real stake. At $62.08, using the strikes in the proxy, they are worth about $130M for Sarandos and $62M for Peters above their exercise prices, and only 44,810 and 21,110 are under water C (calculated).¹¹ Most came as monthly pay from 2018 to January 2024, at an average strike near $38. That is a large stake compared with a year’s pay, and it lines them up with the share price. Its weakness is that an option holder loses little in a further fall compared with an owner who paid today’s price.

The guideline is six times salary, or $18M, within five years of 2024, and the summary I read doesn’t say whether options count.¹¹ I found no purchase by an insider in the Forms 4 from July 2025. Insiders sold about $565M in that time, and $451M of that was Hastings, whose sales ran to 1 June C (calculated).³¹ Sarandos sold about $12M from November 2025 to August 2026, and Peters about $20M since July 2025 C (calculated).²⁹³⁰ In April Hastings held 21.2M shares through a family trust and 16.6M options.¹¹ A founder selling after almost thirty years isn’t a verdict. But nobody at the top has put new money in at these prices, and the company’s own buyback is doing that for them.

What they did with capital over a full cycle

The full record starts with burning cash. From 2016 to 2019 Netflix had owner cash flow of −$10.6B and funded it with debt, to build a content library before rivals woke up.¹⁵¹⁸ That was the right bet: paid memberships went from 89M to 204M by 2020. From 2023 to 2025 it made $22.3B and paid back shareholders with $21.4B of buybacks C (calculated).¹⁶ Over the decade net income was $41.3B and owner cash $13.7B. Netflix pays no dividend.¹⁶

On that record they bought well in 2023, poorly in 2025 and so far poorly in 2026. In the third quarter alone they spent $4.8B at about $75 a share, the largest quarter ever, and the shares have since fallen further.² That is the honest test of a buyback: they bought most when the price was falling, which is what one wants if value holds, and they raised the authorization by $25B in April.³² At $62.08 the remaining $22.3B would retire about 359M shares, 8.7% of the company C (calculated).² My base says that is sensible at this price. I can’t yet say it was sensible at $105.

In their own words, Warner Bros. was “a ‘nice to have’ at the right price.”⁷ I respect the walk. It left them richer than they started. Yet for almost three months they stood behind an all-cash purchase with $42.2B of bridge commitments, nearly three times their gross debt, for a business unlike their own. It tells me that the capital policy has more room for a large deal than the buyback record suggests, and I treat that as a risk below.

VFinancials

Ten years of Netflix show one big change. For the first half of the decade it reported profits and burned cash. Since 2023 profits and cash have nearly met.

Revenue has compounded at 19.9% a year since 2016, operating income at 48% C (calculated). Growth was only 6.5% in 2022 and 6.7% in 2023, and then it recovered to about 16% on the password crackdown, price and ads. It is now slowing again: 17.6% in the fourth quarter of 2025, then 16.2%, 13.4%, 11.8%, and 11.5% forecast for the fourth quarter of 2026.² I wouldn’t call that alarming. I read it as a business returning to its normal rate after a one-off gain from shared passwords.

What an owner could really take out

This year’s reported numbers need the most care of any on this page. 2026 net income comes to about $15.1B, on the nine months plus the fourth-quarter forecast C (calculated).² That includes the $2.8B Warner fee, which I put at about $2.1B after tax, using the 16% tax rate of the second and third quarters as the rate on everything else E (estimate). Underlying net income is about $13.0B, or about $3.07 a share. Free cash flow is guided to about $12.0B.² The fee lifted that guide from about $11B to about $12.5B, so I take off $1.5B. I add back about $0.5B for a real-estate deal in the fourth quarter and the $729M of non-routine Brazil tax payments, and that gives about $11.7B.¹³²³ Take out about $0.5B of stock pay, and $11.2B is left for an owner E (estimate).

So the gap between reported earnings of $15.1B and what an owner could take is about $3.9B, and most of that is one payment. Against underlying earnings the gap is about $1.8B, or 14%. Content explains about all of it: about $2B more cash than amortization, on management’s ratio of about 1.1 times and amortization up about 10%.²⁴ Capital spending above depreciation adds a little, and other items take a little back.

Is that $2B a cost of standing still? I think it mostly is. A service that stops adding to its library doesn’t stay the same size, it shrinks in members’ eyes. The whole $20B a year is closer to maintenance than to growth, far more than the capital spending on most income statements. That’s why I don’t treat cash above amortization as an investment that will pay out later. It belongs in the cost of the business. The good news is that stock pay is small, about 1% of revenue, because staff choose between cash and options.¹⁶ At most software companies this gap runs the other way.

Quality of earnings

Three things deserve a footnote. First, content amortization is an estimate. Netflix chooses how fast to write off each title, and a slower schedule would lift margin without any cash. It discloses no change that worries me, and cash near 1.1 times amortization is my check on it. If the ratio climbs, earnings are being flattered. Second, 2025 carried a $619M Brazil tax charge in cost of revenues, which is why the third-quarter margin was 28.2% and this year’s 34.6% looks so strong: about 5 points of the 6-point rise is that charge C (calculated).³³⁶ Third, the third-quarter margin was ahead of plan on timing, and the fourth is guided to 26.0%.² One quarter tells you little in a business with this much seasonality in title launches.

The balance sheet

Gross debt is $15.3B and cash $8.2B, so net debt is about $7.1B.² The bigger claim sits partly off the balance sheet. Content obligations were $25.2B on 30 September, of which $19.6B wasn’t yet recorded as a liability. $11.8B falls due within a year.³ The company also warns that some deals for future titles have unknown amounts that “are expected to be significant.”³ That is why I do not add these to debt: they are next year’s programme, already in my cash flows. But they do mean that a large part of the $20B can’t be cut quickly if revenue slows. The fixed-cost scale that helps on the way up hurts on the way down.

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, and none needs a surprise.

• Attention moves and price can’t follow. Revenue per hour rose 12% in a year. That can’t go on for a decade if hours per member keep slipping. In a mature market like the US, management says growth skews to price and ads.⁹ If members one day feel the price against the hours, growth falls to mid-single digits in the richest region first, and every other number on this page depends on that.
• It has to spend to stop the slide. Today’s margin story rests on content growing slower than revenue. If Netflix must buy back viewing with sports rights or bigger budgets, that reverses. Content cash went from $17.7B to about $20B this year, up about 13%, as fast as revenue, after four years of barely moving C (calculated).¹⁶⁷ With $25.2B already committed, costs are slow to adjust.³ The result is a bear case that looks like a very good media company and not a compounding one.
• Good cash goes into a large deal. Netflix signed for Warner Bros. with $42.2B of bridge commitments, and its stated priority still puts selective M&A ahead of buybacks.⁵² A bad $80B deal in a soft patch is how a strong balance sheet goes. If such a deal lost half its value, that alone is about $10 a share E (estimate).

All three are versions of one question: is this a utility people keep paying for, or a habit that can fade? They set my bear case. I keep a fourth outside the model, because I can’t size it: AI-made video that makes good long-form content plentiful.

What the price implies

At $62.08 the market values Netflix at $259B on 4,167M diluted shares, the 4,106M outstanding on 30 September plus 61M of dilution from options at the third-quarter rate C (calculated).¹³ That is 20.2 times underlying 2026 earnings, and 16.6 times the Nasdaq consensus of $3.75 for 2027, or 14.0 times $4.43 for 2028.³⁷ The $3.59 consensus for 2026 appears to include the Warner fee, so it is not comparable.³⁷ In June 2025 the same stock sold above 40 times forward earnings, by Pershing Square’s count.²² The owner yield today is 4.3% E (estimate), against the 4.94% ten-year Treasury.³⁸

What does it pay for? Holding margin flat at 33%, slightly above today, the price needs revenue to grow 8.5% a year for ten years. Or on my base growth, the price works with the 2036 margin at 33%, barely above the 2026 guide C (calculated). Either way, the market is pricing growth or margin gains well short of Netflix’s recent record, which is less common on this site than the reverse.

What I can’t judge

Two inputs decide this valuation and I can’t know either. One is how long price can outrun hours. Netflix has done it for three years. Nobody has run a service like this at this scale for long enough to know where the limit lies. The other is the normal margin. Management says it sees plenty of room and intends to raise margin each year.⁵ I know of no streaming service that has reached 40%, so I have no precedent to lean on. So here is the range, not one number. Read my $76 with the grid beside it.

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.³⁸³⁹ My ten-year beta on month-end returns is 1.22, which gives 10.1%. The five-year beta of 1.60 gives 11.7%, and the three-year beta of 0.91 gives 8.8% C (calculated).¹⁴⁰ The range is wide because the stock moved with the market in some years and not in others. I take the ten-year figure, rounded. It is a little under the 10.5% on my Broadcom page, and above the 9.0% on my Visa page, because subscriptions are steadier than chip orders and less steady than card payments.

Three cases

The model is simple on purpose. I set revenue growth and operating margin for each year from 2027 to 2036. I tax operating income at 17%, the nine-month rate. Then I take off cash content above amortization, capital spending above depreciation of 1% of revenue, and 5% of added revenue for working capital. Stock pay is already an expense and I don’t add it back. Amortization is 35.4% of revenue in 2026 and falls by 0.6 of each point of margin gain, because content is where the margin comes from. What is left is owner cash, before interest.

My base 2027 operating income of $18.7B is in line with what the Street’s $3.75 implies, and my 2028 of $21.5B fits $4.43, on my own rough sum for shares and interest E (estimate).³⁷ So the near years are not where I differ from the market. The disagreement is over what happens after that.

Where would $74B of 2030 revenue come from? I think of it as about $8B of ads and about $66B of subscriptions, which is about 8% a year on the subscription side E (estimate). Ads at $8B would be about a fifth of what YouTube’s ads make today, across 27 countries once the planned launches are done.³⁶²¹⁰ It is a stretch but not a fantasy. If ads reach only $5B, subscriptions must grow about 9% a year instead, which is a harder ask with hours at +2%.

The price sits 18% below my base, and that is the plain result. But why might the market be right? Because it can see what I can: hours growing 2% while revenue grows 13%, US share of TV down from December, an ad forecast cut, a new content chief, a thousand jobs gone, a co-CEO saying execution has to improve. Stocks don’t fall 54% on nothing, and the market may be saying that this is a great franchise at a turning point in how people watch. If so, my base growth is too high, and the grid shows what that costs. The $42 bear is a fair picture of that world.

Other uses of the same dollar

The ten-year Treasury pays 4.94%. Against that, my base here returns 11.3% a year and my bear 7.8% E (estimate).³⁸ The closest comparison on this site is Alphabet, because it owns the very thing taking Netflix’s hours. There I have 5.4%, 8.3% and 10.9% for bear, base and bull at $283.82. So the owner of YouTube is priced for a lower return than the company it is pressing. That isn’t an error. Alphabet has more ways to win, and Google Cloud grew 103% in the third quarter C (calculated).¹⁰ For the same dollar, Alphabet offers less on my numbers and much more diversity. Visa is 8.2% on my base and Mastercard 8.4%, with steadier businesses. Those pages use their own prices, rates and methods, so I treat gaps of a point as noise. Netflix’s gap is wider than that. It is what the market charges for the hours question.

What past lows looked like

The stock has fallen more than 30% from a record high three times in ten years: 44% in 2018, 76% from November 2021 to May 2022, and 54% so far C (calculated).¹ The 2022 low was $16.64, about 14.8 times 2021 earnings, in a year when owner cash had been negative.¹¹⁸ The same multiple on underlying 2026 earnings is about $45 C (calculated). That is not a forecast. It lands near my bear, and it shows that the bear is a real price this stock has paid before, not a remote one.

The margin of safety

• Price$62.08Close, Q4 2026 [1]
• Base value$76My estimate E (estimate)
• Margin of safety below$56set from my bear case C (calculated)

I set the margin of safety from the bear case: at $56, my $42 bear value is at least three quarters of the price paid, and the price is 26% under my base value. Why the bear? Because the danger here is the shape of the business changing, and a discount to base wouldn’t protect against that. At $56 my bear returns 8.3% and my base 12.1% a year E (estimate). The stock last closed below $56 on 1 May 2024, in a year when operating income was $10.4B, against about $16.1B guided for 2026.¹¹⁶² At $62.08 the price is 11% above that level, so there is some cushion against my base and no full margin of safety. It doesn’t cover the stress case either: growth of 3%, margin falling to 25%, content at more than 1.1 times amortization, which gives $29.

Two different questions

How good is the business, and how good is the price? They are separate questions. As a business Netflix is excellent on most tests I have: global scale, no serious paid rival, high returns, little stock dilution, cash that matches earnings. Its one real flaw is the most important one: the product is taking a smaller part of each member’s day while the price goes up. As for the price, it was poor at $133, when the stock paid for all of this and more; my base gave 7.2% a year there. At $62.08 it looks reasonable on my numbers, with the shape of the business still in doubt. So this looks like a good business at a fair-to-good price with an open question, which is a very different thing from a good business at a bad one. The first reading depends on the hours, and I can’t prove that part.

Where this could be in 5 and 10 years

Here are my estimates for Netflix in October 2031 and October 2036 E (estimate). The base is $103 at five years and $126 at ten, after $47 a share of owner cash has been paid out.

For the base to hold at five years, revenue must reach about $81B by 2031 at a 37.5% margin. For ten years, growth must still be 6% in 2036 and the margin 40%, with content cash staying near 1.06 times amortization. Nearly all of the year-10 value is the years after 2036, so it is the least sure number here.

The bear and bull differ on two things. Bear growth fades to 3.5% and margin slips to 30%. Bull growth only fades to 7.5% and margin reaches 45%. The bear ends at about today’s price, with only the cash to show for ten years. The discount rate matters too: at 11% the base today is $65, not $76.

The most likely way the base is wrong is hours. If time per member keeps slipping, growth falls first and Netflix must spend more on content to hold attention. That moves margin and growth together, toward the bear. The return column is not the 7.8%, 11.3% and 13.4% earlier, which set the end value at that return.

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

VIIThe case for the price

What most people miss

Three things cut in Netflix’s favour and get little attention. Cash content spending in 2025 was $17.7B, almost exactly the $17.5B of 2021, while revenue rose by $15.5B.¹⁷¹⁶ Second, hours are a poor gauge of what live pays for: just over 5% of spending, about 1% of hours, and six of the top ten sign-up days of the last five years.⁴ Third, the share count is now falling fast. It dropped 2.7% in nine months, and at $62.08 the $22.3B of remaining authorization is 8.7% of the company C (calculated).³²

The bull case

In plain words: Netflix won, the cost of content is spread over the largest base in the industry, and the stock fell on worries about watch time that don’t show up in results. The clearest version comes from Pershing Square USA, Bill Ackman’s fund. Its report for 30 June shows 3,156,394 shares, 4.9% of net assets, at about 21 times forward earnings. It says Netflix “has since effectively won the streaming wars.” It argues that investors fix on hours without weighing quality or geography, and that short video takes time mostly from linear TV and weaker streamers. It expects earnings to compound at close to 20% a year.²² This is a second try: the firm briefly owned Netflix in 2022, at a loss the press put above $400M.²²⁴¹⁴⁸

Others moved at the same time, in both directions. At 30 June, the latest 13F date, Fundsmith, Terry Smith’s firm, held a new 6,990,345 shares. Coatue had cut to 4,710,420 from 6,893,560, Tiger Global had sold its 2,439,000, and ARK held 224,871.^42434445 Those positions are four months old, and they don’t say why.

Against reality

Much of that case is already in the numbers. Margin went from 21% to a guided 31.5% in five years.¹⁷² Consensus has earnings of $3.75 in 2027 and $4.43 in 2028, about 22% and 18% growth on my underlying $3.07 C (calculated).³⁷ My base has operating income growing 14% a year to 2030, so with 3% fewer shares each year it reaches about 17% a share. That is close to Pershing Square’s figure for four years. After that my base slows to about 12% E (estimate). So if Netflix sustains double-digit growth to 2028, reaches about $8B of ads by 2030 and keeps content below revenue, my $76 or better is possible.

Where the argument is weakest is engagement. The report says short video has had no effect on results, and that is true so far. Hours, though, grew 2%, YouTube has about twice Netflix’s TV share, and ad revenue just missed its plan by about 10% C (calculated).⁴⁶² Results lag viewing, and that’s the risk in using them as evidence.

It helps to multiply it out. My base needs five things. Double-digit growth to 2028: 65%. Ads near $8B by 2030: 50%. Content below revenue, so margin reaches about 36% by 2030: 70%. No value-destroying large deal: 85%. Viewing per member stabilizes: 50%. Those are my guesses. Multiplied, they give 10% E (estimate). They overlap, so perhaps 25% in practice. The price needs much less. Growth of about 8.5% with no margin gain, no large deal, and hours that don’t fall outright: say 70%, 75% and 85%, or about 45% before overlap E (estimate). That is why the price can ask for less than my base without my base being likely.

What a 10x would take. Ten times $62.08 is $620.80, about $2.6T of market value on today’s shares. At 25 times earnings that needs about $103B of net income, eight times this year’s underlying figure. At a 35% net margin, well above today’s 25%, that is about $296B of revenue, nearly six times 2026, and more than half of what all of Alphabet will make this year, on about $357B in nine months C (calculated).¹⁰ By 2036 it needs 19% revenue growth a year for ten years, from a company now growing 12%. Buying back 3% of the shares a year would cut those sums by about a quarter, and it would still take about $219B of revenue. The one case I can show of a tenfold rise from this size in my ten-year file is NVIDIA, from about $260B on 29 July 2020 to about $2.6T on 24 May 2024, on about 24.6B split-adjusted shares and a surge in chip demand C (calculated).⁴⁶⁴⁷ I know of no subscription business that has done it. For Netflix, tenfold from here is close to impossible, and I would not build a case on it.

Where I come down

The cash record moved me: four flat years of content spending is stronger than I expected, and it is why my base margin runs to 40%. The argument that hours don’t matter didn’t move me, because results are the last thing to change. A year of hours per member rising, on any published measure, would raise my values most.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• Advertising. Netflix has built its own ad technology, and it missed this year’s plan.⁹ My base needs a business about three times this size by 2030. Ad markets fall in recessions, and subscription revenue mostly doesn’t, so a larger ads line also makes results less steady.
• Leadership. A new content chief, a founder gone from the board, a new chairman and about 6% of staff cut, all within a year.²³²⁷²⁶ Changes that large can help, and they can also unsettle the creative people a studio depends on. Sarandos himself said talent has choices.⁹
• Sports rights. Netflix has an expanded NFL slate and live events. If it moves to full seasons, costs change in kind.⁴
• Currency. 57% of revenue is outside the dollar, against 30% of operating costs, and the company hedges only part of that.³
• Tax disputes. One dispute in Brazil brought a $619M charge in 2025 and $729M of non-routine payments in 2026. Non-income tax audits continue elsewhere.³⁵
• Regulation of content and pricing. The 10-K lists local content quotas, investment obligations and levies as risks.¹⁶
• AI. Netflix uses it, and Pershing Square argues that it stays expensive for long-form work.⁴²² I can’t tell how fast this changes. If good video gets cheap, the value of a large budget falls, and so does the barrier that protects it.
• Suppliers as rivals. Sarandos said: “all of our suppliers are also our competitors.”⁹ Licensed titles can leave. Skydance now owns a much larger library.²⁴