The receipts
Five years of the major music companies' own moves, all on the public record. We just put it in order.
First thing you should know: the major music companies are not failing. They posted record revenues while everything in this file happened.
In the same five years, they sued a two-year-old startup for half a billion dollars and licensed that same company sixteen months later. Warner cut ten percent of its staff the same day it announced its best quarter ever. UMG spent $775 million buying the pipes independent artists use to route around the majors.
Everything below is sourced — court filings, investor documents, earnings calls, press releases. Dates, dollars, names, with links. Click anything and check it.
The exhibits
25 exhibits · 8 categories · every source linked
Litigation as defense
They sued AI to price it, not to stop it.
EXHIBIT A-1The $500M twin lawsuits: RIAA v. Suno, RIAA v. Udio
What happened
All three majors filed nearly identical copyright infringement suits against the two leading AI music generators on the same day, coordinated through the RIAA. The complaints alleged Suno and Udio trained on massive amounts of copyrighted recordings without authorization. The evidence was vivid by design: tracks nearly identical to “My Girl,” “American Idiot,” “All I Want for Christmas Is You,” and 29 outputs resembling “Johnny B. Goode”; outputs carrying authentic producer tags; generated vocals indistinguishable from Bruce Springsteen, Michael Jackson, Lin-Manuel Miranda. RIAA chief Mitch Glazier framed it as defending “responsible AI,” saying the industry was “already partnering and collaborating with responsible developers” — read that sentence again, because it's the whole strategy in one line: the lawsuit was never against AI, it was against unlicensed AI.
The numbers
Damages sought of up to $150,000 per work plus $2,500 per act of DMCA circumvention — headline exposure of ~$500M, with the initial Suno complaint citing 560 recordings (~$84M at the statutory cap). Then it got bigger: in May 2026, using Audible Magic fingerprinting, the labels moved to add 61,026 additional recordings to the Suno case — pushing theoretical maximum exposure past $9 billion. Suno had raised $125M the month before the suit. Udio had raised $10M.
What changed after
Suno conceded in an August 2024 filing that its training data “includes essentially all music files of reasonable quality that are accessible on the open internet,” and argued fair use. That admission reframed the entire fight: the question was no longer whether they trained on the majors' catalogs, but whether they'd pay for it. Which is a licensing negotiation conducted in a courtroom.
THE TELL — You don't file the largest coordinated copyright action in music history against startups a fraction of your size to make them disappear. You file it to establish that the right exists, that you own it, and that the price is whatever keeps them at the table. Glazier said it in the launch statement: we're already partnering with responsible developers. The suit defined “responsible” as “paying us.”
Suno complaint (RIAA PDF) ↗TIME coverage ↗MBW: the 61,026-recordings motion ↗404 Media: Suno's admission ↗
Sources: RIAA press release (riaa.com, June 24, 2024); Bloomberg (June 24, 2024); TIME; LA Times; Music Business Worldwide (Audible Magic motion, May 2026); Suno's August 2024 answer (D. Mass. docket).
EXHIBIT A-2The reversal: sue-then-license
What happened
Roughly sixteen months after calling Suno and Udio thieves in federal court, the majors started signing with them. UMG settled with Udio first, pairing the settlement with recorded-music AND publishing licenses for a new licensed platform launching 2026. Warner followed with Udio three weeks later, then stunned the industry by settling with Suno during Thanksgiving week — a company Warner's own complaint had accused of mass infringement. Suno dropped its fair-use defense, agreed to deprecate every model trained on unlicensed music, and committed to a licensed relaunch in 2026. As part of the deal, Warner sold Songkick — its live-music discovery platform — to Suno. The cascade continued: Merlin (the independent label coalition) signed with Udio in December 2025, Kobalt in January 2026.
The numbers
Financial terms of every settlement: undisclosed. What we can see: Suno raised $250M at a $2.45B valuation in November 2025 — the same month it settled with Warner — and later raised $400M at $5.4B. Songkick's books (UK Companies House, via MBW): £4.5M turnover, zero pre-tax profit in FY2024. The structural terms we know: Udio's UMG deal turned it into a “walled garden” — users can no longer download or export anything; creations live and die on-platform, where UMG controls distribution and collects on every play. Warner's Suno deal was looser: downloads survive for paid users, subject to a monthly cap.
What changed after
The AI music companies the majors sued became more valuable because the majors sued them — the settlements functioned as licensing certifications that unlocked billions in venture funding. UMG is now reportedly trying to subpoena the terms of the Warner–Suno deal (Digital Music News, April 2026) — one major using discovery to find out what price another major set. Sony remains the lone holdout, still litigating both cases; the Suno case headed to a July 2026 summary-judgment hearing.
THE TELL — This is the single most important pattern in the entire file: they sued to establish the right, settled to price it, and licensed to collect on it. Their own signatures prove they were never against AI platforms — they were against being cut out. Which means for a platform that invites them in on day one, the door isn't just open. They built it themselves.
UMG–Udio press release ↗MBW: UMG settles with Udio ↗MBW: Warner–Suno deal ↗
Sources: Billboard Pro (Suno–Warner deal; “What Do the Suno and Udio Licensing Deals Mean,” March 2026); Rolling Stone (Nov 26, 2025); Music Business Worldwide (May 2026); Chartlex AI Lawsuits Tracker (docket-verified, July 11, 2026); UMG/Warner press releases; PRNewswire.
EXHIBIT A-3Publishers v. Anthropic: the lyrics front
What happened
The publishing arms opened a second front: lyrics. The original 2023 suit (~500 songs) alleged Claude reproduced copyrighted lyrics in outputs. Through discovery in the Bartz v. Anthropic authors' case — which settled for $1.5 billion (~$3,000 per book, ~500,000 works, Anthropic agreeing to destroy the pirated files) — the publishers say they found Anthropic had downloaded 20,000+ songs from pirate shadow libraries. When the court denied their motion to fold that into the original case (October 2025), they filed a brand-new suit in January 2026 seeking more than $3 billion — described as potentially the largest non-class-action copyright case in US history — and named CEO Dario Amodei and co-founder Benjamin Mann personally. Anthropic added trial teams from Morrison Foerster and WilmerHale in July 2025; one industry blog called the staffing move “the legal equivalent of a code red.” BMG filed its own suit over 493 compositions in March 2026.
The numbers
$3B+ sought (Concord II). Statutory ceiling of $150,000 per work for willful infringement. 20,000+ works alleged. The Bartz comparison the publishers are anchoring to: $1.5B.
What changed after
Second amended complaint filed July 2026. Anthropic moved to dismiss the output and DMCA claims and Amodei's personal count. The strategic shape is identical to A-1: establish that AI training on lyrics requires a license, then sell the license. The publishers watched the authors get $1.5B and immediately re-filed bigger.
THE TELL — Naming individual executives personally isn't how you behave when you want a company destroyed — it's how you maximize settlement pressure. The publishing arms are running the same sue-to-price playbook the recorded-music side ran on Suno and Udio, one asset class over.
TechCrunch: the $3B suit ↗IPWatchdog analysis ↗MBW: amended suit + Bartz ↗
Sources: TechCrunch (Jan 29, 2026); Music Business Worldwide (Jan 28, 2026; July 24, 2025); Shopifreaks; Nine North Label Group weekly (July 24, 2026).
EXHIBIT A-4GEMA opens the European front
What happened
GEMA sued OpenAI over ChatGPT reproducing lyrics, then sued Suno in Munich Regional Court over compositions including “Forever Young,” “Mambo No. 5,” “Cheri Cheri Lady,” and “Daddy Cool.” Critically, GEMA didn't just sue — it simultaneously launched the first AI licensing model for music (2024). Sue with one hand, hold out the rate card with the other. On July 31, 2026, the Munich court ruled for GEMA against Suno: cease-and-desist, disclosure obligations, damages — with the EU AI Act's copyright-transparency duties cited in the ruling.
The numbers
Damages to be assessed; the precedent is the prize. EU AI Act enforcement begins August 2026.
What changed after
The first merits ruling anywhere in the world on AI music training went against the AI company — in Europe, before any US court reached the question. Every AI music company now needs a European licensing answer, and GEMA already printed the menu.
THE TELL — The rights-holder side didn't wait for US courts. They opened a second jurisdiction where the law was friendlier and built the licensing product before winning the case — a product launch that used a lawsuit as its marketing.
GEMA: the AI lawsuits ↗Music Ally: the ruling ↗JUVE Patent case report ↗
Sources: GEMA press releases (Jan 21, 2025); AI Musicpreneur GEMA–Suno case tracker; Chartlex tracker (July 2026 ruling); financialcontent/PRLeap.
EXHIBIT A-5The AFM suit: their own musicians want to see the AI money
What happened
The musicians' union sued two of the three majors, alleging they licensed AFM-member session recordings to Suno and Udio for AI training “without compensation or credit” — triggering the “new use” clause of the collective bargaining agreement — and then refused to disclose which recordings were licensed when the union asked.
The numbers
Scope tied to decades of session work across both catalogs; damages to be determined.
What changed after
The settlements from Exhibit A-2 became a liability inside the building within months. The money the majors extracted from AI companies is now the subject of litigation from the people who played on the records.
THE TELL — The AI licensing revenue line materialized so fast that the majors' own paperwork couldn't keep up. You only get sued by your own musicians over AI proceeds if the AI proceeds are real, immediate, and being booked. This exhibit is third-party proof that Exhibit A-2 is a revenue strategy, not a legal defense.
MBW: the AFM suit ↗PYMNTS report ↗
Sources: Chartlex AI Lawsuits Tracker (docket-verified); trade press coverage June–July 2026.
EXHIBIT A-6UMG's own legal weather: Drake, antitrust, the royalty class
What happened
UMG's biggest artist of the streaming era sued his own label for defamation over its promotion of Kendrick Lamar's “Not Like Us” — while his contract was up for renegotiation, per his own complaint. Judge Vargas dismissed it as “non-actionable opinion”; Drake appealed. Meanwhile: a late-2025 ruling allowed a nationwide artist royalty class action against UMG to proceed — opening the door to potentially tens of thousands of artists who received UMG royalty payments in the past six years. And the DOJ's antitrust interest in UMG (~31% global recorded-music share) remains at the investigation stage, with private plaintiffs already filed, raising exactly the theories in this file: exclusive platform deals that disadvantage independents, distribution structures that lock playlist access, and acquisitions “designed to absorb and neutralize competitors rather than compete with them.”
The numbers
UMG global share ~31%. Royalty class: potentially tens of thousands of artists, six-year lookback.
What changed after
The artist-label bargain is being contested at every level simultaneously — superstar (Drake), rank-and-file (royalty class), session player (AFM), and regulator (DOJ). No confident incumbent fights on four internal fronts at once.
THE TELL — When your flagship artist litigates against you during a contract renegotiation, the leverage has moved. Artists can now credibly threaten to walk into an independent infrastructure that didn't exist at scale ten years ago — which is exactly why UMG spent $775M buying that infrastructure.
ABC News: dismissal ↗Sidley: dismissal announcement ↗MBW: the appeal ↗
Sources: Lawfold UMG litigation overview (April 2026); S.D.N.Y. docket (Drake dismissal, Oct 9, 2025); trade coverage of the class certification and DOJ interest.
Panic restructuring
Record profits and thousands of layoffs, announced in the same press release.
EXHIBIT B-1UMG's “strategic organizational redesign”
What happened
UMG top brass told investors in October 2023 they were preparing a “cut to grow” strategy. On February 28, 2024, UMG confirmed a cost-saving “redesign” including headcount reduction. Layoffs began within 48 hours of the earnings call, reaching into the high hundreds, concentrated in promotion, publicity, A&R, and catalog — the actual artist-facing machinery. Confirmed departures included Interscope promotion legend Brenda Romano, Interscope head of publicity Cara Donatto, Def Jam EVPs, Capitol's EVP of global marketing. Billboard confirmed ~50 named layoffs across Interscope, Republic, Capitol, Def Jam, Island, UMe, and corporate — with the true total far higher.
The numbers
€250M ($270M) in annual run-rate savings by 2026, phased €75M (2024) → €125M (2025) → €250M (2026), “all of which is accretive to EBITDA.” Restructuring charges: €125M expected in 2024, ~€100M in Q1 alone. Baseline headcount: 9,992 at end of 2022 — a base that had GROWN by 487 in 2022 and 322 in 2021. They hired into the streaming boom and cut the moment it plateaued.
What changed after
2024 full-year results showed the redesign flowing straight to the bottom line: $12.8B total revenue (+6.5%), adjusted EBITDA +13.8% — profit growth outrunning revenue growth because the cost base was gutted. And buried in the same results: ad-supported streaming revenue DOWN 5.1% year-over-year.
THE TELL — If they believed in the model, a record year would have gone back into A&R and promotion. Instead they cut promotion, publicity, and catalog staff — the machinery of the old business — and moved the money to “Streaming 2.0” and superfan products that don't exist yet. The redesign is a confession written in severance checks: the streaming-era org chart was built for a growth curve that ended.
MBW: the $270M redesign ↗Billboard: the layoffs ↗Variety: named departures ↗MBW: 2024 results ↗
Sources: MBW (Jan 16 and Feb 28, 2024); Billboard (Feb 28 and Mar 1, 2024); Variety (Mar 4, 2024; Mar 6, 2025); Digital Music News.
EXHIBIT B-2The collapse of the legacy label system
What happened
Grainge's February 1 memo reorganized US labels into an East Coast–West Coast structure: Republic's Monte Lipman took Republic, Def Jam, Island, and Mercury; Interscope's John Janick took Interscope, Geffen, Capitol, Motown, Priority, Verve, and Blue Note. Days later, Capitol chair-CEO Michelle Jubelirer stepped down — and in one of the most brutal juxtapositions in modern industry history, she was named Billboard's Women in Music Top Executive of 2024 the same morning UMG confirmed the cuts. Then the UK: Island and EMI merged, Polydor absorbed Capitol UK, with UK chief David Joseph writing internally that “developing artists now requires more creativity and patience than ever before” — a sentence that admits the old artist-development economics stopped working.
The numbers
Seven historic frontline labels folded under two US executives. Two “powerhouse” groups replacing the UK label system.
What changed after
Def Jam — the most important hip-hop label in history. Island — Bob Marley's label. Capitol — the tower on Vine. Motown. Blue Note. All still exist as brands; none still exist as autonomous companies. Unlike normal consolidations where the winning division absorbs the losers' staff, ALL of UMG's labels lost staff.
THE TELL — These aren't operational tweaks — this is the demolition of the label system as an organizational form. They kept the logos because catalog marketing needs them, and deleted the companies because the margin couldn't carry them. If the frontline-label model still worked, Capitol would still be a company and its CEO wouldn't have been laid off on the morning of her own award.
MBW: the Grainge memo ↗Billboard: the restructure ↗Variety: Jubelirer steps down ↗MBW: Island–EMI merger ↗
Sources: Billboard (Feb 28, 2024); Variety (Mar 4 and Jul 9, 2024); MBW.
EXHIBIT B-3Warner: record quarter, 600 pink slips, same day
What happened
March 2023: Warner cut 4% of global staff (~270 of ~6,200 roles). Then February 7, 2024: Kyncl announced ~600 roles cut — roughly 10% of the company — on the SAME DAY Warner reported net income of $193M for the quarter, up from $124M the prior year. The cuts gutted Warner's owned media brands (Uproxx, HipHopDX, IMGN, Interval Presents — acquired only years earlier). Kyncl's framing: the cuts came “from a position of strength.” In 2025, Warner announced a further restructuring targeting ~$170M in additional cuts — announced alongside a $1.2B catalog joint venture with Bain Capital.
The numbers
2023: ~270 roles (4%). 2024: ~600 roles (10%), ~$200M annual savings. 2025: ~$170M more. Cumulative: roughly $370M/year in permanent cost extraction from a company that kept calling itself strong.
What changed after
Warner exited the media-brand strategy entirely, concentrated spending on catalog (the Bain JV), and — 21 months later — became the first major to license BOTH Suno and Udio. The company that cut the most, partnered with AI the fastest. That correlation is the strategy.
THE TELL — “Best quarter in our history” and “we're cutting 10% of you” in the same news cycle means leadership does not believe the quarter is durable. Companies signal their real forecast with headcount, not press releases — and three rounds of cuts in three years is a forecast.
MBW: the 2023 4% cut ↗MBW: 600 roles cut ↗MBW: the 2025 round ↗Bain Capital: the $1.2B JV ↗
Sources: MBW (Feb 2024; 2023 coverage); Billboard; Variety; trade coverage of the 2025 restructuring and Bain JV.
EXHIBIT B-4The industry-wide cut (nobody was confident)
What happened
The cuts weren't a UMG story or a Warner story — they were an industry posture. SoundCloud cut 8% of staff. BMG cut 3–5%. Every major music employer trimmed within the same 18-month window that streaming growth decelerated.
The numbers
SoundCloud 8%; BMG 3–5%; Warner 4% then 10%; UMG hundreds-to-thousands against a €250M target.
THE TELL — One company cutting staff is a management story. When every company cuts at the same time, the entire industry has read the same data — streaming's flattening curve — and reached the same conclusion together.
MBW: the industry roundup ↗MBW: SoundCloud cuts 8% ↗
Sources: MBW (Jan 16, 2024, comparative layoff roundup).
Leverage plays and platform standoffs
When you start taking hostages, you've admitted you're losing the negotiation.
EXHIBIT C-1The TikTok blackout
What happened
UMG let its TikTok license expire and published “An Open Letter to the Artist and Songwriter Community” — accusing TikTok of trying to “bully” it into a deal “worth less than the previous” one, and citing artist pay (revealing TikTok accounted for “only about 1% of our total revenue”), AI, and platform safety. UMG pulled roughly 4 million songs. Taylor Swift, Bad Bunny, Drake, The Weeknd — silenced on the biggest music-discovery engine on earth. Then the escalation: the publishing takedown muted songs by artists NOT signed to UMG's labels, if any writer in the credits was a UMPG writer. TikTok's counter-statement accused UMG of putting “greed above the interests of artists and songwriters.” Notably, Warner did NOT join. The standoff ended May 1, 2024 with a new deal touting “improved remuneration” and AI protections.
The numbers
~4M songs pulled. ~60% of TikTok videos use music (CNBC). TikTok = ~1% of UMG revenue — the number UMG itself disclosed to justify the fight, which simultaneously proved how little leverage the catalog gave them: the platform survived three months without the biggest catalog in music, and user growth didn't blink.
What changed after
UMG got a better rate and AI language — and litigation over back royalties from the blackout period continues. But the deeper result: the industry watched its most powerful company go nuclear on a platform and discover the platform could live without it. The era when catalog was unconditional leverage ended in public, between January and May of 2024.
THE TELL — Pulling your entire catalog is the last card in the deck, and you only play it when the normal cards stopped working. And when your open letter has to lead with “they only pay us 1% of revenue,” you've told the world the discovery layer of the music business now runs through platforms you don't control and can't out-wait.
UMG: official announcement ↗TechCrunch: the pull ↗Exclaim: the resolution ↗
Sources: UMG Open Letter (Jan 30, 2024); eMarketer; EPGD Law case summary; Exclaim (May 1, 2024); Lawfold (ongoing back-royalty litigation).
EXHIBIT C-2The Spotify bundling fight: the platform repriced them unilaterally
What happened
Spotify added 15 hours of audiobooks to Premium and reclassified the whole subscription as a “bundle” under the 2022 Phonorecords IV settlement — a paperwork move that mechanically reduced US songwriter/publisher royalties. The MLC sued in May 2024; the case was dismissed in January 2025, then reinstated on appeal in September 2025. At the NMPA annual meeting in June 2026, COO Danielle Aguirre put the bill on the table: “by Spotify's own admission,” bundling had cost publishers roughly $480M since 2024, including a $230M first-year loss.
The numbers
~$480M publisher losses since 2024. ~$230M in year one. One platform, one reclassification, no negotiation.
What changed after
The publishers are still fighting it in court and at the Copyright Royalty Board — but the money is gone and the precedent is set: a streaming platform can restructure the industry's royalty economics with a product bundling decision the industry finds out about after the fact.
THE TELL — For twenty years the labels set terms and platforms complied. Spotify's bundling move is the moment the direction of power reversed in writing: the platform repriced the rights holders unilaterally, and the rights holders' only recourse was litigation. That's what dependency looks like on a balance sheet.
Billboard: MLC sues ↗MBW: dismissal ↗MBW: case revived ↗NMPA: the $480M figure ↗
Sources: NMPA annual meeting remarks (June 2026); MLC v. Spotify docket; trade coverage.
Defensive acquisitions
They're buying the exits, not growth.
EXHIBIT D-1UMG buys Downtown: $775M for the independent escape hatch
What happened
UMG agreed to pay $775M for Downtown — which is not a label. It's the infrastructure of independence: CD Baby (DIY distribution used by millions of self-releasing artists), FUGA (the B2B distribution backbone behind hundreds of independent labels), Songtrust (global publishing administration), and Curve (the royalty-accounting platform independent labels run their books on). The independent sector went to war: over 200 independent executives signed opposition, IMPALA urged the EU to block the deal outright. The European Commission opened a full Phase II investigation and issued a Statement of Objections over a specific, devastating concern: through Curve, UMG would gain access to rivals' commercially sensitive data — the royalty books of the very independents competing with it. The EU cleared the deal on February 13, 2026 ONLY after UMG agreed to fully divest Curve.
The numbers
$775M. Part of a 2024 in which UMG's total investment spending exceeded $1 billion.
What changed after
The world's largest label now owns the two biggest routes that artists use to NOT need the world's largest label. And the EU put on the public record that regulators believe UMG buying indie infrastructure creates a surveillance problem, not just a market-share problem.
THE TELL — You don't spend three-quarters of a billion dollars buying CD Baby because DIY distribution is your future business model. You buy it because DIY distribution is your competitors' future business model — 8.2 million self-releasing artists' worth of it — and owning the pipe means the escape route pays you a toll. The EU's forced Curve divestiture is the receipt: even the regulators read the deal as absorb-and-monitor, not compete.
EC: clearance decision ↗EC: Statement of Objections ↗MBW: EU clears the deal ↗MBW: IMPALA's opposition ↗
Sources: Variety (Mar 6, 2025); MBW; ALERA analysis; European Commission Phase II filings and Feb 13, 2026 conditional clearance; IMPALA statements.
EXHIBIT D-2UMG's global land grab
What happened
In parallel with Downtown, UMG completed its acquisition of [PIAS] — one of Europe's last great independent label groups — took a majority of Nigeria's Mavin Global (Rema, Ayra Starr; the crown jewel of Afrobeats, estimated ~$150–200M), took full ownership of Thailand's RS Group music assets (~$65–70M), invested in Outdustry (China/India artist services), put $240M into the Chord Music catalog vehicle, and took stakes in NTWRK/Complex.
The numbers
$1B+ total 2024 investment spend, per UMG's own earnings call.
What changed after
The highest-growth music markets on earth (Afrobeats, Southeast Asia, South Asia) and Europe's flagship independent now route through UMG. The “independent alternative” in multiple regions is now a UMG subsidiary.
THE TELL — Every one of these deals buys something UMG couldn't build: authentic independence, regional credibility, and growth curves steeper than its own. When your organic growth engine slows, you buy other people's — and when the DOJ's stated concern is “acquisitions designed to absorb and neutralize competitors rather than compete with them,” this is the shopping list they're describing.
MBW: Mavin majority stake ↗MBW: [PIAS] completed ↗Chord press release ↗MBW: the $1B+ spend ↗
Sources: MBW (“Universal spent $1bn+ on investments in 2024”); Variety (Mar 6, 2025); Billboard; Lawfold.
EXHIBIT D-3Sony's multi-billion-dollar bet on the past
What happened
Sony went on the biggest catalog-buying run in history: Bob Dylan's recordings ($150–200M est., 2021); Bruce Springsteen's catalog (~$500M+, December 2021 — then the largest single-artist transaction ever); a 50% stake in the Michael Jackson catalog (reported February 2024, valuing it at $1.2B+); Queen (~£1B / $1.27B, June 2024 — the largest catalog deal ever, full stop); Pink Floyd's recorded music plus name-and-likeness rights (~$400M, October 2024). Stringer, 2021: “Anywhere where we think there is a deal that enhances our ecosystem, we will be there.”
The numbers
North of $3.5B on five artists, all of whose defining work is 30–60 years old.
What changed after
Sony became the only major to GROW market share in 2024 (to 21.7%, per MIDiA, while UMG lost ~100 basis points) — a share gain built substantially on owning more of the past. Sony is also the lone AI holdout and sent the 700-letter opt-out barrage: the purest “own the scarce assets, litigate the infinite ones” strategy of the three.
THE TELL — Billions deployed into proven, dead-or-legacy catalog is the asset allocation of a company hedging the future, not backing it. When your biggest investments are in music recorded before your employees were born, you've told the market where you think the reliable value is — and it isn't in the new-artist pipeline the old model was built to run.
MBW: Dylan recordings ↗MBW: Springsteen $500M+ ↗MBW: Queen $1.27B ↗MBW: Pink Floyd $400M ↗
Sources: LA Times (Springsteen, Dec 2021); trade reporting on Jackson (Feb 2024), Queen (Jun 2024), Pink Floyd (Oct 2024), Dylan (2021); MIDiA market shares (Mar 13, 2025).
EXHIBIT D-4The Hipgnosis collapse: the bubble called its own top
What happened
The publicly-listed fund that turbocharged the catalog gold rush imploded: shareholders voted no confidence, new independent valuer Shot Tower Capital slashed the portfolio's value by ~26%, NAV write-downs cascaded, and Blackstone took the whole thing private for ~$1.6B ($1.30/share) — a fraction of the ambitions the fund IPO'd with.
The numbers
~26% valuation haircut. ~$1.6B exit.
THE TELL — The catalog-buying wave the majors rode has a ceiling, and the public markets found it first. “Own the past” is a hedge, not a future — and the one pure-play vehicle for that thesis got repriced 26% downward and sold for parts. If catalog were the answer, Hipgnosis would be the biggest company in music. It's a Blackstone line item.
MBW: the Shot Tower cut ↗MusicTech: Blackstone $1.6B ↗MBW: done deal ↗
Sources: The AIC (Shot Tower 26% cut); Reuters (Blackstone $1.6B bid); trade coverage of the shareholder revolt.
The AI pivot
From “AI is theft” to a licensed AI revenue line in under two years.
EXHIBIT E-1UMG's AI deal sheet (while suing over AI)
What happened
Count the partnerships: BandLab (2023, “ethical AI” alliance). Roland — co-published the “Principles for Music Creation with AI” (2024). SoundLabs (2024, voice-model tools). ProRata.ai (attribution tech). Endel (functional soundscapes). NVIDIA and KDDI (infrastructure). Then the post-settlement sprint: Udio licensing deal (October 29, 2025), Stability AI partnership (October 31, 2025 — TWO DAYS LATER), KLAY Vision (November 20, 2025), Splice (December 2025). UMG also partnered with SoundPatrol to detect unauthorized AI music — building the radar while selling the planes.
The numbers
Terms undisclosed across the board. Cadence is the data: three major AI deals in 23 days (Oct 29 – Nov 20, 2025).
What changed after
UMG went from plaintiff to platform-partner across every layer of the AI stack — generation, tools, attribution, detection, standards — inside 30 months.
THE TELL — No company builds a twelve-partner AI portfolio to stop AI. UMG built the full stack of a company planning to operate in AI music — while its lawyers argued in court that AI music was theft. The litigation set the price and the partnerships collect it; it was always one strategy.
MBW: Udio settlement ↗Stability AI: the alliance ↗UMG–Stability press release ↗
Sources: Variety (Mar 6, 2025); Hollywood Reporter (KLAY deal); UMG press releases (Oct 29, Oct 31, Nov 20, 2025); MBW.
EXHIBIT E-2Sony's 700 letters: opt-out as rate card
What happened
Sony mass-mailed 700+ companies declaring it was “opting out” of AI training on its content, warning recipients they “may already have made unauthorized uses” — while stating in the same breath that Sony had “been embracing the potential for responsibly produced AI.” Eighteen months later, Sony licensed KLAY Vision.
The numbers
700+ letters, one rate card.
THE TELL — An opt-out letter at scale works like an invoice template: you formally establish non-consent so that every future use requires a license from you, at your price. Sony wrote to 700 companies to create 700 future negotiations, and “embracing responsibly produced AI” in the same letter tells you the door was always for sale.
MBW: the 700 letters ↗TechCrunch coverage ↗MusicTech: letter details ↗
Sources: MBW (May 16, 2024); Variety (May 2024).
EXHIBIT E-3KLAY Vision: all three majors sign the same AI company
What happened
KLAY Vision became the first AI music company to license all three majors and their publishing arms — a full-house licensing sweep that not even the settling AI generators achieved. KLAY's leadership tells you why the majors trusted it: co-founded/led by Thomas Hesse, the former president of global digital business at SONY MUSIC itself, alongside Björn Winckler, former music lead at Google DeepMind. The majors licensed an AI company run by one of their own.
The numbers
3 of 3 majors. Terms undisclosed.
What changed after
Even Sony — the litigation holdout, the 700-letter sender — signed. The last major's last objection was never “AI shouldn't exist.” It was “AI shouldn't exist without our deal.”
THE TELL — This is the control group of the entire experiment: the same technology category as Suno and Udio, but a founding team led by an ex-major-label executive who came asking permission first — and the result was no lawsuit and all three signatures. How the majors treated KLAY versus Suno proves the variable was never the technology. It was the invitation. That sentence is the thesis of this whole page.
KLAY: press page ↗Music Week: all three majors ↗
Sources: Hollywood Reporter (Nov 20, 2025); UMG/Sony/Warner press releases; MBW.
Narrative pivots and admissions
Listen to what they tell investors, not what they tell artists.
EXHIBIT F-1The Grainge memos: a five-year mood ring
What happened
Read the memos in sequence and watch the story change. January 2023: the New Year memo calls for an ecosystem “in which our artists can thrive” and launches the “artist-centric” push — the first public admission that the pro-rata streaming model UMG spent fifteen years defending was broken. October 2023: on the Q3 earnings call, Grainge dismisses critics of artist-centric as “those whose business model is based on being merchants of garbage.” 2024: the vocabulary becomes “superfan,” “Streaming 2.0,” “responsible AI.” 2025: Grainge declares “the Streaming 2.0 era has arrived.” January 2026: the memo attacks “the exponential growth of AI slop” and “irresponsible business models” — published weeks after UMG licensed Udio and Stability AI, and weeks before its Downtown deal closed.
The numbers
Five years of memos, all traveling in the same direction.
THE TELL — Each memo is a course correction announced as a victory lap. Artist-centric (2023) admits pro-rata failed. Superfan/Streaming 2.0 (2024–25) admits subscription growth plateaued. The AI-slop attacks (2026) admit the upload flood is drowning catalog economics — while the company simultaneously licenses the flood's biggest engines. The memos are the most honest documents UMG produces, as long as you read what each one is quietly burying.
MBW: the Jan 2023 memo ↗MBW: “merchants of garbage” ↗MBW: the 2025 memo ↗MBW: the 2026 AI memo ↗
Sources: UMG internal memos as reported by MBW, Billboard, Variety (Jan 2023 – Jan 2026).
EXHIBIT F-2Capital Markets Day at Abbey Road: the growth math confession
What happened
UMG's investor day pitched “Streaming 2.0”: consumer segmentation, premium and superfan tiers, geographic expansion, and a target of ONE BILLION global subscribers by 2028, underwriting an 8–10% streaming revenue CAGR through 2028. Translation into plain English: the product that carried the last fifteen years (the $10.99 all-you-can-eat subscription) has saturated its rich-world market, so growth must now come from charging superfans more, poorer markets at lower ARPU, and products that don't exist yet.
The numbers
1B subscriber target by 2028. 8–10% CAGR promise. Meanwhile in UMG's own 2024 results: ad-supported streaming DOWN 5.1%; and per MIDiA, streaming's share of total recorded revenue FELL for the first time ever in 2024, to 61.3% from 62.4%.
THE TELL — “Streaming 2.0” is a name you only need if Streaming 1.0 stopped growing. Companies don't rebrand their core revenue engine at investor days when the engine is fine. The superfan pivot is the plateau, spoken in investor-relations dialect.
MBW: the CMD address ↗Music Week: the investor deck ↗MIDiA: streaming share dips ↗
Sources: Billboard Pro (Capital Markets Day coverage); iMusician; Variety (Mar 6, 2025); MIDiA (Mar 13, 2025).
EXHIBIT F-3Artist-centric: redrawing the royalty pool in their favor
What happened
UMG's “artist-centric” model, launched with Deezer, introduced a radical change: tracks under a threshold (the widely-adopted version: 1,000 annual streams) earn NOTHING — their royalties are redistributed up the pool to bigger artists. Spotify adopted the 1,000-stream minimum in 2024. UMG also negotiated contract provisions to keep “AI slop” and functional noise out of the same royalty pools as its artists.
The numbers
Per Luminate, roughly half of the ~202M tracks on streaming services were played 10 times or fewer — the long tail this model defunds. The redistributed money flows disproportionately to major-label catalogs, because majors own the head of the demand curve.
THE TELL — When you can't grow the pie, you redraw the slices. Artist-centric is marketed as protecting artists; mechanically, it's the incumbents' market share defending itself against 8.2 million self-releasing artists and an AI upload flood by changing the payout rules. It's a defensive rezoning of streaming economics — and the fact that the majors needed it is itself the admission.
Deezer–UMG announcement ↗MBW: Spotify's 1,000-stream rule ↗MBW: the ≤10-plays long tail ↗Luminate year-end report ↗
Sources: UMG–Deezer announcements (2023); Spotify royalty model changes (2024); Luminate 2024 Year-End Report; trade analysis.
The threat data
The numbers they're reacting to — the why behind every exhibit above.
EXHIBIT G-1The independents are outgrowing them on their own turf
What happened
Global recorded music grew 6.5% to $36.2B in 2024 — decelerating from 9.7% in 2023 (Mark Mulligan: “The much-anticipated streaming revenue deceleration – despite recent price increases – has now arrived”). Inside that slowing market: non-major labels and artists grew 8.2% to $10.7B, lifting their share to 29.7% — the THIRD consecutive year of share gains. In streaming specifically — the majors' core engine — non-majors grew 8.4%, outpacing the majors' 5.4%. On an ownership basis, non-majors were 46.7% of the market. Self-releasing artists hit 8.2 MILLION in 2024. Among the majors, only Sony gained share; UMG lost ~100 basis points. And streaming's share of total revenue fell for the first time ever: 61.3%, down from 62.4%.
The numbers
Non-majors: 29.7% share by label, 46.7% by ownership, 8.4% streaming growth vs the majors' 5.4%. 8.2M self-releasing artists.
THE TELL — Every exhibit in this file is a reaction to this table. The independents aren't a niche anymore — they're nearly half the market by ownership and growing faster than the majors on the majors' best surface. Now re-read Exhibit D-1: UMG didn't buy CD Baby and FUGA despite this data. It bought them BECAUSE of it.
MIDiA: the market shares ↗MBW coverage ↗RouteNote breakdown ↗Music Week coverage ↗
Sources: MIDiA Research (Mar 13, 2025); MBW; RouteNote; MIDiA “State of the Independent Music Economy.”
EXHIBIT G-2The flood: infinite supply vs. scarcity economics
What happened
Luminate's 2024 Year-End Report: 99,000 new tracks delivered to streaming services EVERY DAY. Roughly half of all ~202M tracks on streaming have 10 or fewer lifetime plays. Then the AI multiplier — Deezer's own detection data: fully AI-generated tracks as a share of daily uploads went ~10% (Jan 2025) → 18% (Apr 2025) → 28% (Sep 2025) → 34% (Nov 2025) → 44% (Apr 2026) → past 50% and ~90,000/day by June 2026. Deezer also found up to 70% of plays on fully-AI tracks were FRAUDULENT streams. And Spotify, September 25, 2025: over 75 MILLION “spammy” tracks removed in twelve months — against a total catalog of ~100M — alongside new policies on AI voice-clone impersonation and spam. Spotify's own words: at its worst, AI is used to “push 'slop' into the ecosystem… and often attempts to divert royalties to bad actors.”
The numbers
99,000/day human-era baseline. 50%+ AI share of uploads by mid-2026. 75,000,000 removed by one platform in one year. 70% fraud rate on AI-track plays (Deezer).
THE TELL — The entire major-label model is scarcity economics: control the valuable recordings, meter the access. AI broke the scarcity. When the daily upload flood exceeds what a human could hear in several lifetimes and half of it is machine-made, owning a catalog stops being a moat and starts being a needle in a stack of needles. THIS is what every lawsuit, layoff, acquisition, and royalty-rule change in this file is actually about — and the platforms, not the labels, are the ones holding the mop.
Spotify: For the Record ↗TechCrunch: Deezer's 50% ↗Deezer newsroom data ↗MBW: the upload flood ↗
Sources: Luminate 2024 Year-End Music Report; Deezer disclosures via TechCrunch (Jul 21, 2026) and MBW; Spotify “For The Record” (Sep 25, 2025); MBW, Hollywood Reporter, Variety, Rolling Stone, eWeek.
Regulatory and political moves
Lobbying for new rights with one hand, fighting antitrust with the other.
EXHIBIT H-1The two-front regulatory posture
What happened
Offense: the majors and RIAA championed the NO FAKES Act (a federal digital-replica right — statutory damages of no less than $5,000 per violation) and Tennessee's ELVIS Act (the first state law protecting voice against AI cloning) — creating brand-new intellectual property rights they would own and license. Defense: simultaneously, UMG absorbed an EU Phase II merger investigation, a Statement of Objections, and a forced divestiture on Downtown; sits under DOJ antitrust interest; and the industry operates in the shadow of the UK CMA's streaming market study. The EU AI Act's copyright-transparency obligations (enforcement August 2026) — which the rights-holder side lobbied to strengthen — were cited in GEMA's Munich win over Suno.
The numbers
NO FAKES: statutory damages of no less than $5,000 per violation. EU AI Act enforcement: August 2026.
THE TELL — A confident incumbent doesn't need new laws — the existing ones made it rich. Lobbying to create fresh rights (voice, likeness, training-consent) is the move of an industry that knows its current rights don't cover the world that's coming. Fighting antitrust scrutiny at the same time is the move of an industry that knows its response to that world — consolidation — looks like what it is.
Senate: NO FAKES release ↗ELVIS Act analysis ↗EC: AI Act transparency ↗Cooley: AI Act obligations ↗
Sources: RIAA/NMPA legislative statements (NO FAKES, ELVIS Act); European Commission Downtown filings; Lawfold (DOJ); UK CMA study; GEMA ruling coverage.
The timeline
Here's what five years of receipts add up to.
The biggest companies in music sued the future and then licensed it. They cut the old machine while shopping for the new one. They pulled their catalog off the biggest discovery platform on earth and found out the platform didn't need it. They put billions into buying the past and three-quarters of a billion into buying the independents' exits, rewrote the royalty rules when the pie stopped growing, and lobbied for brand-new rights because the old ones don't cover the world that's coming.
None of that is failure. It's three companies telling you — in filings, severance checks, and settlement signatures — that they know.
And the most important exhibit in the record is KLAY: the same AI technology the majors sued Suno over, except KLAY asked first, and all three majors signed. No lawsuit, no blackout, no war.
The majors were never against what's next — they were against being cut out of it.
So we're not cutting them out. We built PLTFRM to be fair to both sides of this industry — artists and the businesses that fund them — with a seat at the table for the big boys from day one: licensed, attributed, and paid from the start, without the sixteen-month lawsuit-to-license cycle or the walled-garden hostage terms. Their own five-year record shows this is the outcome they've been fighting to reach.
You've read the receipts. They know — and now they have an answer.
The fine print — we show our work
Undisclosed terms: The financial terms of the UMG–Udio, Warner–Udio, Warner–Suno, and KLAY licensing deals are not public. Sony catalog figures (Dylan, Jackson, Queen, Pink Floyd) are trade-press estimates, not company-confirmed; Springsteen (~$500M+) is the exception that was effectively confirmed. The Michael Jackson figure carries a reporting conflict: dominant reporting says $1.2B+ total valuation / ~$600M for the half; an outlier report said $1.6B for the half.
Interpretation vs. fact: Every date, dollar figure, filing, and quote on this page is documented public record. The “TELL” lines are our argument. The majors describe these same moves as strength, discipline, and “responsible AI” leadership. Read both framings and decide — that's why the sources are cited.
Ongoing matters: Sony v. Suno/Udio, the UMG–Suno impasse, the Drake appeal, the Anthropic suits, the AFM suit, the UMG royalty class action, the Spotify/MLC case, and the DOJ's antitrust interest (investigation stage — no complaint filed) are all live as of August 2026. Outcomes may change the record. We'll update the page when they do.
Data notes: MIDiA revised historical market-share figures in a 2024 methodology change. Spotify's 75M removals include pre-AI spam categories, not exclusively AI tracks. Deezer's AI-share figures measure its own upload stream via its proprietary detector.
Last updated: August 10, 2026.
Primary: RIAA press releases and complaints (D. Mass. / S.D.N.Y. dockets, June 24, 2024) · UMG earnings releases and Capital Markets Day materials · UMG “Open Letter to the Artist and Songwriter Community” (Jan 30, 2024) · Warner Music earnings and Kyncl memos · Sony Music opt-out letters (as reported) · Spotify “For The Record” (Sep 25, 2025) · Luminate 2024 Year-End Music Report · European Commission merger filings (UMG/Downtown) · Munich Regional Court ruling (GEMA v. Suno, Jul 31, 2026) · S.D.N.Y. dockets (Drake v. UMG; AFM v. UMG & WMG) · N.D. Cal. dockets (publishers v. Anthropic). Trade press of record: Music Business Worldwide · Billboard / Billboard Pro · Variety · Digital Music News · Hollywood Reporter · Rolling Stone · TechCrunch · Bloomberg · TIME · LA Times. Analysts: MIDiA Research (Mark Mulligan) · Chartlex AI Lawsuits Tracker (docket-verified) · Lawfold litigation summaries · Dynamoi · AI Musicpreneur case trackers.