Business Overview

Spotify Technology S.A. (NYSE: SPOT)

6 September 2026

Evidence base: Spotify's SEC filings held in the SPOT research folder — the FY2021–FY2025 Forms 20-F (most recent filed 10 February 2026), interim 6-Ks through the quarter ended 30 June 2026 (filed 4 August 2026), and material-event 6-Ks through 3 September 2026. Industry structure, competitor economics and regulatory status are drawn from independent primary sources: IFPI, RIAA, MIDiA Research, the UK Competition and Markets Authority, the US Copyright Royalty Board and Federal Register, the Ninth Circuit, and the reported accounts of Universal Music Group, Warner Music Group, Sony, Tencent Music and Deezer.

This document explains how the business works and where its economics come from. It is not a valuation and contains no recommendation.

1. Executive Snapshot

2. What the Company Does

The customer problem, and the trade Spotify makes

A listener wants any recording, immediately, on any device, without owning or storing it. Spotify solves that by licensing catalogue rather than owning it: it holds essentially no copyright in the music it streams. Rights to sound recordings from Universal Music Group, Sony Music Entertainment, Warner Music Group and Merlin accounted for approximately 72% of streams of audio content delivered by record labels in FY2025 (FY2025 20-F). That single sentence defines the business. Spotify's product is distribution, discovery and convenience; the underlying goods belong to somebody else, and are rented on multi-year agreements that are not automatically renewable.

The company describes the transformation it caused as moving listeners from “a 'transaction-based' experience of buying and owning music to an 'access-based' model” (FY2025 20-F). The economic consequence of access is that the cost of goods sold is a contractual share of revenue rather than a manufactured item, so it does not fall as volume rises.

The unit: one Premium Subscriber account

Spotify defines a Premium Subscriber as a user who has completed registration and activated a payment method for its Subscription Offerings (FY2025 20-F). The definition counts accounts, not paying households: every member of a Family Plan (up to six) and a Duo Plan (up to two) is counted separately, as is a subscriber in a grace period of up to 30 days after failing to pay. This is why realised revenue per subscriber is far below the headline plan price.

Tracing one unit through Q2 2026: the account generated Premium ARPU of €4.89 per month. Premium cost of revenue ran at 65% of Premium revenue, leaving €1.71 of gross profit per subscriber per month. Nothing below that line is allocated to segments — the company states plainly that operating expenses are “managed on an overall Group basis” (FY2025 20-F, Note 23), so segment operating income is not disclosed and cannot be computed.

The cash timing matters more than the margin. Subscription Offerings are “paid for on a monthly basis in advance” (FY2025 20-F), while royalties are accrued and paid in arrears — accrued fees to rightsholders stood at €1,971m at 30 June 2026, roughly two months of Premium cost of revenue (inferred). Netting operating current assets against operating current liabilities gives approximately –€3.6bn of working capital at 31 December 2025 (inferred). Each new subscriber and each price increase therefore contributes cash before it contributes cost, which is why free cash flow (€2,874m in FY2025) has run ahead of net income and why the business has never needed capital to grow.

What it sells, and to whom

Two reported segments. Premium — 89% of FY2025 revenue and 94% of gross profit — sells ad-free, offline, on-demand access on Individual, Duo, Family, Student and Basic plans, with an Audiobooks+ add-on introduced in 2025. Ad-Supported — 11% of revenue and 6% of gross profit — gives the catalogue away with advertising and sells that inventory to brands, increasingly through the programmatic Spotify Ad Exchange launched in April 2025.

The company's own account of why both exist is that they “live independently, but thrive together”: the free tier is a funnel that drives “a significant portion of our total gross added Premium Subscribers” (FY2025 20-F). The funnel is working on its own terms — Ad-Supported monthly users grew 14% year on year to 494m at 30 June 2026 — but conversion is not improving. Premium subscribers were 38.6% of total monthly users at both December 2025 and June 2026, against 39.0% a year earlier (inferred from disclosed counts). The free tier is widening faster than it converts.

Lines entered and exited, and why the economics differ

Between 2019 and 2022 Spotify bought roughly €0.9bn of podcast assets — Anchor, Gimlet, Parcast, The Ringer, Megaphone, Podsights, Chartable — and told investors that exclusivity “helps differentiate our Service” (FY2021 20-F). Owned content is a fixed cost: the FY2021 filing warned that “given the multiple-year duration and largely fixed-cost nature of such commitments”, margins could suffer if usage disappointed. It did, and they did.

The reversal is visible in the balance sheet rather than in any announcement. Capitalised podcast content assets fell from €187m at end-2022 to €28m at end-2025, and content-asset amortisation from €208m in FY2023 to €125m in FY2025 (FY2025 20-F). The FY2021 rationale for exclusivity has been deleted from the filing; the FY2025 text describes podcast expansion in the past tense and leads with content “that we license from others.” The replacement model, the Spotify Partner Program launched 2 January 2025 and now in 19 markets, pays creators from consumption rather than commissioning shows — converting a fixed cost into a variable one. Crucially, none of the €0.9bn of podcast goodwill has ever been impaired; only €29m of produced content assets was written off, in 2023.

Audiobooks moved the other way. Findaway was acquired for €117m in 2022; a bundled allowance of listening hours was added to Premium during 2023 and is now in 22 markets against a subscriber catalogue of 500,000 titles. The economics differ from music in two ways: audiobook licences are consumption-based rather than a fixed revenue share, and — far more importantly — bundling audiobooks into Premium reclassified the US subscription as a “bundle” for statutory publishing-royalty purposes, lowering the mechanical rate Spotify owes. Section 7 quantifies what that is worth and why it is contingent.

3. Industry, Competitive Position & Moat

What the industry sells, and where the profit sits

Recorded-music streaming rents access to copyright. Global recorded-music trade revenue reached $31.7bn in 2025, up 6.4% and the eleventh consecutive year of growth, with streaming close to 70% of the total and 837m paid subscription accounts (IFPI Global Music Report 2026). MIDiA Research, using a broader definition, counts 921.6m subscribers at Q4 2025, up 10.1%.

The value chain runs artist and songwriter, to label and publisher, to platform, to listener. The decisive fact about it is not that the labels take a large share but that the share ratchets. The UK Competition and Markets Authority measured publishers' and songwriters' share of streaming revenue rising from 8% in 2008 to about 12% in 2012 and 15% by 2021, while record companies' share in 2021 sat “at a broadly similar level to what it was in 2008” (CMA, Music and Streaming Market Study Final Report, November 2022). The platform absorbs the increase.

The clearest way to see who captures the profit is to compare the wrong lines on purpose. Universal Music Group earned a 22.5% adjusted EBITDA margin on €12,507m of FY2025 revenue; Warner Music Group 21.5% adjusted OIBDA on $6,707m; Sony's Music segment a 21.1% operating margin. Spotify's record gross margin — before a euro of research, marketing or administration — was 33.4% in Q2 2026. A major label's EBITDA margin is two-thirds of the largest platform's gross margin. Deezer, the only listed pure-play comparator without scale, earned a 25.4% gross margin on €534m of FY2025 revenue and converted it into €8.5m of net profit — the first annual profit in the company's history, in a year its revenue fell 1.4% and its subscriber base declined.

Why the structure produces that outcome

• Non-substitutable input. The CMA found that streaming services “cannot meet consumers' needs without obtaining licences to the large catalogue of each major record label and publisher” and that “there is no credible alternative to each major's catalogue.” Four counterparties supply roughly 72% of Spotify's label-delivered streams.
• A one-sided walk-away option. Universal withdrew its entire catalogue from TikTok for three months in 2024. A peer-reviewed difference-in-differences study of 235,741 tracks found no significant impact on aggregate demand for Universal's music (May et al., Marketing Science, 2024); Universal returned in May 2024 with improved remuneration. No platform has ever run a mainstream music service without a major's catalogue.
• Cost of goods that does not scale. Royalties are a percentage of revenue with per-user minimums, so a platform with thirty times the subscribers pays roughly thirty times the royalties. The software operating-leverage curve applies to Spotify's research and marketing lines; it does not apply to two-thirds of its cost base.
• A price lever held by third parties. US publishing mechanicals are set under a compulsory licence by three federal judges, not negotiated — 15.1% of revenue in 2023 stepping to 15.35% by 2027 under Phonorecords IV (Federal Register, 30 December 2022).

The CMA examined precisely this question and concluded that it found “low or negative operating margins for the music streaming services whose accounts we have been able to analyse”, no evidence of “substantial and sustained excess profits by the majors”, and that the imbalance stems from “bargaining power of rightsholders inherent in the market which will not be overcome by more intense competition.” It then declined to intervene. A competition regulator has stated on the record that the platform's weak position is structural and permanent absent legislation.

Spotify's position inside that structure

Spotify holds 31.4% of global music subscribers, against Tencent Music 13.8%, Apple Music 12.6%, YouTube Music 12.4% and Amazon Music 8.5% (MIDiA Research, Q4 2025). The share data contains the most useful competitive evidence in this document: since 2020, Apple Music's share has fallen from 18.4% to 12.6% while YouTube Music's has risen from 7.9% to 12.4%. Because catalogue is effectively identical across every service by regulatory and commercial necessity, something other than content explains a six-point share swing at a company with hardware defaults, an identical catalogue and, until July 2026, a lower price. Product and discovery is the most plausible explanation, and Spotify is the beneficiary of Apple's loss.

Three of Spotify's four largest competitors do not run music as a business that must earn a return. Apple Music is an attach to hardware and to Apple One and has disclosed no subscriber number since 2019; Amazon Music is bundled into Prime; YouTube Music is sold inside YouTube Premium, and Alphabet has said it accepts advertising cannibalisation to grow subscriptions. In theory that should cap industry pricing. In practice it has not: Apple raised US prices to $10.99 in 2022 and $11.99 in July 2026, explicitly citing “rising licensing costs”; Amazon followed the same ladder. Spotify moved to $10.99 in July 2023, $11.99 in July 2024 and $12.99 effective February 2026, and currently carries a $1.00 premium to both. The bundled players behave as price followers, because their royalty base rises with the same rates and none is trying to win music on standalone economics. The competitive pressure they exert is on distribution, not on price.

What would be hardest for a well-funded competitor to reproduce is not the catalogue, which anyone can license, nor the technology. It is the installed base itself: 300m paying accounts and 777m monthly users generating 211bn hours of listening a year (FY2025), which produces the behavioural data behind recommendation, gives Spotify the volume to negotiate marketplace programmes with rightsholders, and supports a $1.00 price premium. Deezer is the control experiment for what the same business looks like without it.

What could weaken the position is not a competitor taking share. It is a licence renewal, a Copyright Royalty Board determination, or a court reinstating an App Store commission — none of which is decided by anything Spotify does to its product.

What kind of company wins here

Owning copyright wins; distributing it does not. On that test Spotify is the wrong kind of company — it owns almost no copyright and cannot buy its way to any at a plausible price. But it has escaped the fate the CMA described for the category: it earned €2,198m of operating income in FY2025 at a 32% gross margin, it is the only platform demonstrating unilateral pricing, and it runs roughly 800 basis points of gross margin above the only comparable pure-play. It is the best available seat on the wrong side of the table, and the distinction between those two statements is the whole investment question.

4. Growth Engine

Reported growth is organic growth

Acquisitions are not a factor and the memo says so explicitly rather than omitting the test. Cash spent on business combinations across FY2021–FY2025 totalled €430m; there were no acquisitions at all in FY2023 or FY2024, and FY2025's single unnamed transaction cost €9m against €17,186m of revenue. Reported revenue growth and organic revenue growth are the same number.

The decomposition that does matter is currency, and it is large enough to invert the headline. FY2025 revenue grew 10% as reported; at FY2024 exchange rates Premium revenue would have been about €502m higher and Ad-Supported about €83m higher, implying roughly 13% constant-currency growth (inferred). In H1 2026 the gap was wider still — 11.1% reported against approximately 14.4% at prior-year rates, a €280m drag (inferred from disclosed FX effects in the Q2 2026 6-K). By Q2 2026 the drag had almost gone: €30m, on 14% reported growth.

Source: FY2024 and FY2025 20-Fs; Q1 and Q2 2026 interim 6-Ks. Constant-currency revenue growth is not published by Spotify; the figures marked (inf.) are inferred from the euro FX effects the company does disclose for revenue and cost of revenue. ARPU bridges are as disclosed.

The drivers, ranked

• 1. Price increases — management-driven, and currently the entire ARPU story. Spotify has raised US prices in three consecutive cycles: $9.99 to $10.99 in July 2023, to $11.99 in July 2024, and to $12.99 effective February 2026, with parallel moves across dozens of markets. Price contributed +€0.49 of ARPU in Q2 2026, more than the entire +€0.32 net move. The filings never name a repriced market or a price point — that detail comes from the public record, not from disclosure.
• 2. Subscriber additions — structural, but decelerating. Premium subscribers reached 300m at 30 June 2026, up 24m year on year. The growth rate has stepped down for four consecutive quarters, from 12% to 10% to 9% to 9%, and Q1 2026 added only 3m net against 5m a year earlier. Monthly users grew faster, at 12%, with the Rest of World region up 21% in FY2025 — which is the mix problem in the next bullet.
• 3. Geographic mix — structural, and a permanent drag on ARPU. Product and market mix has subtracted €0.12 to €0.15 of ARPU in every recent period without exception. Europe is 26% of monthly users and grew 6%; Rest of World is 37% and grew 21% (FY2025 20-F). Growth is arriving where realised price is lowest. Tencent Music's monthly ARPPU of RMB 11.9, roughly $1.70, is what deep emerging-market monetisation looks like.
• 4. Currency — cyclical, and reversing. The euro reporting currency turned a strong constant-currency year into a modest reported one in FY2025 and the first half of 2026. It is not a business driver, but it has been large enough to make reported ARPU fall 1% in a year when price rose €0.25, and no reader should mistake one for the other.
• 5. Non-music revenue — management-driven, real, and small. Audiobooks, the Partner Program and the Spotify Ad Exchange all grow faster than music: US podcast advertising reached $2.9bn in 2025, up 17.6% (IAB/PwC), and US audiobook sales $2.43bn, up 9% (Audio Publishers Association), against recorded music's 6.4%. But each of those markets is roughly a quarter the size of US recorded music, and Spotify shares both. Spotify discloses no audiobook revenue at all.
• 6. Advertising — currently negative. Ad-Supported revenue fell 1% in FY2025 and 2% in H1 2026 while Ad-Supported users grew 14%. Direct music and podcast ad sales fell €116m on lower fixed CPM rates, offset by only €103m from automated channels (FY2025 20-F). Higher fill at lower realised price is not growth. The US industry corroborates it: ad-supported music revenue fell 0.6% in 2025 (RIAA).

5. Margin, Cash & Capital Allocation

How the royalty is calculated — the mechanism behind two-thirds of the cost base

Subscription music royalties are “based on the greater of a percentage of relevant revenue and a per user amount”, and the applicable percentage “is generally dependent upon certain targets being met” — the number of Premium subscribers, the ratio of ad-supported users to subscribers, and subscriber churn rates (FY2025 20-F). Three consequences follow, and they explain almost everything about the margin.

First, because the obligation is a percentage of revenue, gross margin cannot be improved by volume. Second, because there is a per-user floor beneath the percentage, low-price plans and low-ARPU markets are structurally dilutive — Spotify says it has “negotiated lower per user amounts for our lower priced subscription plans”, which concedes the floor binds. Third, because the percentage steps with subscriber and mix targets, scale does convert into a slightly better rate, which is the one place where size helps on the cost line. Most-favoured-nations clauses run the other way: a concession granted to one licensor can “cause our payments or other obligations under those agreements to escalate” across the others, so the marginal cost of settling any single negotiation is multiplied.

Against that, Spotify runs voluntary marketplace programmes under which artists, labels and distributors accept “a discounted royalty rate to streams generated in specified recommendation contexts” in exchange for algorithmic promotion. These are cited as a gross-margin driver in every filing from FY2021 to FY2025 and in every recent quarter. They are never quantified in any amount, in any filing. On the company's own account this is one of the two largest levers behind the margin expansion, and it is the single largest disclosure gap in the business.

What the margin has actually done

Consolidated gross margin went 27% in FY2021, 25% in FY2022, 26% in FY2023, 30% in FY2024, 32% in FY2025, and 33.4% in Q2 2026 — the highest figure in any period these filings cover. Expressed per unit, gross profit per Premium subscriber per month rose from €1.25 in FY2021 to €1.27 in FY2023, €1.56 in FY2025 and €1.71 in Q2 2026 (computed from disclosed ARPU and segment margins). Two-thirds of the improvement is price outrunning the royalty base; the remainder is podcast cost reduction and the audiobook bundle.

Below the gross line, the 2023 cost reset did the work. Spotify cut about 6% of staff in January 2023, realigned podcast operations in the second quarter, and cut a further 17% in December, taking €212m of severance and a further €123m of real-estate impairment. Average headcount fell from 9,123 in FY2023 to 7,287 in FY2025 while revenue rose 30%, lifting revenue per employee from €1.45m to €2.36m. Combined operating expenses fell from 29% of revenue in FY2023 to 19% in FY2025. No comparable restructuring has been taken since.

One large distortion has to be stripped out of any recent margin read. Spotify accrues Swedish and other payroll taxes on the intrinsic value of unvested equity awards and remeasures the accrual to its own share price each quarter. Share-based-compensation social costs were €291m in FY2024 and €125m in FY2025 — a €166m swing booked entirely inside R&D, sales and marketing, and administration, with nothing to do with operations. In Q2 2026 the year-on-year credit was €113m, or 45% of the €249m increase in operating income; for H1 2026, €227m of a €455m increase. Underlying Q2 2026 operating income grew roughly 34%, not the 61% reported. The accrual has drained to €142m and a 10% share-price move now shifts it by about €20m — so the credit is close to exhausted, and a share-price recovery puts the cost straight back into operating expenses.

The financial spine

Source: FY2021, FY2023 and FY2025 20-Fs; Q2 2026 interim 6-K. Years chosen to show the shape of the change — the pre-reset business, the loss-making trough, the reset business, and the current run rate — not every year available.

Three comparability breaks sit inside that table and must be carried by any reader of it. From FY2025, podcast costs attributable to the new Premium video experience moved from the Ad-Supported segment into Premium, and prior years were not restated; the amount is not disclosed. From 1 January 2026 a further transfer of revenue-generating activities moved from Ad-Supported to Premium, and this time prior periods were restated in the 6-K — so the segment splits printed in the FY2025 20-F are not comparable with those in the 2026 interims. From FY2023, audiobook content costs were allocated to Premium. Separately, the disclosed share of streams from the major labels changed denominator in FY2023, from “music streams” (77% in FY2021) to “streams of audio content delivered by record labels” (72% in FY2025); the apparent five-point decline is not like for like.

Cash conversion and capital intensity

Operating cash flow was €2,933m in FY2025 against €2,212m of net income, and €1,652m in H1 2026. Capital expenditure was €61m in FY2025 — 0.35% of revenue — because the compute sits in a Google Cloud service agreement inside purchase obligations rather than on the balance sheet. Free cash flow was €2,874m in FY2025 and €1,621m in H1 2026. The gap between cash flow and profit is the prepaid float described in Section 2, and it widens with every subscriber and every price increase.

Where the cash has gone

Ranked over FY2021–FY2025, the answer is that it mostly did not go anywhere. Cash and short-term investments rose from roughly €3.5bn to €9.5bn. Share repurchases totalled €530m across five years, of which €439m fell in FY2025 alone and nothing at all in FY2023 or FY2024. Employee tax withheld on share releases — economically a buyback — came to €538m. Acquisitions were €430m, lease payments €286m, capital expenditure €194m. No dividend has ever been declared. Against €6.3bn of operating cash flow, returns of roughly €1.1bn were themselves largely offset by €1.9bn of option-exercise proceeds coming in, and shares outstanding still rose 7.1% over the period.

That posture changed after the FY2025 20-F was filed, and the change is the most consequential thing in the recent record. The US$1.5bn zero-coupon Exchangeable Notes matured on 15 March 2026 and were settled in cash for €1,304m, leaving Spotify with no debt other than leases against €9,388m of cash and short-term investments at 30 June 2026. Repurchases ran €547m in H1 2026 against zero in H1 2025. The April 2026 annual meeting approved a fresh five-year authorisation for 10,000,000 shares — an item absent from the 2025 agenda — and on 20 August 2026 the board added a further US$1.5bn to the programme, taking cumulative authorisation to US$3.5bn against US$1,253m spent through June. Read together: a company that hoarded cash for five years began returning it in the first year after it turned structurally profitable, and accelerated twice within seven months. The intent is now buybacks, not a dividend.

Two items post-date the FY2025 20-F and belong here. The Tencent Music stake, carried in long-term investments, fell from €2,111m at 31 December 2025 to €1,034m at 30 June 2026 — a €1,077m fair-value loss routed through other comprehensive income, which is why total equity barely moved despite €1,266m of first-half earnings. And Daniel Ek ceased to be Chief Executive on 31 December 2025, becoming Executive Chairman; Alex Norström and Gustav Söderström became Co-Chief Executives on 1 January 2026. No transaction has been announced and not closed.

6. Cyclicality, Constraints & What to Monitor

What a downturn actually looks like here

This is not an economically cyclical business in the ordinary sense, and the evidence is the 2020–2025 record: monthly users rose from 345m to 751m and subscribers from 155m to 290m without a single down year, through a pandemic, an advertising recession and a rate shock. A €12.99 monthly subscription is small enough, and habitual enough, that discretionary-spending pressure shows up as trading down between plans rather than as cancellation. Spotify does not disclose a churn rate in any filing, so this is an inference from the subscriber series rather than a measured fact.

The exception is advertising, which is 11% of revenue and behaves like the rest of digital audio: US ad-supported music revenue fell 0.6% in 2025 (RIAA) even as podcast advertising grew 17.6%. An advertising downturn costs Spotify roughly a tenth of revenue and rather less than a tenth of gross profit, and it did not stop the margin rising in 2025.

The cycles that do matter are contractual and regulatory. Label licences run “frequently between one and three years”, are not automatically renewable, and set the rate on two-thirds of the cost base. US publishing rates are reset every five years by statute. Spotify's minimum guarantees stood at €2,613m at 31 December 2025, of which €1,123m falls due within a year, with a further €202m of new commitments signed after the year end. Those obligations are fixed while subscribers “may cancel their subscriptions at any time”, and the company states the guarantees are “not always tied to our revenue and/or user growth forecasts.” That is the one place a demand shortfall would transmit directly and immediately into gross margin.

Where the business sits right now, against its own history

• Margin: at its peak. 33.4% consolidated and 34.9% Premium in Q2 2026 are the highest figures in any period these filings cover, against a 25–27% band from FY2019 through FY2023.
• Price: at its peak, and accelerating. Premium ARPU of €4.89 in Q2 2026 is the highest quarterly figure disclosed, and +7% year on year is the fastest — against –4% in Q3 2025 and –1% for FY2025 as a whole.
• Volume: past its peak growth rate. Subscriber growth has decelerated for four straight quarters to +9%, and Q1 2026's +3m net additions were the weakest quarter in the observable series.
• Profit: flattered. Roughly 45% of the Q2 2026 increase in operating income was the share-price-driven social-cost credit. The comparison of net income is worse: Q2 2025's loss reflected €421m of Exchangeable Note fair-value losses that cannot recur, the instrument having matured.

The downside mechanism, stated as a mechanism

The Copyright Royalty Board opened the Phonorecords V proceeding in December 2025 to set US publishing mechanical rates for 2028–2032. A partial settlement published in the Federal Register on 10 July 2026 covers physical, downloads and ringtones only; the streaming rate is unsettled and objections were filed in August 2026. The historical ratchet is 8% in 2008, roughly 12% in 2012, 15.35% by 2027. A move to 17% would be about 165 basis points of additional cost of revenue with no offsetting benefit and no ability to negotiate, since this is a compulsory licence — roughly a 5% reduction in gross profit, permanent, effective 1 January 2028. The same proceeding is where the “bundle” definition that underpins Spotify's audiobook treatment can be redrafted. The downside is therefore correlated, exogenous and calendared.

Durable versus borrowed

Leading indicators, and where each is published

7. Risks, Unknowns & Questions for Deeper Work

Cyclicality and the statutory rate cycle are covered in Section 6 and are not restated here. What follows is what cyclicality does not capture, ranked by whether the risks compound.

• Renewal asymmetry, amplified by most-favoured-nations clauses. Four counterparties supply roughly 72% of label-delivered streams under multi-year agreements that are not automatically renewable and are terminable on non-payment, material breach, or a change of control. Every 100 basis points of Premium royalty rate is about €154m of Premium gross profit at FY2025 volumes (inferred), so a rate outcome 300 basis points worse than the current one removes roughly €460m — more than a fifth of FY2025 operating income. Most-favoured-nations provisions mean any concession granted to settle one negotiation ratchets across the others, so the cost of a single bad renewal is multiplied rather than isolated. The FY2025 filing adds a phrase absent from FY2021: rightsholders “could impede our business by withholding content, discounts and bundle approvals, and the rights to launch new service offerings.” Spotify's product roadmap requires supplier consent.
• The per-user floor collides with emerging-market growth. Royalties are the greater of a revenue percentage and a per-subscriber amount. As growth migrates to Rest of World — 37% of monthly users, growing 21% — blended ARPU falls while the floor does not, so the effective royalty rate rises above the headline percentage. This is the mechanism by which the company's fastest-growing markets can be revenue-accretive and margin-dilutive at the same time, and it compounds with any statutory rate increase rather than offsetting it.
• Two margin wins came from courts and regulators, and can be withdrawn by them. The audiobook bundle reclassification survived only because a federal judge read the Phonorecords IV regulation as unambiguous; the MLC filed an amended complaint on 1 October 2025 and is seeking interlocutory appeal, and Spotify's own disclosed exposure for March 2024 to June 2026 has grown from €358m to €473m, unprovided, accreting roughly €57m a quarter. Separately, the collapse of Apple's App Store commission on external link-outs — from 30% to zero in the US and to 15% in the EU from October 2026 — is worth more on iOS-acquired subscribers than Spotify's entire gross margin on them, and the US rate is provisional pending the district court's determination of a “reasonable” cost-based commission. Neither win was earned by building anything, and neither is settled.
• The advertising segment is structurally deteriorating, not cyclically weak. Revenue fell 1% in FY2025 and 2% in H1 2026 while the audience grew 14%. The stated cause is a shift from fixed-CPM direct sales to biddable inventory at lower realised prices. At 11% of revenue and 6% of gross profit, and almost certainly loss-making after any reasonable allocation of the €3,298m operating-expense base (inferred), the segment's value is as the conversion funnel rather than as a business — and Section 2 shows conversion is flat.
• Segment reporting has been recut twice in two years. Podcast costs moved from Ad-Supported to Premium in FY2025 without restatement, and further activities moved on 1 January 2026 with restatement. Neither amount is quantified. The FY2025 Ad-Supported margin improvement from 12% to 18% and the Premium improvement from 33% to 34% are therefore not clean comparatives, and an analyst building a segment margin series across 2024, 2025 and 2026 is drawing a line across two redefinitions.
• Reported book equity depends materially on one Chinese-listed share price. The Tencent Music holding fell €1,077m in six months through other comprehensive income, absorbing most of €1,266m of first-half earnings. This does not touch the income statement or cash, but it means equity is a poor guide to the operating business.
• Governance and an untested executive structure. The founders hold 309,932,980 beneficiary certificates, so voting control is unaffected by any buyback or issuance. The co-Chief Executive structure took effect on 1 January 2026; director remuneration has drawn the largest dissenting vote at each of the last two annual meetings; the Chief Accounting Officer changed in May 2026 and a nine-year director resigned on 3 September 2026.

What the sources could not answer

Each of these is a finding rather than a gap in the research — in most cases Spotify has chosen not to disclose it.

• Churn. No churn rate appears in any filing, in any year, despite churn being an input to the royalty rate formulas Spotify describes. Every statement about retention in this document is inferred from subscriber counts.
• The size of the marketplace programmes. Cited as a leading gross-margin driver in every year since FY2021, never quantified in any amount.
• The composition of cost of revenue. Only year-on-year movements are given — absolute royalty expense, payment processing fees and streaming delivery costs are nowhere disclosed.
• Revenue by region, and subscribers by region. Revenue is reported only as United States, Luxembourg and “other countries”; regions appear only as a percentage of monthly users.
• Audiobook economics. No revenue, no cost, no attach rate, four years after entry.
• Which markets were repriced, when, and by how much. The filings confirm price increases only through the ARPU bridge; the price ladder in this document comes from the public record.
• The ad-supported to Premium conversion rate, described qualitatively as “a significant portion” of gross additions and never measured.
• Segment operating income, segment assets and segment capital expenditure — not disclosed, so the standalone economics of the advertising business cannot be established from the filings.
• Price elasticity above roughly $13. No rigorous independent study of music-streaming churn response to price was found; the low-elasticity conclusion in this document is inferred from aggregate industry data through the $10–$12 band and does not extend beyond it.
• Competitor economics. Apple, Amazon and Alphabet disclose no music subscriber count or segment financials; all competitor shares here are third-party estimates.

What would need resolving before forming a thesis: the terms on which the current label agreements were last renewed and when each expires; the euro value of the marketplace-programme discount; and a defensible view of the Phonorecords V streaming rate, since that single number moves gross margin more than any operating decision management can make.

8. Investor Takeaways

• What this business really is. A rental counter for other people's copyright, with the best distribution position in its industry and no ownership of the thing it sells.
• The core economic engine. 300m prepaid monthly accounts at €4.89 each, of which €1.71 is kept as gross profit, collected before the royalty is paid — a float that funds growth with capital expenditure of 0.35% of revenue.
• The main growth lever. Price. Subscriber growth has slowed for four consecutive quarters while price contributed the entire +€0.32 of the Q2 2026 ARPU move; the durable question is how much headroom remains above $12.99.
• What could break the story. A label renewal or a Copyright Royalty Board determination that moves the royalty rate. Roughly €154m of gross profit turns on every 100 basis points, and Spotify negotiates neither number from a position of strength.
• What to monitor. The quarterly ARPU price-versus-mix bridge, Premium gross margin ex the social-cost credit, and the Phonorecords V streaming rate for 2028–2032.