Spotify Technology

IThe business

Spotify rents the world’s music and resells it by the month. In the twelve months to 30 September 2026 revenue was €18.8B, and 90% of it came from 306M Premium subscribers C (calculated).⁵²³⁴ The other 10% came from ads sold against 499M people who listen for free.⁵ The fiscal year ends on 31 December, so TTM on this page means the four quarters to 30 September. Spotify reports in euros and its shares trade in dollars on the NYSE, so this page uses both, and I say which.

The model is simple and has one hard feature. Spotify doesn’t own what it sells. It pays rights holders a share of revenue, or a minimum per subscriber, whichever is greater.² Cost of revenue, mostly those royalties, took 67% of revenue in the TTM, which is why a business with 788M monthly users has a 33% gross margin C (calculated).⁵² At its Investor Day, Spotify said it paid more than $11B to the music industry in 2025 alone.⁸

What changed is that the business stopped being only a growth story. From 2016 to 2023 Spotify lost money at the operating line in every year but one. In the TTM it earned €2.7B of operating income and €3.2B of free cash flow C (calculated).⁹²⁵ Two things did it: price rises, and a cut in costs that took average headcount from 9,123 in 2023 to 7,287 in 2025.⁹²

What it sells

• Premium. Monthly plans for individuals, couples, families and students, with music, podcasts and a monthly allowance of audiobook hours. Revenue was €4,487M in Q3, up 15% at constant currency, on 9% more subscribers and average revenue per user, or ARPU, of €4.95 a month, up 7.5%. Gross margin was 34.5%.⁵
• Ad-Supported. Free listening with ads. Revenue was €482M in Q3, up 9% at constant currency after 3% in each of the two quarters before it. Gross margin was 18.7%.⁵⁶ Automated, bid-for sales are now about 40% of the business, by the CFO’s account.⁶
• The Marketplace. Tools that let labels and artists pay to promote their music. The 6-K calls the benefit to music costs “marketplace programs”, so much of it shows up as lower cost and not as a revenue line.³ Spotify said 2025 gross profit from it was four times its 2021 level, and gave no euro figure.⁸
• Add-ons. Audiobooks+ sells extra hours and was on track for $100M of annual recurring revenue in July, by company statement. A paid tool for AI covers and remixes, licensed by Universal Music, is the next one.⁸¹⁰

Where growth comes from

Growth is two multiplications: more subscribers, and more revenue from each. Subscribers have gone up by at least 25M every year since the 2018 listing, through a pandemic, price rises and a bad ad market.⁶⁹² The Q4 guide is 316M, which would be 26M for 2026 C (calculated).⁵ Norström put paid penetration at nearly 50% of the population in Sweden, 20% to 25% in the UK and Germany, about 11% in Latin America and under 3% across emerging markets.⁶ So the room is real, and most of it is in places that pay less.

That is why ARPU matters more than the subscriber count. It fell for years, from €6.20 in 2016 to €4.29 in 2021, as family plans and cheaper countries came in C (calculated).⁹ Then Spotify began raising prices. The US Individual plan went to $11.99 in July 2024 and $12.99 in early 2026, and last year’s round covered more than 150 markets.¹²⁶ Net additions held up through all of it. That is the most important fact on this page, and I come back to it.

The new shape of the costs

2026 was the year Spotify spent again. Operating expenses rose 26% in Q3, or 19% without currency and payroll taxes tied to the share price. R&D rose 35%, of which €58M was cloud and AI.⁵³ Marketing rose too. Management told investors this would peak in Q3 and moderate from Q4.⁶ The guide for Q4 implies it: €938M of costs on €5.2B of revenue, up 18% on Q4 2025 C (calculated).⁵ Q4 is seasonally strong, so one quarter proves little. The operating margin has gone 15.8%, 13.7%, 13.5% this year, on a guide of 15.8% for Q4 C (calculated).⁵

One line can mislead. Spotify pays payroll taxes in some countries on the value of share awards, so its reported costs move with its own stock. These Social Charges were €22M in Q3 and −€16M a year earlier, and a 10% move in the share price is worth about €20M in a quarter.⁵ When the stock falls, profit looks better, and the reverse. Nine-month R&D rose only 4% partly because of this.³

Spotify’s product is excellent and its supply is rented. Everything hard on this page comes from that gap.

IIThe moat

A rival can license the same songs. Apple, Amazon and Google all do. So the catalog protects no one, and I think what protects Spotify is what happens after the licence: the habit, the record of your taste built up over years, and being on every device you own. Those are real. They are narrower than what a firm gets from owning its product, and they have to be defended against companies with far more money.

The evidence is in behaviour under price rises. In Q3 ARPU rose 7.5% at constant currency, and subscribers rose 9%, with every region up.⁵ A business with weak loyalty can’t do both at once. Over 100M people use it about 28 days a month, Söderström said on the call.⁶ Those users have playlists, years of history and, for many, a family plan shared with others. Leaving costs effort, and the product gets better the more you use it. Luiga told Fortune that it’s much more costly if people leave than to win new ones.¹³

Evidence it is widening

Paying more, leaving less. Spotify says US customer lifetime value is up more than 70% since 2022.⁸ That is a company claim with no published method. The two things I can check, ARPU and net additions, point the same way.

More reasons to open the app. Audiobook listening hours rose 60% from 2024 to 2025. Podcasts are in their second year of gross profit. Both now run above a 30% gross margin along with music, the company says.⁸ Audiobooks go from 22 markets to more than 180 by year end.⁵ Management says audiobook users come on more days of the month, which is the measure it cares most about.⁶

Taste as an asset. Spotify’s model of what people like now sits behind a talking interface, and it says tens of millions of people use it.⁶ If AI lowers the cost of building a music app, the thing left that is hard to copy is the data, and I think Spotify has more of it than any rival. I can’t size that. I believe the direction.

Evidence it is narrowing

The suppliers are getting paid first. Warner’s chief executive told investors in August that per-subscriber minimum increases now cover 88% of its subscription streaming revenue, up from zero two years ago, across all its major streaming partners, and that this added 3.5 points of growth in the quarter.¹⁴ Those clauses move some of each price rise to the label by contract. Warner calls them a win for everyone. For Spotify they cap the benefit of pricing.

YouTube is catching up. Alphabet’s count of YouTube Music and Premium subscribers, trials included, reached more than 150M, up from 125M.⁷ It sits alongside the largest video app in the world and needs no music profit to justify itself.

AI floods the supply. Deezer says fully AI-generated tracks passed half of its daily uploads at a June peak, about 90,000 a day. They were only 1% to 3% of streams, and up to 85% of those streams were fraudulent in 2025.¹⁵ Listening hasn’t moved, so this is a cost and a nuisance today. If AI music ever became a real share of listening, the labels’ share of the pool would fall, which would help Spotify. That risk cuts both ways, and I can’t call it.

Spotify is protected against people leaving. It is much less protected against the people it pays.

IIIThe competitive advantage

The last section asked what keeps customers in. This one asks why Spotify earns more than a rival would in its place. Scale is the first answer. With about twice the subscribers of YouTube’s two products combined, it has more data, more leverage over a fixed product and engineering budget, and more audience to offer anyone with something to promote.⁵⁷ Costs prove it: from 2023 to the TTM, research and development fell 16% while revenue rose 42% C (calculated).⁹²³⁴ Revenue per average employee went from €1.45M in 2023 to €2.36M in 2025 C (calculated).⁹²

The share data agrees. MIDiA Research put Spotify at 31.4% of the 921.6M music subscribers worldwide at the end of 2025. Tencent Music had 13.8%, Apple Music 12.6%, YouTube Music 12.4% and Amazon Music 8.5%. Apple and Amazon publish no counts, so MIDiA estimates those, and it counts YouTube Music with YouTube Premium. Two points matter more than the rank. Spotify’s share is bigger than the next two services combined. And YouTube Music has gone from 7.9% at the end of 2020 to 12.4%, while Apple Music fell from 18.4% to 12.6%, so the gap is closing from the side that has no need to earn a music profit.⁴² Treat these as a third party’s estimates.

The second answer is focus. For Apple, music sells phones. For Amazon it fills out Prime. For Google it is a line inside YouTube. Spotify has to get audio right or it has nothing, and for twenty years it has shipped the better product. That is an advantage while audio stays a product decision. If it becomes a bundling decision, it turns into a weakness, because the bundlers don’t need to earn a return on music.

Netflix is the comparison investors reach for, and it shows the limit. Netflix’s operating margin was 33.5% over nine months of 2026, against 14.6% here C (calculated).¹⁸ The difference is who owns the content. Netflix pays a fixed cost and keeps the upside when it adds subscribers. Spotify pays a share, so more revenue means more cost, and only part of the gain falls through. Spotify’s board knows this comparison better than anyone: Netflix co-CEO Ted Sarandos is a director.²¹⁹

The four licensors matter most. Universal, Sony, Warner and Merlin supplied about 72% of streams of audio delivered by labels in 2025.² The licences run for several years, don’t renew on their own and can include most-favoured-nation clauses, so a concession to one can spread to others.² Asked about the Universal and Warner deals reaching their two-year mark in Q1 2027, Norström called it a recurring two-year rhythm and said interests are aligned.⁶ They may well be. Labels need Spotify’s growth, and Spotify needs their catalogs. Neither side can walk away. So the split is what’s up for discussion every two years.

The third answer: what the money does next

The newest source of advantage is the one I’d watch: products that sit on top of the licence. The Marketplace, audiobooks, podcasts, Audiobooks+, Creator Memberships and paid remixes each earn money that doesn’t run through the old royalty formula, or runs through a better one. Luiga said about a third of the gross margin gain came from music, so about two thirds came from elsewhere.⁸ If that continues, the gross margin can rise even if the labels hold their share of music. That is the honest bull thread here, and the gap between my base and bull is mostly a question of how far it goes.

Spotify’s edge is that it does one thing better and at greater scale than anyone. Its limit is that the things it does best are licensed from others.

IVManagement

Daniel Ek ran Spotify from its founding until 31 December 2025 and is now Executive Chairman. Alex Norström and Gustav Söderström became co-CEOs on 1 January 2026.² Both were already co-presidents, so this was a handover to insiders. Two chief executives is usually a warning. Their first year has gone to plan: Q3 met or beat every guide in the update, and revenue and users have held up. One year proves little.⁵ Christian Luiga is chief financial officer.² A new chief accounting officer started in May, and Heidi O’Neill left the board in September, with no disagreement, the company says.²⁰²¹

Follow the incentives

Here is how they are really paid. Cash is small: salaries of about $340,000 for each co-CEO and no annual bonus, which the 20-F says is on purpose.² Nearly everything is long-term equity. In 2025 each co-CEO received $12.5M of it, split between options and restricted stock, and it vests monthly over four years on service alone.² There is no performance test on revenue, margin or return on capital. Options last only about five years. Every employee picks a mix of options, stock or cash, so people self-select into how much of the share price they want.²

So what does that pay them to do? Get the share price up over the next few years. That lines up with owners on growth and with not wasting money, and the option form rewards volatility too. It doesn’t pay them for per-share cash over ten years, and it doesn’t punish them for a bad label deal in 2031. The 25th-percentile cash and the 50th-to-75th-percentile total pay against peers is sensible.² I’d prefer a performance hurdle. Few tech companies have one, so I don’t mark this down much.

How much of their own money is in? Less than the headline suggests. Ek’s 13.9% in the table of holders includes 16.6M shares belonging to Tencent, on which he only holds a voting proxy. His own stake is 11,993,298 shares, about 5.8% of the company, or $7.1B at the close C (calculated).²²² Since 2017 he has taken no salary, bonus or new equity from the company.² That is a real alignment. He has also been a steady seller: Form 144 notices add up to about $808M from 2023 to 2025 C (calculated).²⁴ I found no sale by him in 2026 in Forms 4 or 144. Lorentzon’s holding company filed notices of about $1.2B over 2024 and 2025 C (calculated).²⁴

The co-CEOs own little stock outright and a lot of options. Söderström had 19,908 shares after a 12 October sale, worth about $12M, and in 2026 he has sold 145,988 shares for about $72M on a Rule 10b5-1 plan, much of it from options that expire in March 2027.²³²⁴ Norström sold 27,180 shares for about $13M.²⁴ At the 20-F date Söderström’s options had about $263M of value above strike at today’s price, and Norström’s about $102M C (calculated).² Those are large numbers, and they keep the two of them tied to the stock. Exercising before expiry is normal. The point is that the alignment is through options that will be gone in a few years, more than through shares they mean to keep.

Control sits with the founders. Each beneficiary certificate carries a vote and no economic right, and the two founders hold all 309.9M of them.²³ Lorentzon has 40.5% of the votes and Ek 28.8%, so together 69.3% C (calculated).² Outside owners can’t change anything without them. At the April annual meeting Lorentzon took 427.7M votes for and 44.9M against.¹⁹ Ek no longer runs the company and, with Lorentzon, can still decide its future.

What they did with capital over a full cycle

Spotify gives us one full cycle to judge. It raised money cheaply in 2021, its stock fell about 80% from the 2021 high of $364.59 to a close of $71.05 in November 2022, and it then rebuilt.¹

I give that record decent marks with one blot. The early buybacks at about €117 a share are very good, though they came when the company was still losing money. The stop during the 2022 fall is the opposite of what an owner wants: repurchases ended just as the shares got cheap, though with cash flow near zero and notes coming due it was defensible.⁹ The blot is 2018 to 2023. Spotify spent over €1B on podcast and audiobook deals while gross margin stuck between 25% and 27% for six years and losses ran on.⁹ Then it cut a fifth of its average headcount. That was late, and it was right. Podcasts now make money.⁹⁸

The current plan is clear: organic first, then deals, then return of excess cash, on what Luiga calls an opportunistic basis.⁶ Repurchases of €851M in nine months are the fastest ever, and the board added $1.5B to the program in August.³²⁵ About $1.88B was left at 30 September.³ The company has €9.9B and no debt, and it pays no dividend, so that cash earns interest that its owners could earn themselves.³ I would like to see a stated view on how much cash is enough. Strategic flexibility, the phrase on the call, can mean anything.⁶

The people are good, and the record since 2023 is better than good. The structure asks you to trust two founders who are no longer running the business, with no vote of your own.

VFinancials

The ten-year record has two halves. From 2016 to 2023 revenue grew 24% a year and Spotify lost €2.1B at the operating line C (calculated).¹¹⁹ From 2023 to the TTM revenue grew about 14% a year and operating income went from −€446M to €2.7B C (calculated).⁹²³ Gross margin did most of the early work, from 13.6% to 25.0% by 2022. Operating costs did the late work: 19.2% of revenue in 2025 against 29.0% in 2023 C (calculated).⁹²

What an owner can really take out

Spotify’s headline number is free cash flow of €3.2B. It needs almost no capital, just €62M in the TTM, so that looks like nearly all cash.⁵ It is a very good number. It also overstates what is left for an owner, in four ways.

• Interest. €214M of the TTM came from the cash pile.²³⁴ I count that cash separately, so I take the income out.
• Tax. Spotify paid €200M of tax in the TTM on €2.7B of operating income, about 7%, thanks to old losses and timing differences. The nine-month accounting rate was 24.5% C (calculated).²³⁴ Cash tax has more than doubled in a year and is going to catch up.
• Leases. €70M of lease payments sit below the cash flow line.²³
• Stock pay. €286M of expense is added back, and it is a cost.²³⁴

So I put what an owner keeps at about €2.2B, which is 32% below reported free cash flow and 27% below the €3.0B of net income E (estimate). Net income itself includes €320M of net finance income in nine months.³ At €99.3B for the operating business, the price is 46 times that number and 31 times reported free cash flow C (calculated). That is the first thing I’d want a reader to take away: Spotify looks cheaper on free cash flow than it is.

There is a second, harder layer. Stock pay expense understates what it costs to keep the share count flat. In nine months shares outstanding fell only 0.15%, from 205.8M to 205.5M, and it took €775M of cash to do that C (calculated).²³ Staff exercised 1.38M options at an average strike of $167, when the stock averaged over $500.³ That is the old cheap equity coming home. On the call an analyst put the dilution from staff awards at about one million shares and management did not dispute it, and 1.74M shares went out in nine months C (calculated).⁶³ So while those options run off, the honest figure is nearer €1.4B E (estimate). It should improve: only 3.97M options remain, 2.59M exercisable, and they are in my share count already.³ New grants come at today’s prices, about 692,000 restricted units and 787,000 options so far this year.³ But if you value Spotify on €3.2B, you are counting cash that left through the financing line.

The balance sheet

At 30 September Spotify had €9,888M of cash and short-term investments and no debt but leases, after settling its exchangeable notes for €1,304M in March.³ It also owns about 9% of Tencent Music, carried at €990M after a 53% fall in that stake’s value this year, and an interest in the distributor DistroKid, part of which it sold.³ The cash isn’t all free. €2,025M is owed to rights holders, and €794M is deferred revenue.³ Spotify collects from subscribers before it pays labels, and that float funds it for as long as revenue grows.

Commitments are small compared with the AI builders on this site, and that is worth saying plainly. Minimum guarantees are €1,740M, down from €2,613M at December, and purchase obligations, including Google Cloud, are €1,996M.³ Spotify rents compute. If AI costs fall, it benefits; if they rise, it pays more in a cost line, and it isn’t sitting on a data center. It also has a dependence: it relies on Google Cloud for the vast majority of its computing, from a company that competes with it.²

The accounts are clean and the capital needs are tiny. The cash an owner keeps is about two thirds of the number on the slide.

VIValuation

How this goes badly wrong

Before any numbers I ask what would most likely wreck this, and work back. I see three ways.

1. The labels keep the gains. Spotify raises prices, and rising per-subscriber minimums pass much of each rise to rights holders, so ARPU climbs while the gross margin stays near 33%. Nothing breaks. Revenue grows and the promised margin never arrives. This is the most likely of the three, because it needs only polite negotiation. I build my bear case on it. Each point of gross margin held for good is worth about $20 a share E (estimate).
2. The customer moves up a layer. Rich Greenfield put it to management on the Q3 call: an assistant could compare prices, switch services or cancel ones that go unused.⁶ If people ask an agent for music and don’t care whose app answers, Spotify becomes a supplier, and pricing power goes to whoever owns the agent. Spotify’s answer is that it monetizes from the background and that people use it daily.⁶ That answer is good. It is also untested at scale.
3. A rich rival gives it away. Apple, Google or Amazon could price music as a loss-making part of a bundle, as they already partly do. YouTube’s count is growing. The 20-F says plainly that these rivals can offer services at little or no profit.⁷²

All three share a feature. None shows up in this year’s accounts. So I read the valuation for what it pays for, and how it would look if one came true.

What I can’t judge

I can’t tell you churn. The F-1 showed monthly Premium churn of 5.1% in Q4 2017, and the current 20-F prints no figure.¹¹² For a subscription business that is the key number, and it is hidden. I can’t tell you the label terms. They are confidential, and they decide my gross margin. I can’t tell you how much of the Marketplace benefit is durable, because Spotify gives no euro figure. I can’t tell you what AI does to the size of the pool for human music. So I show a wide range, and the $516 below is the middle of it, not a measurement.

What the price implies

At $593.44 Spotify’s market value is $123.9B on 208.8M shares, which count 205.3M outstanding after October repurchases, 1.26M restricted units and the 3.97M options by the treasury method C (calculated).³¹ The close is confirmed on a second feed.¹²⁶ That is €110.9B at the ECB rate of $1.1172 C (calculated).²⁷ The shares are 24% below the 2025 closing high of $775.90 and 44% above this year’s low of $412.75.¹ They rose 6.5% on the last day of the month, and I found no company news to explain it.¹ The price is 43 times Nasdaq’s 2026 consensus of $13.74 and 37 times 2027’s $15.97, with the basis unstated.²⁸ My own 2027 is about $15.56 with interest E (estimate), so I’m near the Street on next year. The gap is what happens after.

Take €2.2B as the true start and the price needs 18.4% a year for ten years. Take the reported €3.2B and it needs 13.1%. I’d use the first. Free cash flow grew far faster than that in the last few years, but it started from almost nothing. Operating income can grow faster than revenue for a while, and I expect it to. The test is whether it does so for a decade, from a level that already reflects price rises.

The discount rate

I use 10.0%. The capital asset pricing model with the 4.94% ten-year Treasury and Damodaran’s 4.20% premium gives about 8.5% on the three-year beta of 0.86, 11.2% on the five-year beta of 1.48 and 11.3% on the 1.51 since listing C (calculated).^3031321 Betas use month-end price returns against SPY, without dividends. That spread is too wide to treat as a measurement. Recurring revenue and net cash pull the rate down, and a supplier base that can change the economics every two years pulls it up. So I take 10.0%, a beta of about 1.2. That sits between the 9.0% on my Mastercard and Visa pages and the 10.5% on Broadcom. One point on the rate moves the base by about $73 E (estimate).

Three cases

Each case is built by year, 2027 to 2036, valued at 1 January 2027 in euros: subscribers, ARPU, ad revenue and add-ons, then gross margin, operating costs including stock pay, and 24% tax. Cash after that, plus working capital, is owner cash. I add net cash and investments and convert to dollars at $1.1172.

The base case takes management’s plan seriously and shades it.

• Subscribers rise 27M in 2027 and fade to 14M by 2036, to 527M E (estimate). That is 1.7 times today’s guide and continues a nine-year record, with fewer big easy markets left.
• ARPU grows 4.5% for two years and fades to 3%, to €7.02 a month E (estimate). That is a real price rise on a falling country mix. ARPU fell 2.4% a year from 2016 to 2026 and has risen 2.6% a year since 2021 C (calculated).⁹⁵ So my base assumes the recent pricing era lasts and gets a bit better.
• Ads grow 11% fading to 7%, to €4.4B E (estimate). Management is guiding toward double digits after 3% in the first half.⁶
• Add-ons reach €2.7B by 2036 E (estimate). No accounts exist to test that. It is about 5% of revenue.
• Gross margin goes from about 33.3% in 2026 to 36% in 2030 and 38% in 2036, at the low end of the 35% to 40% target for 2030 E (estimate).⁸ By a third-party account, Spotify reached its 2018 gross margin range late and beat the nearer goals set in 2022.³³
• Operating costs fall from 18.6% of revenue to 15%, so the operating margin is 19.5% in 2030, just under the target of above 20%, and 23% in 2036 E (estimate).

So my revenue for 2030 is €31.8B, or 13% a year from 2025. Mid-teens would be about €34.6B at 15% C (calculated). Jefferies modeled about 14% after the Investor Day, by a press report of its note.³⁴ Mine is lower because I expect growth to slow as price rises lap and I count euro revenue, not constant currency.

The bear case is way one: gross margin stays at 33.3%, net additions slow to 8M a year, ARPU rises 2% and ads about 5%. Costs still fall to 16.5% of revenue, so operating margin reaches 17% in 2036 E (estimate). It also pays €700M on the audiobook royalty dispute E (estimate).³ This is not a disaster case. It is a good company that doesn’t get its margin. The bull case is management’s plan delivered and then extended: €36.6B of revenue in 2030, a 38.5% gross margin that year and 42% in 2036, €9.0B of add-ons and a 28.5% operating margin E (estimate).

Weighted 25%, 50% and 25%, the cases give $533 E (estimate). At 20%, 50% and 30% it is $557, and at 30%, 50% and 20% it is $509 E (estimate). The value after 2036 is 59% of my base C (calculated), so most of the answer rests on a business a decade from now that I can only sketch.

What closes the gap

The simple way to read the gap: $77 a share is about €14B of value. No single small change closes it. Each of these does, on its own.

The first is the one bulls need to argue. ARPU growing about 5.9% for ten years means revenue of €61.2B by 2036. It would also need labels to accept that without taking more of it. A combination reads easier: ARPU one point faster every year and gross margin two points higher by 2036 is worth $585, about the price E (estimate).

That needs two things plus one absence. Say a 35% chance that ARPU beats my base by a point for ten years, 40% that gross margin reaches 40% through two or three rounds of renewals, and 75% that neither agents nor bundling cost it the customer. Multiplied, that is 10%, about one in ten E (estimate). They aren’t independent, since strong pricing helps the margin, so call it one in six. These odds are my judgment and I can’t defend the second digit. Against that sits the bull case above the price, and other routes, so the price is not absurd. It simply needs more to go right than my base assumes.

The margin of safety

What does the same dollar earn elsewhere? My base return at $593.44 is 9.1% a year, against 4.94% on a ten-year Treasury that takes no business risk E (estimate).³⁰ That is a 4.1-point premium for equity exposed to all three failures above. The closest comparables on this site are the toll-takers on daily habits. Mastercard is 13% above my base and Visa 18%, and those pages put the rate at which base meets price at about 8.4% and 8.2%. So Spotify’s return looks about a point better on my numbers, with faster growth. But Mastercard and Visa own their networks and have no one in Spotify’s supplier position. Tencent, which owns 8.1% of Spotify, sits 24% below my base value on its page.² Alphabet, the owner of its fastest rival, is 43% above. So on my numbers Spotify isn’t the stretched name in this group, and it isn’t the cheap one either.

How good the business is and how good the price is are two questions. How good is the business? Very good, and better than it was three years ago. The habit is deep, it needs almost no capital, and it has shown it can raise prices. It would be great if it owned its product. How good is the price? Not bad, and not generous. It is 15% above my base and pays for most of management’s 2030 plan before it happens. This looks like a very good business at a price that leaves little room for the renewal risk. The 31 times free cash flow that makes it look reasonable is 46 times what an owner keeps.

• Price$593.44Close, Q4 2026 [1]
• Base value$516My estimate E (estimate)
• Margin of safety below$413set from my bear case C (calculated)

I set the margin of safety so that my bear value of $310 is at least three quarters of the price paid, which gives $413, 20% under my base C (calculated). I use the bear because the risk here is one slow, contractual loss of margin. At $413 my base returns about 11.8% a year and the bear about 8.0% E (estimate). That price isn’t fanciful. The stock closed at $412.75 within the past year.¹ At $593.44 there is no margin of safety: the price sits 44% above that level. The cushion is for what the model doesn’t hold: a surprise in the renewals, the agent question, and my own error. I don’t cut the value again for the risks already in the cases.

Where this could be in 5 and 10 years

These are my model’s values at later dates. The value in a year is the present value then of the owner cash after it, at 10.0% and 3% growth after 2036, plus today’s net cash held at 10.0%. It is per today’s 208.8M shares, in dollars at $1.1172 to the euro, so it is not a price.

For the base to hold at five years, revenue needs to be about €35B in 2031 and the operating margin about 20% E (estimate). At ten, it is €50.9B and 23%. Owner cash goes from €3.0B in 2027 to €9.2B in 2036, and capital needs stay near 0.75% of revenue. The return is 8.3% against the 9.1% earlier on this page because that figure also lets the rate on later years fall. Here, value after 2036 is capitalized at 10.0%.

The bear is the labels keeping the gains: gross margin stays at 33.3%, net additions slow to 8M a year and the margin ends at 16.8%. The bull is management’s plan delivered and extended, with €9.0B of add-ons and a 28.5% margin. Growth and margin make the spread. The rate matters too. One point more on it takes about $90 off the base value at the end of 2036 E (estimate).

The most likely way the base is wrong is the labels keeping the gains. Gross margin stays near 33% and not 38%. Each two points of it is worth about $41 a share today E (estimate). The five-year value rests 74% on what comes after 2036, and the ten-year value on nothing else, so both lean on a decade I can only sketch.

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

Most of the margin gain hasn’t come from squeezing music. Luiga said about a third of it came from music, and music and non-music now both run above 30%.⁸ So the thing bears fear, that labels cap the margin, doesn’t cap all of it. In Q3, podcast costs and marketplace programs both helped.⁵

The minimum-guarantee burden is falling. It was €2,613M at December and €1,740M at September, ahead of the renewals, and that is the opposite of a firm being squeezed for more up-front money today.³ The number could jump in 2027, which is why it is in the table.

The two-year renewal is old news to Spotify. Norström called it a rhythm the company has had through its life, and gross margin has gone from 13.6% to 33% across those rounds.⁶⁹

The bull case

Management makes the strongest argument. Spotify is a daily habit with room to raise prices, and new products earn money outside the old royalty formula. Söderström says it is in ‘one of the few categories that actually stands to thrive’ in the agent age, because nobody wants a song summarized.⁶ The 2030 targets, set on 21 May, are a mid-teens revenue growth rate, a 35% to 40% gross margin and an operating margin above 20%.⁸ Those are company goals.

Jefferies lifted its target to $600 from $540 after that day, on the view that the targets eased fears about AI diluting margins. That is a press report of its note, dated 22 May.³⁴ Filings show who has backed it with capital, as of 30 June. Baillie Gifford had 7.93M shares, down from 8.22M.³⁵ Coatue had 1.31M, Tiger Global 1.17M, down from 1.58M, and D1 569,000, up from 340,000.³⁶³⁷³⁸ ARK had 75,108.³⁹ Lone Pine reported none, after 68,844 in March.⁴⁰ Those filings are months old. I found no public view on Spotify from Brad Gerstner, Gavin Baker, Cathie Wood, Bill Gurley or Bill Ackman, and Altimeter, Atreides, Pershing Square and Berkshire showed no position in the 13Fs I opened.

Against reality

If ARPU runs a point above my base every year and gross margin ends two points higher, the price is possible. That path is worth $585 E (estimate). It means about 5% ARPU growth for a decade, a 36.8% gross margin in 2030 and 40.0% in 2036, and an operating margin of 20.3% in 2030 E (estimate).

Today, pricing is ahead of that: 7.5% in Q3. But management guides to 4.5% to 5% for Q4 as last year’s round laps, so it is just about on the line.⁶ Gross margin is 33.0%, up 1.4 points in a year, and the full-year pace of about 1.3 points would reach the 2030 number if it held for three more years C (calculated).⁵⁶ Operating margin is the laggard: 14.7% for 2026 on the guide, so it needs to add about 1.5 points a year C (calculated).⁵ So the gross margin half is on track and untested by the renewals. The cost half depends on 2027 doing what management promised.

What a 10x would take. Ten times $593.44 is about $5,934, or $1.24T of equity on today’s share count. By 2036 that is about 26% a year C (calculated). At 25 times earnings it needs about $50B of net income. At a 30% operating margin and 24% tax that is about €195B of revenue, more than twice the $100B long-term goal and 2.8 times my bull case’s 2036 figure C (calculated).⁸ The nearest example is Netflix. It first closed above Spotify’s market value in February 2018 and is worth about $255B today, about twice that, in almost nine years C (calculated).⁴¹¹⁸ Chip makers on this site have gone ten times from this size, on a once-in-a-generation build. I can’t name a consumer subscription that has. For a business that pays most of its revenue to suppliers, it is close to impossible.

Where I come down

The pricing record and the falling guarantees moved me. They are why my base has ARPU rising near 4% and not 2.5%. They didn’t move my $310, $516 and $789, because the renewals are still ahead and I can’t see the terms or the churn. I would raise my values if the 2027 gross margin guide is above 35% after the renewals and net additions stay above 25M.

VIIIRisks, and what would change my view

1. Supplier power. Four licensors account for about 72% of label-delivered streams, and most-favoured-nation clauses can spread terms across them. Royalty formulas can also move with targets such as subscriber numbers and churn.²
2. Audiobook royalties. The Mechanical Licensing Collective claims Spotify underpaid US songwriters by treating Premium as a bundle. A court dismissed the first claim and an amended one followed. If the MLC won entirely, the liability through September would be about €538M, plus penalties and interest that Spotify says it can’t estimate, partly offset by direct publisher deals.³
3. Agents and AI. See above. Spotify has also tied itself to third-party assistants and said it is confident in that. The data stays with it, it says.⁶
4. Currency. The company earned 36% of TTM revenue in the US and reports in euros. In Q3 currency added about 190 basis points to growth, in the first half it took some away, and the Q4 guide assumes a 230 basis point tailwind.⁵³ My dollar value rises when the euro does.
5. Tax. The rate is headed up from a low cash base. At 28% my base is $493 E (estimate). Assessments of €43M on non-income taxes are being contested, and later periods aren’t estimated.³
6. Costs of AI. Q3 cloud and AI costs rose €58M in R&D alone. Management says they stay near these levels while moderating as a growth rate.³⁶
7. Tencent Music. The stake lost €1.1B of value in nine months and could move further, and China risk comes with it.³ Revenue there grew 5.8% in Q2.¹⁶
8. Control. Founders hold 69.3% of the votes, and as a foreign private issuer Spotify gives less disclosure than a US company.²
9. Key people. A new co-CEO structure, a founder who has stepped back and a board change this year.²²¹
10. Family plans. Management says tighter address checks are routine, and a large share of subscribers are on them.⁶ Done well, this helps. Done badly, it costs net additions.
11. Cash use. €9.9B invites a large deal. The record on deals is mixed.³
12. Dependence on Google Cloud, a rival’s platform, and on the app stores of Apple and Google.²

What would raise my base value: a gross margin guide above 35% for 2027 after the renewals, ARPU above 5% with net additions above 25M, or real add-on revenue. What would lower it: a jump in minimum guarantees, gross margin stuck near 33% through 2027, costs that don’t moderate as promised, or evidence that assistants are taking the customer. I’d also change my view on seeing churn. If Spotify ever prints that number, it will tell me more than any model here.