Publishing, music & rights-based IP

Publishing, music & rights-based IP

Industry ID: 24 | Slug: publishing-music-ip | Researched: 2026-09-15 | Analyst: agent

Research-capacity note. The shared 200-call WebSearch pool for this session was already exhausted when this sector began; the single permitted search returned a budget message and zero results. Per the standing instruction in the macro brief, the entire dossier was built by WebFetch against primary sources — SEC EDGAR, the US Copyright Office, Justia federal dockets, company IR and newsroom pages, IFPI, RIAA, AAP, the Authors Guild, Pew Research Center, NAVER and Kakao Korean-language IR, the European Commission and gov.uk. This raised the average source tier materially (the source registry is 22 Tier A of 25) but narrowed discovery: I could not sweep trade press for events that no primary source announces, which in this sector means deal terms, private licensing prices and executive changes. Affected areas are itemised in §13. Several hosts were blocked outright by the egress proxy and are recorded there too. This sector should be flagged for re-run with restored search capacity, principally to recover Universal Music Group and News Corp primary disclosure.

0. The merge question, answered up front

This sector was created by merging music with publishing on the stated rationale that both halves' dominant 2026 trend is the same one — who owns, licenses and is paid for IP used in AI training and generation. The brief asked me to test that rationale rather than assume it.

My finding: the rationale is half right, and the half that is wrong is load-bearing enough to justify splitting in Phase 2.

What genuinely is shared:

• The legal substrate is identical. Both halves litigate under 17 U.S.C. §§ 101/501, both depend on timely registration under §§ 411–412 for statutory damages, both are governed by the same fair-use four-factor test, and both are affected by the same US Copyright Office reports and the same EU AI Act Article 53(1)(d) transparency template. A ruling in the OpenAI MDL binds the reasoning applied to song lyrics and to novels alike.
• The counterparties are the same five companies. OpenAI, Anthropic, Google, Microsoft and Meta are the defendants on both sides. Anthropic alone is being sued by book authors (Bartz, Kwon, Shakespeare) and by music publishers (Concord) in the same district.
• The policy fight is one fight. IFPI's September 2026 EU submission and AAP's amicus in In re Mosaic LLM make the same argument in the same words: licensing not exceptions, transparency not remuneration rights.

What is not shared, and this is where the merge breaks:

1. The two halves are at opposite ends of the AI event. Book publishing has already been paid: $1.5bn, final approval 2026-07-20, roughly $3,000 per work. Music has not been paid in any disclosed amount — every music AI licence (Udio–UMG, Udio–WMG, Suno–WMG, Suno–BMG, Suno–Believe) is confidential with no terms published, and the only auditable music number is Warner's $16m of settlement revenue in a $1,864m quarter, which is 0.9%. Publishing's AI trend is a realised cash event; music's is an unpriced option. Those require different analysis, different metrics and different monitoring cadence.
2. The direction of travel is opposite. For book and news publishing, AI is a threat being converted into compensation — the underlying businesses grow 2.5% (books) or shrink 4.0% (local news). For music, AI is a supply-side nuisance attached to a healthy business growing 6.4% globally on price increases that have nothing to do with AI. A trend platform that scores "AI copyright" as one trend across this sector will systematically mis-weight both.
3. Comics and webtoons do not belong with either. This is the sharpest finding in the dossier. Webtoon revenue is contracting — WEBTOON Entertainment −1.5%, Kakao Story −15.8%, NAVER Contents +0.5% — while Kakao's Music line inside the same segment of the same issuer grew 7.8%. AI is not the operative variable for webtoons at all; a Korea/Japan consumption cycle and a film-and-TV commissioning pullback are. Webtoons sit inside this sector only because "comics are published," which is a taxonomy accident, not an economic relationship.
4. News publishing's dominant trend is not the training-data question at all. It is the inference-side destruction of referral traffic: Pew measures click-through falling from 15% to 8% when an AI summary appears, and source clicks inside the summary at 1%. That is a distribution problem, not a rights problem. Compensation litigation is a response to it, not the trend itself. Music has no analogue — nobody's discovery funnel was removed.

Recommendation for Phase 2: split into three. (a) Music and rights-based IP — recorded music, publishing, catalogue as an asset class, generative audio; a growing, concentrated, price-led sector whose AI story is a supply shock and an unpriced licensing option. (b) Book and news publishing — a low-growth sector whose AI story is a realised compensation event and a distribution collapse, and whose key variables are court rulings and registration formalities. (c) Comics, webtoons and Korean IP — either its own sector or folded into sector 14 (media/film/TV), which owns the adaptation demand that actually drives it.

Keep one thing merged across all three: a cross-cutting AI training-data rights trend line, monitored once, feeding all of them. That is the genuinely common object. It is a trend, not a sector.

1. Definition and boundaries

In scope. The creation, ownership, licensing and monetisation of text, music and static narrative IP: trade, educational and professional book publishing; comics, manga and webtoons; news and magazine publishing; recorded music and music publishing (compositions); podcasts, audiobooks and spoken-word audio; the streaming and retail platforms through which these are distributed insofar as they set the sector's economics; catalogue acquisition as an asset class; and the collecting societies, rate-setting bodies and copyright registries that administer the whole.

Explicitly out of scope, and who owns it.

• Film, television and premium video — sector 14. Music licensing into film and TV is a revenue line here (WMG licensing $111m in the quarter); the moving-image business is not. The boundary is porous at webtoon IP adaptation, where sector 14's commissioning cycle is the demand driver for this sector's supply (see T-24-18) — that trend should be cross-read.
• Video games and interactive — sector 15, including music licensing into games.
• Creator-economy platforms and social video as a business model — sector 17. YouTube appears here only as a music and podcast distribution counterparty.
• Advertising technology and the ad market — sector 16. Ad-supported music and news revenue is in scope; the ad stack is not. Note sector 16 established that on 2026-09-02 the court declined to break up Google's ad stack, a behavioural-only remedy — relevant here because news publishers' remaining ad revenue depends on it.
• AI foundation models themselves — sector 01. The labs appear here as licensees and defendants only.
• Live music, touring, ticketing and merchandise. Genuinely adjacent and arguably misplaced: live is now the largest single revenue pool in music and is not covered by any sector in this programme. Flag for Phase 2.

Boundary disputes worth naming. (a) Is Spotify a publishing/music company or a platform? It now distributes music, podcasts and audiobooks and sets royalty terms for all three, which makes it a price-setter in three of this sector's subcategories at once. (b) Is catalogue acquisition a media business or an asset-management business? Reservoir's 42% adjusted EBITDA margin and 15.5x entry multiples argue the latter; its A&R function argues the former. (c) Are webtoons publishing or media? Argued at §0. (d) Is news publishing a content business or a distribution-dependent advertising business? The Pew data says its economics are set at the search layer, which sits in sector 16.

2. Subcategories

1. Trade book publishing — consumer titles across hardback, paperback, ebook and digital audio. Distinguished by a hit-driven frontlist funded by a predictable backlist, and by being the half of the sector that has actually been paid by an AI developer.
2. Educational, academic and professional publishing — textbooks, journals and reference. Distinguished by institutional rather than consumer buyers, subscription licensing and much higher margins than trade.
3. News and magazine publishing — national, regional and local. Distinguished by dependence on third-party distribution (search, social) for audience acquisition, which is precisely what AI search has removed.
4. Comics, manga and webtoons — serialised visual narrative, overwhelmingly Korean and Japanese in origin and economics. Distinguished by per-episode micropayment monetisation and by an IP-adaptation option that has just stopped paying.
5. Recorded music — masters and their exploitation. Distinguished by three-firm concentration and by streaming subscription as ~70% of revenue.
6. Music publishing — compositions, and the mechanical, performance, sync and print rights attached. Distinguished by statutory rate-setting (CRB) and collective administration, which makes it far more regulated than recorded music.
7. Catalogue and royalty investment — the purchase of rights streams as a financial asset. Distinguished by being underwritten on discount rates rather than on creative judgement, and therefore uniquely exposed to the rate environment.
8. Music streaming platforms — Spotify, Apple, Amazon, YouTube, Deezer. Distinguished by setting the sector's realised price and by owning the fraud and AI-detection problem.
9. Podcasts and spoken-word audio — advertising- and subscription-funded talk audio, now substantially video. Distinguished by direct advertiser relationships rather than royalties.
10. Audiobooks — the fastest-growing book format and the one most exposed to AI narration cost deflation.
11. Generative audio and text platforms — Suno, Udio, KLAY. Distinguished by being simultaneously defendants, licensees and a new distribution channel.
12. Rights administration and collective management — collecting societies, the MLC, SoundExchange, the Copyright Office registry. Distinguished by being infrastructure whose failure modes (see registration, §10) now determine who gets paid.

3. Market structure

Concentration varies sharply by half, which is itself an argument against the merge.

Recorded music is a textbook oligopoly. Three groups — Universal, Sony Music and Warner — hold the majority of global recorded-music revenue, and IFPI's $31.7bn 2025 trade figure is substantially a measure of their reporting. Warner alone booked $1,864m in a single quarter. The three negotiate individually with DSPs and, as this dossier shows, individually with AI developers. Barriers to entry are catalogue, not capability: you cannot build a back catalogue.

Music publishing is an oligopoly constrained by a regulator. The same three groups dominate, but the Copyright Royalty Board sets statutory mechanical rates and announced final rules on 2026-03-13 for digital sound-recording performance covering 2026–2030. Pricing power is therefore partly administrative rather than commercial — a structural difference from recorded music that most analysis elides.

Book publishing is a concentrated oligopoly with a fragmented tail. Five trade groups dominate US consumer publishing, sitting atop a very long tail of independents and an unmeasured self-publishing sector that AAP's ~1,200-publisher panel largely excludes.

News publishing is fragmenting into two tiers. The New York Times grew Q1 2026 digital-only subscription revenue to $389.0m; USA TODAY Co (the renamed Gannett, now NYSE:TDAY) shrank 4.0% with print circulation down 12.0% and carries $988.3m of principal debt. There is no middle.

Webtoons are a duopoly in their home market — NAVER and Kakao — with Japan (Piccoma, LINE Manga) as the profit pool. WEBTOON Entertainment's Japan ARPPU of $22.5 against Korea's $7.8 and Rest of World's $6.8 shows where the economics actually live.

Streaming is winner-take-most but not winner-take-all. Spotify at 300m premium subscribers and 777m MAU is the price-setter; Deezer at EUR268.2m of half-year revenue survives as a differentiated niche and, notably, as the sector's information source.

Where margin actually sits. Not with creators. Spotify's own Loud & Clear data shows that in 2025 more than 1,500 artists generated over $1m and over 13,800 over $100,000, but the 100,000th-highest-earning artist generated just over $7,300 — and all of those are gross to the rightsholder, before label and distributor splits. Spotify itself ran a 33% gross margin and EUR655m of quarterly operating income. Warner ran a 23.2% adjusted OIBDA margin. Reservoir, a pure rights owner with no A&R obligation to speak of, ran 42%. The margin sits with whoever owns the right in perpetuity and has no marginal cost — which is why catalogue trades at 15.5x and why the whole sector's strategic behaviour is rights accumulation.

Barriers to entry. Catalogue ownership; statutory licensing complexity; distribution relationships with a handful of DSPs and retailers; and, newly, the ability to indemnify an AI developer, which only large rightsholders can do.

Pricing power. With the majors and with Spotify, jointly. Spotify's Q2 2026 ARPU rose 7% with EUR0.49 from price increases; Warner attributed 12.5% subscription growth partly to "improved terms with DSP partners." Both sides raised price at once. Neither creators nor consumers have countervailing power.

Market size figures — with modeller, date and basis named.

• Global recorded music trade revenue $31.7bn in 2025, +6.4% — IFPI, published 2026-03-18. Trade (wholesale) basis; excludes live, merchandise and most unreported independent revenue.
• US recorded music ~$6.0bn mid-year 2026, +6.9% at estimated retail value — RIAA, 2026-09-01. A wholesale figure of ~$4.9bn circulates from the same cycle; see §13.
• US book publishing $33.4bn calendar 2025, +2.5%; trade $21.7bn — AAP StatShot, published 2026-08-10. Panel of ~1,200 reporting publishers; excludes most self-publishing.
• EU recorded music EUR6.0bn in 2025, +5.1%, 21.3% of global — IFPI, September 2026.
• Global label investment in A&R and marketing EUR8.3bn in 2025 — IFPI, September 2026.

No market-size figure in this dossier comes from a commercial "market report" vendor. Where I could not verify a figure from a primary issuer, it is absent and recorded in §13.

4. Who matters

Leading companies. Universal Music Group (https://www.universalmusic.com/) · Warner Music Group (https://www.wmg.com/) · Sony Music Entertainment (https://www.sonymusic.com/) · Spotify (https://www.spotify.com/) · Deezer (https://www.deezer.com/) · BMG (https://www.bmg.com/) · Believe (https://www.believe.com/) · Concord (https://concord.com/) · Reservoir Media (https://www.reservoir-media.com/) · Penguin Random House (https://global.penguinrandomhouse.com/) · News Corp / HarperCollins (https://newscorp.com/) · The New York Times Company (https://www.nytco.com/) · USA TODAY Co, formerly Gannett (https://www.usatodayco.com/) · WEBTOON Entertainment (https://www.webtoons.com/) · NAVER (https://www.navercorp.com/) · Kakao (https://www.kakaocorp.com/)

Notable startups. Suno (https://suno.com/) — $400m+ Series D at $5.4bn, 2026-06-03 · Udio / Uncharted Labs (https://www.udio.com/) · KLAY — identified by IFPI as the first AI music startup licensed by all three majors pre-launch.

Active investors. Bond Capital (https://www.bondcap.com/) led Suno's Series D, with IVP, Forerunner, Union Square Ventures, Alkeon, Quiet, Matrix, Lightspeed, Menlo Ventures and Schroders Capital participating. Bain Capital (https://www.baincapital.com/) is WMG's catalogue joint-venture partner; the associated Beethoven Financing 1 LLC facility was raised from $500m to $750m on 2026-05-05 with Goldman Sachs, Fifth Third, Barclays, MUFG, Bank of America and Mizuho as lenders. Blackstone owns Recognition Music Group (formerly Hipgnosis Songs Fund) — referenced but not verifiable from a primary source in this session.

Platforms and standards bodies. Spotify · Apple Music · Amazon Music · YouTube Music · Audible · Deezer · TuneCore · IFPI's ISRC and GRid identifier systems · DDEX (recording credits/metadata) · the EU AI Office's Article 53(1)(d) training-data summary template.

Regulators. US Copyright Office (https://www.copyright.gov/) · Copyright Royalty Board (https://www.crb.gov/) · European Commission DG CNECT and the AI Office (https://digital-strategy.ec.europa.eu/) · UK Intellectual Property Office and DSIT (https://www.gov.uk/government/organisations/intellectual-property-office) · the federal courts, which in 2026 are doing more rule-setting than any of the above.

Research institutions. Pew Research Center (https://www.pewresearch.org/) — the only Tier A quantification of AI search's referral effect. Edison Research with SSRS (https://www.edisonresearch.com/) — The Infinite Dial, the reference US audio series since 1998.

Trade organisations. IFPI (https://www.ifpi.org/) · RIAA (https://www.riaa.com/) · Association of American Publishers (https://publishers.org/) · News/Media Alliance (https://www.newsmediaalliance.org/) · National Music Publishers' Association · The MLC · SoundExchange · ASCAP and BMI.

Consumer and civil-society groups. The Authors Guild (https://authorsguild.org/) — a membership body for authors and a named MDL plaintiff, and the source of the registration survey at §10. Note the genuine gap: there is no organised consumer voice in this sector's AI debate. Every organised participant is a rightsholder, a platform or an AI developer. Listeners and readers are represented by nobody, which is worth recording as a structural feature.

Named individuals. Maria A. Pallante (President and CEO, AAP; former Register of Copyrights) · Shira Perlmutter (Register of Copyrights) · Judge Sidney H. Stein (S.D.N.Y., presiding over MDL 25-md-3143) · Alexis Lanternier (CEO, Deezer) · Namsun Kim (director, WEBTOON Entertainment). Where I could not verify an executive's name against a directly retrievable primary source — most notably at UMG, Sony Music, WMG and Suno — the entity records say unverified rather than guessing.

5. Products, business models, technologies, customers

Major products. Recorded music catalogues and new releases; musical compositions and the mechanical, performance, synchronisation and print rights over them; trade, educational and professional books across print, ebook and digital audio; serialised webtoon episodes; daily and periodical journalism; podcasts; and, newly, licensed generative music models as a product in their own right (Suno v6, shipped 2026-09-09).

How money is actually made today.

• Music: subscription streaming is the engine — 69.6% of global recorded-music income and 52.4% from paid subscription alone. Revenue accrues to the rightsholder as a share of a platform's revenue pool, not per stream, which is why "per-stream rate" is a derived and unstable number that no platform discloses.
• Books: unit sales through retail and library channels, with print still 50.6% of industry revenue against 14.2% digital. The business is frontlist risk funded by backlist annuity.
• News: bifurcated between direct digital subscription (NYT: $389.0m in a quarter) and a declining print circulation and advertising base (USA TODAY Co: print circulation −12.0%).
• Webtoons: per-episode micropayments — $261.4m of WEBTOON Entertainment's $320.9m quarter — plus advertising and IP adaptation licensing.
• Catalogue: buy a royalty stream at a multiple of net publisher share, finance it with debt, and earn the spread plus whatever administrative uplift you can generate.

How that is changing.

1. Price replaces volume. Detailed at T-24-03. Spotify ARPU +7% with EUR0.49 attributable to price rises; Warner subscription streaming +12.5% on renegotiated DSP terms.
2. The free tier stops working. Spotify ad-supported revenue grew 1% while its ad-supported MAU grew 14% (T-24-17). Revenue per free user is falling outright.
3. Litigation and licensing become a revenue line. Warner recognised $16m of digital revenue from copyright infringement settlements in one quarter. Small, but it is now a line item.
4. Metered export replaces unlimited generation. Suno's download tiers — 7 lifetime free, 20 Pro, 60 Premier, unlimited Studio, effective 2026-09-03 — are the first observable mechanism by which rightsholder deals reshape generative-AI unit economics (T-24-10).
5. Detection becomes a product. Deezer now licenses its AI music detector to Hungarian and Dutch collecting societies (T-24-11).

Technologies that matter. Generative audio models (Suno v6, Udio, KLAY); AI content detection and provenance classification; recommendation and playlist systems, which are the actual allocation mechanism for music attention; text-to-speech for audiobook narration, the main cost deflation vector in publishing; rights-management and registration infrastructure, whose failure is documented at §10; and retrieval-augmented AI search, which is the mechanism destroying news referral traffic.

Customer segments and what they buy on. Consumers buy music on convenience and catalogue completeness, not on price discovery — which is why coordinated price increases stick. Book buyers buy on author and title, making publishing a brand business at the author level. News subscribers buy on habit and identity. Institutional buyers (schools, libraries, universities) buy on curriculum fit and licensing terms, and are the least AI-exposed segment. AI developers — a genuinely new customer segment since 2025 — buy on legal certainty, not on content quality, which is why the licensing deals are structured as indemnity-bearing partnerships rather than as content supply contracts.

6. Geography

Production. Music A&R and recording concentrate in the US, UK, Sweden, South Korea and increasingly Latin America and Nigeria; IFPI reports Latin America as the fastest-growing region at +17.1% and MENA and Sub-Saharan Africa both at +15.2%. Book publishing concentrates in the US, UK and Germany. Webtoon and manga production is overwhelmingly Korean and Japanese, and this is the sector's most geographically concentrated subcategory by a wide margin.

Capital. US and UK private capital dominates catalogue acquisition; Bond Capital and the Silicon Valley syndicate fund generative audio. Korean platform capital (NAVER, Kakao) funds webtoons, and both are currently retrenching rather than deploying.

Demand. The US and Canada remain the largest revenue pool but grew only 3.5% in 2025. Europe grew 5.6% at 30.4% of global, Asia 10.9%. Penetration is inverted against revenue: IFPI reports EU paid-subscriber penetration at 54%, the UK at 47% and the USA at just 27% — so the largest market is the least penetrated, which is a materially more optimistic picture for US subscriber growth than the "saturation" narrative allows, and a materially more pessimistic one for European volume growth.

Regulation. The EU is the only jurisdiction with an operative AI training-transparency instrument. The UK has a consultation that has produced nothing in over a year. The US is regulating through Article III courts. Korea and Japan regulate webtoon platforms mainly through app-store and platform rules — note from sector 17 that Apple's China commission moved to 25%/12% from 2026-03-15 and EU CTC to 5% from 2026-10-01, which affects webtoon micropayment economics jurisdiction by jurisdiction.

Non-US market, covered with regional sources. South Korea. Two Korean-language primary disclosures were retrieved. NAVER (KRX:035420) reported Q2 2026 group revenue of KRW3,388.8bn (+16.2%) and operating profit of KRW520.3bn (−0.2%, margin down to 15.4% from 17.9%), with the Contents segment — which contains Webtoon — at KRW464.3bn, up just 0.5%. NAVER's own framing is that Webtoon must "focus on expanding user base via content expansion including new AI services," which is turnaround language. Kakao (KRX:035720) reported Q2 2026 revenue of KRW2,098.5bn (+9.4%) and operating profit of KRW277.0bn (+35.9%), with its Content segment up 0.7% — inside which Story (webtoon and web novel, including Piccoma) fell 15.8% to KRW210.7bn while Music grew 7.8% to KRW558.4bn. Two Korean issuers, reporting in won, both show the same divergence: Korean music is healthy, Korean webtoon is not. This is the single most useful regional finding in the dossier and it is the empirical basis for the split recommendation at §0.

European Union. IFPI's Music in the EU report (September 2026) supplies EUR6.0bn 2025 revenue at +5.1%, 21.3% global share, nine of the global top 20 markets inside the EU accounting for EUR5.1bn, fastest growth in Romania, Slovakia and the Baltics at +15.9% to +16.1%, and a domestic repertoire share of 53.5% of year-end top-10 tracks.

7. Historical trend patterns

Over 25 years this sector has run one very long cycle and several instructive false positives.

The actual cycle. Recorded music revenue peaked around 1999, fell for roughly fifteen years under unlicensed file-sharing, bottomed in the mid-2010s and has now grown for eleven consecutive years to $31.7bn. The lesson that matters for 2026: the industry's last existential technology shock was resolved by licensing, not by enforcement — Napster was shut down and it changed nothing; iTunes and then Spotify licensed the behaviour and that changed everything. The current bilateral AI licensing wave (T-24-02) is the same institutional reflex, and the people running it lived through the first one. That is a genuine reason to weight licensing outcomes above litigation outcomes, and it is under-appreciated.

Prior hype waves in this sector and how they resolved.

• The ebook disruption thesis (2010–2014). Ebooks were forecast to displace print. They did not: in 2025 ebooks were $2.2bn against $16.2bn of print, and print is 50.6% of industry revenue. Format substitution in books has repeatedly been forecast and has repeatedly plateaued. This is the closest available analogue to "AI music will replace human music," and it argues strongly for the sceptical reading at T-24-19.
• The "vinyl is dead" and then "vinyl is the future" cycles. Vinyl has now grown for 19 consecutive years and physical grew 8.0% in 2025 — but it remains a small share. Both the obituary and the revival narrative overshot.
• The catalogue bubble (2019–2022). Zero rates drove catalogue multiples to levels that required permanently cheap money. Hipgnosis Songs Fund's NAV write-downs and eventual sale to a Blackstone vehicle ended it. The sector is now living the sober version: Reservoir buying at 15.5x with disclosed leverage. Anyone extrapolating 2021 catalogue multiples into 2026 is repeating a documented error.
• The NFT and Web3 music wave (2021–2022). Forecast to disintermediate labels and pay artists directly. It produced no measurable revenue line in any IFPI, RIAA or major-label disclosure before or since. It is the sector's cleanest recent example of a trend that existed entirely as attention — exactly what the contract's persistence and evidence_quality dimensions are designed to catch.
• The podcast gold rush (2019–2022). Large advances and platform exclusives were written on an assumption of continued audience growth and advertising scaling. Reach did keep growing — Infinite Dial 2026 puts monthly podcast reach at 58% of Americans, 167m people — but the monetisation did not follow the content-spend, and the exclusivity model was abandoned. The parallel to today's generative-music valuations is direct: audience reach was real and the business case was still wrong.
• "Streaming will pay artists fairly once it scales." It scaled. Spotify paid over $11bn in 2025 and nearly $70bn cumulatively, and the 100,000th-ranked artist still earned just over $7,300. The distributional question was never a scale question.

The pattern worth generalising. In this sector, supply-side shocks are absorbed and demand-side shocks are not. Home taping, CD burning, file-sharing, self-publishing, SoundCloud, DistroKid, and now generative AI have each multiplied the amount of available content without destroying the industry, because attention — not supply — is the scarce good. The shocks that did damage incumbents were the ones that changed distribution and discovery: iTunes unbundling the album, and now AI search removing the click. That is why T-24-05 (news referral collapse) deserves more weight than T-24-04 (AI upload flood), even though the latter gets far more coverage.

8. What is changing now (as of 2026-09-15)

Grounded in the macro brief:

• Capital is expensive and that is now visible in this sector's balance sheets. The FOMC held at 3.50–3.75% on a 9–3 vote with three dissents in favour of a hike, and the 2026-09-16 decision is unresolved one day after this research date. Catalogue is a long-duration royalty-stream asset and is therefore the most rate-sensitive thing in the sector. Reservoir carries $430m of net debt against $74m of adjusted EBITDA (~5.8x); Warner carries $4,710m of long-term debt; USA TODAY Co carries $988.3m. The catalogue trade has not stopped — the Beethoven facility grew from $500m to $750m on 2026-05-05 — but 15.5x entry multiples with value-enhancement targets have replaced yield-buying.
• AI funding concentration distorts everything adjacent. Per the macro brief, OpenAI and Anthropic alone took 43% of H1 2026 global VC. The relevant consequence here is that the sector's counterparties can pay a $1.5bn settlement out of petty cash, which is why settlement is available as a resolution mechanism at all. It is also why $5.4bn valuations attach to generative-music companies with no disclosed revenue.
• The paper/cash divergence from sector 11 applies directly. Suno's valuation rose from $2.45bn to $5.4bn in about seven months with no disclosed revenue, ARR or subscriber figures. Against a 2021-vintage DPI of 0.05x and a seed→Series A graduation rate of 16%, that mark should be read as sentiment, not as validation of the demand thesis.
• Courts and states, not federal policy, are setting boundary conditions — the pattern Addendum 2 identified in Google's ad-stack remedy and the Paramount/WBD block. This sector is the purest case: the $1.5bn Bartz approval, the pending Authors Guild v. OpenAI summary judgment, and the Concord complaint naming two Anthropic officers personally are doing all the rule-setting. The US Copyright Office recommended no legislation; the UK has produced nothing in over a year.
• Regulatory-deadline slippage is confirmed again. The macro brief recorded the EU AI Act high-risk deferral to 2027-12-02. This sector adds a second instance: the UK copyright-and-AI consultation closed 2025-02-25 and was still "analysing your feedback" at its 2026-03-19 revision. Any thesis in this sector that depends on an announced regulatory date should be scored low on persistence by default.
• Public-market disclosure is shrinking here too. Sector 15 noted EA's take-private removing a top-five publisher's reporting. This sector's equivalent is structural rather than transactional: Universal Music Group's investor site and News Corp's corporate site were both unreachable through this session's egress proxy, which — combined with Sony Music reporting only as a segment of a conglomerate — means only one of the three major music groups (Warner) publishes quarterly, auditable, segment-level economics. That is a serious and permanent data-availability constraint on this sector, not an artefact of this session.

9. The five lists

Five most important current trends

1. Court-supervised settlement setting the price of pirated training data (T-24-01)
2. Bilateral, label-by-label AI music licensing replacing litigation (T-24-02)
3. Music revenue growth coming from price, not listeners (T-24-03)
4. AI-mediated search breaking the news referral economy (T-24-05)
5. Copyright registration formalities becoming the gating asset (T-24-08)

Five fastest-growing signals

1. Licensed generative-music models as a distinct product tier (T-24-09)
2. Download metering as the rightsholder monetisation lever (T-24-10)
3. AI detection and provenance tooling as a licensable product (T-24-11)
4. Personal liability for executives in AI copyright actions (T-24-12)
5. Local and regional publishers organising into collective AI actions (T-24-13)

Five trends most likely to affect businesses

1. T-24-01 — settlement as price discovery; it sets every subsequent negotiation
2. T-24-03 — price-led growth changes the whole revenue model's risk profile
3. T-24-07 — catalogue discipline and cost of debt determine capital allocation
4. T-24-16 — the webtoon growth thesis needs rebuilding from the current numbers
5. T-24-15 — disclosure-not-exception is the operative regulatory regime to plan against

Five trends most likely to affect consumers

1. T-24-03 — consumers are paying the price increases that are funding sector growth
2. T-24-04 — the upload flood degrades discovery and fills catalogues with fraud
3. T-24-05 — fewer people reach journalism, with the sharpest effect on local news
4. T-24-10 — generative music moves from free-and-unlimited to metered-and-paid
5. T-24-14 — AI labelling will change what consumers can see about what they hear

Overlaps, stated explicitly. T-24-03 appears on three of the four lists because it is the sector's actual economic engine and touches business model, consumer price and geography at once. T-24-01 appears twice — it is both the most important current trend and the one with the largest business consequence. T-24-04 and T-24-05 both appear on the consumer list and underpin overhyped/cooling entries elsewhere; the trend records carry the reconciliation. No trend appears on all four lists, which is deliberate: nothing in this sector is simultaneously the fastest signal and the most durable force.

10. Overhyped / overlooked / cooling / reversing

Most overhyped

1. "AI music is taking over streaming." (T-24-19) The claim rests entirely on an upload statistic — Deezer's >50% of daily new uploads at peak in June 2026, about 90,000 tracks a day — detached from the same disclosure's demand figure: fully AI-generated music is 1–3% of total streams. It is further inflated because most of that sliver is not listening at all: Deezer identified up to 85% of streams on fully AI-generated tracks as fraudulent in 2025 and demonetised them, against platform-wide fraud of 8%. Net of fraud, genuine human demand for fully AI-generated music is plausibly in the low tenths of a percent of streams. Meanwhile global recorded music grew 6.4%, US revenue grew 6.9%, and Spotify added 24m premium subscribers — no industry series shows substitution. The specific evidence that hype outruns substance: the two numbers come from the same press release, and only one of them is ever quoted.

2. Collective or statutory licensing for AI training as the imminent settlement. (T-24-20) Widely promoted, and every primary source points the other way. IFPI's September 2026 EU submission explicitly opposes statutory remuneration rights, compulsory licensing and mandatory collective management. The US Copyright Office examined compulsory and collective licensing in Part 3 and recommended neither — and Part 3 is still a pre-publication draft sixteen months on. The UK consultation has produced nothing in over a year. Every actual deal is bilateral and confidential. The evidence that hype outruns substance: the largest rightsholders, who would be the principal beneficiaries of a levy, are lobbying against one, because bilateral leverage pays them more.

Most overlooked

1. Publishers did not register their authors' copyrights, and it decided who got paid. (T-24-08) The Authors Guild's 2026-08-20 survey of 500+ authors covering 1,400+ books across nearly 400 imprints found that of the 864 books with compliance data, 804 — about 93% — showed publisher failure to register, even though 54% were under contracts requiring it. Because Bartz eligibility required registration within three months of publication or before the infringement, a large share of authors were excluded from a $1.5bn fund by an administrative omission. Why attention has missed it: it is a contract-compliance story with no villain and no technology, published by a membership body in August, while the coverage was all on the headline settlement number. It is nonetheless the single most consequential distributional fact in the sector — and it generalises: statutory damages, and therefore the economic viability of every AI copyright claim, depend on a formality that most of the corpus fails.

2. Spotify's ad-supported business has stopped working. (T-24-17) Ad-supported revenue grew 1% in Q2 2026 while ad-supported MAU grew 14%, and H1 revenue actually fell 2%. Revenue per free user is declining outright. Why attention has missed it: it is four lines in a 6-K during a quarter with record headline profit, and it contradicts the industry's preferred story that the free tier is a conversion engine rather than a cost. It matters because ad-supported was the answer for the two-thirds of the world that will not subscribe, and it is not answering.

3. Only one of the three major music groups publishes auditable quarterly economics. Warner files with the SEC; UMG reports to Euronext on a different cadence and its investor site was unreachable here; Sony Music is a segment of a conglomerate. Analysts routinely generalise from Warner to "the majors." That is a one-firm sample. Record it as a data-availability trend in its own right, in the same spirit as Addendum 2's note on EA.

Trends that appear to be cooling

1. Webtoons as a growth story. (T-24-16) The indicator that turned: WEBTOON Entertainment's Q1 2026 revenue fell 1.5% year over year (+0.2% constant currency) with adjusted EBITDA margin at 3.0% and Q2 guidance of 0.0–1.5%; Kakao's Story line fell 15.8%; NAVER's Contents segment grew 0.5% against group growth of 16.2%. Three issuers, two currencies, one direction.

2. Ad-supported music streaming. (T-24-17) The indicator that turned: Spotify's H1 2026 ad-supported revenue declined 2% in absolute terms while its free user base grew 14%.

3. Comics IP adaptation revenue. (T-24-18) The indicator that turned: WEBTOON Entertainment's IP adaptations line fell 22.8% to $19.8m — the steepest decline in any of its reported segments, and the mechanism on which the entire "webtoons are an IP factory" thesis rested.

Trends that may reverse, and the mechanism

• The licensing détente could revert to litigation. Mechanism: the bilateral deals are confidential and none has disclosed a rate. If artist and songwriter constituencies conclude the terms are inadequate — or if Sony's Uncharted Labs suit produces a large damages award — rightsholders who settled cheaply will have an incentive to reopen. Sony's 2026-07-20 filing after UMG and WMG settled is the first instance of exactly this dynamic.
• The industry's opposition to collective licensing could invert overnight. Mechanism: a broad fair-use ruling for developers in the OpenAI MDL would destroy bilateral leverage, at which point the same trade bodies now opposing a statutory remuneration right would demand one. Watch the summary-judgment ruling — briefing completes in early November 2026.
• Catalogue could re-inflate. Mechanism: if the September 2026 FOMC decision and those following it resolve toward cuts, the discount rate on long-duration royalty streams falls and 15.5x becomes 18x again quickly. The Beethoven facility increase to $750m suggests acquirers are positioning for precisely this.
• The AI-music demand thesis could yet prove right in a place nobody measures. Mechanism: substitution would appear first in low-attention functional music — background, ambient, production, fitness — which no published series separates out. The 1–3% aggregate would mask a much higher share in exactly the categories where per-track royalties are most fungible. This is the strongest argument against my own overhyped classification and it is recorded as a risk on T-24-19.

11. Risks and major uncertainties

Sector-specific risks.

• Concentration risk in disclosure. One of three majors reports properly (see §10). Conclusions about "major-label economics" rest on Warner.
• Single-source risk on the sector's defining statistic. Deezer is the only platform publishing AI-upload and AI-fraud data, and it sells the detection product those numbers justify. There is no second series to triangulate against. Every AI-music figure in this dossier inherits that fragility, which is why T-24-04 is marked single_source and capped accordingly.
• Refinancing risk. Reservoir at ~5.8x net leverage, Warner at $4,710m, USA TODAY Co at $988.3m, in an environment where three FOMC members dissented in favour of a hike.
• Concentration of settlement counterparties. The compensation thesis depends on five companies remaining solvent and willing to settle. The macro brief's warning about AI-firm equity valuations and leveraged infrastructure financing applies directly: if AI capital markets correct, the sector's new revenue line evaporates before most of it is collected.
• Termination-of-transfer recapture. Reservoir's own 10-K flags that US authors can recapture rights in compositions and recordings. Catalogue is not a perpetual asset in the US and models that treat it as one are wrong.
• FX distortion in Asian disclosure. WEBTOON's headline decline is −1.5% but +0.2% in constant currency. Won and yen translation is large enough to flip the sign of the story.

Genuine unknowns — and the distinction the contract asks for.

Things we don't know but could find out: the financial terms of every music AI licence (confidential, but they exist and will surface in filings or litigation); the participation and revenue of the News/Media Alliance licensing programme (undisclosed, but knowable); UMG's and Sony Music's current segment economics (published, but unreachable in this session); the signatory list for the EU GPAI Code of Practice (exists, not on the page I could fetch); AI music share on Spotify and Apple Music (measured internally, not published).

Things nobody can know yet: whether Judge Stein rules for or against fair use on a licensed-corpus theory; whether the UK ever concludes its consultation; whether AI-generated music finds genuine demand in functional categories; whether the Concord individual-officer claims survive a motion to dismiss; whether generative music is a durable consumer behaviour or a novelty with a 2021-NFT shaped adoption curve. These are not research gaps. They are unresolved facts about the future, and a trend platform should present them as branch points rather than forecasts.

One editorial-policy risk specific to this platform, recorded as instructed. This programme's standing rule — do not reproduce lyrics, poems or substantial passages from copyrighted works — is not merely an internal style guideline in this sector. It is the exact conduct at issue in Concord Music Group v. Anthropic, which concerns an AI system reproducing song lyrics. A trend platform covering this sector will routinely be asked to quote the material whose reproduction is being litigated. The rule should therefore be treated as a compliance control with a documented legal rationale, not a preference: describe and analyse, quote only short excerpts of commentary (not of works), and never reproduce lyrics or verse at any length. Phase 2 should apply this at the ingestion layer as well as the output layer, because a corpus that stores full lyrics inherits the same exposure regardless of what it emits.

12. Scenarios to 2030

Base — "licensed coexistence." The OpenAI MDL produces a mixed ruling or a settlement; the bilateral music licences extend to Sony and to the independents; generative audio settles as a creation tool inside professional workflows rather than a consumer substitute; music grows mid-single-digits on price and emerging-market volume; books grow low-single-digits with audio outperforming; local news continues to consolidate and shrink. AI licensing becomes a real but modest revenue line — low single-digit percentages of major-label revenue, consistent with Warner's $16m on $1,864m today. Falsifiable early indicator: a major label discloses AI licensing revenue as a named line item in a quarterly filing, at a scale above 2% of revenue.

Upside — "the third revenue stream." AI training and generation licensing matures into a genuine third pillar alongside streaming and sync. Terms become semi-public through filings; per-work and per-catalogue rates emerge; the publishing side converts its settlement precedent into recurring licences rather than one-off damages; detection and provenance become standard infrastructure sold to collecting societies globally. Falsifiable early indicator: two or more rightsholders disclose recurring (not settlement) AI licensing revenue in consecutive quarters, or a collecting society publishes an AI training tariff with rates attached.

Downside — "fair use wins and the leverage evaporates." Judge Stein grants summary judgment for OpenAI and Microsoft on fair use for lawfully acquired corpora. Bartz is re-read as a piracy-specific outcome with no licensing implication. The bilateral music deals are renegotiated downward or abandoned; catalogue re-rates lower on the expectation that AI output dilutes the long tail; the compensation thesis on which publishers are currently planning collapses. Falsifiable early indicator: the summary-judgment ruling itself, expected after briefing completes in early November 2026 — this is the single highest-information event in the sector's next twelve months.

Disruption — "discovery moves inside the model." The referral collapse Pew measured for news extends to music and books: people ask an assistant for something to listen to or read and never visit a catalogue surface. Whoever controls the assistant controls discovery, and the sector's platform layer — Spotify included — is disintermediated the way search disintermediated publishers. Falsifiable early indicator: a DSP reports a decline in search-and-browse-originated listening share, or a major retailer reports assistant-originated book sales as a distinct channel.

Regulatory — "the EU legislates against the incumbents' wishes." Despite IFPI's opposition, the EU adopts a statutory remuneration right or extended collective licensing for AI training, and the UK finally follows with a TDM exception plus rights reservation. Independent artists and mid-list authors get paid; major-label bilateral leverage is diluted; compliance costs rise for every developer. Falsifiable early indicator: a Commission legislative proposal — not a consultation or a code — naming remuneration or collective management for AI training.

Failure — "the fraud economy wins." Detection does not keep pace; AI catalogue floods every platform; royalty pools are diluted faster than fraud can be demonetised; identity verification fails; consumer trust in charts and playlists erodes; and the metadata layer becomes unreliable enough that rights cannot be matched to payments at all. This is the scenario in which T-24-04 turns out to matter far more than T-24-19 allows. Falsifiable early indicator: a second platform publishes AI-upload and fraud statistics that are materially worse than Deezer's, or a collecting society reports a rising unmatched/black-box share attributable to synthetic catalogue.

13. Data gaps and limitations

1. Discovery was severely constrained. Zero WebSearch calls were available; the one permitted attempt returned a budget-exhausted message. All findings come from direct primary-source fetches. Events that only trade press reports — private deal terms, executive moves, unannounced settlements — are systematically absent from this dossier.
2. Universal Music Group could not be reached. investors.universalmusic.com, universalmusic.com and umusic.com were all blocked by the egress proxy. The world's largest music company therefore has no primary financial disclosure in this dossier; its role is documented only through Udio's announcements and court dockets. This is the single largest gap.
3. News Corp / HarperCollins could not be reached. newscorp.com was blocked. The group that best exemplifies the book-plus-news merger rationale has no 2026 segment results here, and its reported AI content agreements could not be verified from any primary source.
4. Sony Music has no standalone disclosure and sony.com was blocked. Sony's music segment figures are published inside 6-K materials I could not isolate; the 6-K accessions I sampled resolved to Form 4 insider filings. Sony Music appears here only through litigation records.
5. HYBE could not be reached (403 across three URL patterns), so K-pop label economics — material to the Korean picture — are absent. Korean coverage rests on NAVER and Kakao.
6. CourtListener is robots-disallowed through this proxy, so docket documents (complaints, orders, settlement agreements) were unavailable. All litigation facts rest on Justia docket metadata plus trade-body and party descriptions. No complaint, order or settlement agreement was read in full. Phase 2 must licence RECAP/CourtListener API access — in this sector that is not optional.
7. No AI music licence has disclosed terms. Not one of Udio–UMG, Udio–WMG, Suno–WMG, Suno–BMG, Suno–Believe or KLAY–majors has published a rate, a revenue share, an advance or a term length. Any valuation of "the AI licensing opportunity" in music is therefore unfalsifiable at this date, and is marked as such throughout.
8. AI-music share is a one-platform series. Deezer is the sole source, and it monetises the detection technology its numbers justify. Spotify, Apple and Amazon publish nothing comparable. T-24-04 and T-24-19 are both capped at single_source for this reason.
9. Conflicting RIAA bases. The mid-year 2026 cycle yields both ~$6.0bn (estimated retail value) and ~$4.9bn (wholesale), and a 2025 mid-year figure of $5.6bn that is not comparable to a 2026 wholesale number. Recorded as a contradiction on T-24-03; always state the basis.
10. WEBTOON FX distortion. −1.5% reported versus +0.2% constant currency. Recorded as a contradiction on T-24-16. Do not cite either figure without the other.
11. The Authors Guild registration survey is self-selected. 500+ respondents from an advocacy body's membership. The 93% describes respondents, not the author population, and has no independent replication. T-24-08 is marked single_source accordingly — but note that no publisher-side data exists at all, which is itself informative.
12. Podcast and audiobook economics are thin here. Infinite Dial gives reach; no primary revenue disclosure for SiriusXM, iHeart, Audible or Storytel was obtainable (investor pages were JS-rendered listings). Audiobook revenue rests on AAP's $2.5bn digital-audio line alone.
13. Spotify Loud & Clear figures are rounded cumulative disclosures. "Nearly $70bn cumulatively" cannot be differenced against prior years to derive an annual series — the same trap Addendum 2 identified with YouTube's $100bn creator figure.
14. No non-English source outside Korea. Japanese webtoon and manga economics (Piccoma, Shueisha, Kodansha) are referenced only through Kakao's segment reporting. Japan is where webtoon ARPPU is highest ($22.5) and it has no direct source here.
15. The EU GPAI Code signatory list is unpublished on the page retrieved, so which AI developers have actually committed to the copyright chapter is unverified.
16. UK policy outcome is genuinely unknown, not merely unretrieved — the government itself has published no outcome more than a year after the consultation closed.
17. Blocked hosts recorded for Phase 2 procurement: investors.universalmusic.com, universalmusic.com, umusic.com, newscorp.com, sony.com, investors.wmg.com, investors.nytco.com, hybecorp.com and ir.hybecorp.com, courtlistener.com (robots), ir.webtoons.com (DNS). SEC data.sec.gov and www.sec.gov/Archives/ both work well and should be the backbone of Phase 2 collection for this sector.

14. Ranking scorecard

15. Sources

1. IFPI Global Music Report 2026 — IFPI — https://www.ifpi.org/global-music-report/ — 2026-03-18 — A
2. Music in the EU Report 2026 — IFPI — https://www.ifpi.org/wp-content/uploads/2026/09/Music-in-the-EU-Report-2026.pdf — 2026-09-01 — A
3. RIAA Revenue Reports (mid-year 2026; year-end 2025) — RIAA — https://www.riaa.com/reports/ — 2026-09-01 — A
4. AAP StatShot Annual Report, calendar year 2025 — Association of American Publishers — https://publishers.org/news/aap-statshot-annual-report-publishing-revenues-totaled-33-4-billion-for-calendar-year-2025/ — 2026-08-10 — A
5. AAP June 2026 StatShot Report — Association of American Publishers — https://publishers.org/news/aap-june-2026-statshot-report-overall-publishing-industry-up-11-0-for-the-month-of-june-and-up-3-9-year-to-date/ — 2026-09-09 — A
6. AAP Welcomes Court's Final Settlement Approval in Bartz v. Anthropic — Association of American Publishers — https://publishers.org/news/aap-welcomes-courts-final-settlement-approval-in-bartz-v-anthropic/ — 2026-07-20 — A
7. Publishers and Authors File Class Action Lawsuit Against Google (Gemini) — Association of American Publishers — https://publishers.org/news/publishers-and-authors-file-class-action-lawsuit-against-google-for-willful-copyright-infringement-to-develop-gemini-ai-models/ — 2026-07-10 — A
8. Book, News and Journal Publishers File Amicus Brief in In re Mosaic LLM Litigation — Association of American Publishers — https://publishers.org/news/book-news-and-journal-publishers-file-amicus-brief-in-in-re-mosaic-llm-litigation/ — 2026-08-27 — A
9. Copyright and Artificial Intelligence, Part 2: Copyrightability — US Copyright Office — https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-2-Copyrightability-Report.pdf — 2025-01-29 — A
10. Copyright and Artificial Intelligence, Part 3: Generative AI Training (Pre-Publication Version) — US Copyright Office — https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-3-Generative-AI-Training-Report-Pre-Publication-Version.pdf — 2025-05-09 — A
11. US Copyright Office Annual Report, Fiscal Year 2025 — US Copyright Office — https://www.copyright.gov/reports/annual/2025/ar2025.pdf — 2026-01-01 — A
12. Justia Dockets & Filings — federal copyright case index (Anthropic, OpenAI, Suno, Uncharted Labs, Concord, Databricks searches) — Justia — https://dockets.justia.com/ — 2026-09-15 — A
13. Warner Music Group Corp. Q3 FY2026 earnings release, Exhibit 99.1 to Form 8-K — WMG / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1319161/000131916126000031/q32026ex991.htm — 2026-08-05 — A
14. Amendment No. 3 to Beethoven Financing 1 LLC credit agreement, exhibit to WMG Form 10-Q — WMG / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1319161/000131916126000032/bainwmgjvco-amendmentno3to.htm — 2026-05-05 — A
15. Warner Music Group Corp. Form 10-Q, quarter ended June 30, 2026 — WMG / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1319161/000131916126000032/wmg-20260630.htm — 2026-08-05 — A
16. Spotify Technology S.A. Form 6-K, interim report Q2 2026 — Spotify / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1639920/000162828026052543/spot-20260630x6xk.htm — 2026-08-04 — A
17. Loud & Clear 2026 (2025 royalty data) — Spotify — https://loudandclear.byspotify.com/ — 2026-01-28 — A
18. AI-Generated Music Exceeds 50% of Daily New Uploads on Deezer — Deezer Newsroom — http://newsroom-deezer.com/2026/07/21/ — 2026-07-21 — A
19. Deezer H1 2026 Financial Results — Deezer SA — http://newsroom-deezer.com/2026/07/28/ — 2026-07-28 — A
20. WEBTOON Entertainment Inc. Q1 2026 shareholder letter, Exhibit 99.1 to Form 8-K — WEBTOON / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1997859/000199785926000070/wbtn-20260511xexx991.htm — 2026-05-11 — A
21. NAVER Corporation Q2 2026 earnings release — NAVER Corporation — https://www.navercorp.com/en/investment/earningsRelease — 2026-08-01 — A
22. Kakao Corporation Q2 2026 fact sheet — Kakao Corporation — https://t1.kakaocdn.net/kakaocorp/admin/ir/results-announcement/6031.pdf — 2026-08-01 — A
23. Reservoir Media investor presentation, June 2026, Exhibit 99.1 to Form 8-K — Reservoir Media / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1824403/000110465926074884/tm2618101d1_ex99-1.htm — 2026-06-17 — A
24. Reservoir Media Inc. Form 10-K, fiscal year ended March 31, 2026 — Reservoir Media / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1824403/000110465926067615/rsvr-20260331x10k.htm — 2026-05-28 — A
25. The New York Times Company Form 10-Q, quarter ended March 31, 2026 — NYT Co / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/71691/000007169126000025/nyt-20260331.htm — 2026-05-06 — A
26. USA TODAY Co., Inc. Form 10-Q, quarter ended March 31, 2026 — USA TODAY Co / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1579684/000157968426000033/gci-20260331.htm — 2026-04-30 — A
27. Introducing v6: A New Generation of Music Models, in Partnership with the Music Industry — Suno, Inc. — https://suno.com/blog/introducing-v6 — 2026-09-09 — A
28. Suno partners with Warner Music Group — Suno, Inc. — https://suno.com/blog/wmg-partnership — 2025-11-25 — A
29. Suno Announces a Landmark Global Partnership with BMG — Suno, Inc. — https://suno.com/blog/suno-partnership-bmg — 2026-08-12 — A
30. Suno partners with Believe and TuneCore — Suno, Inc. — https://suno.com/blog/believe-partnership — 2026-09-08 — A
31. Suno Raises $400M at a $5.4B Valuation — Suno, Inc. — https://suno.com/blog/series-d-announcement — 2026-06-03 — A
32. An update to our downloads policy and Terms of Service — Suno, Inc. — https://suno.com/blog/suno-updates-tos — 2026-08-10 — A
33. Udio blog: A New Era of Music — Udio with Universal Music Group (2025-10-29); Udio with Warner Music Group (2025-11-19) — Uncharted Labs, Inc. — https://www.udio.com/blog — 2025-11-19 — A
34. Survey Results Detail Publishers' Failure to Register Authors' Copyrights — The Authors Guild — https://authorsguild.org/news/survey-results-anthropic-registration-failure/ — 2026-08-20 — A
35. Authors Guild and Co-Plaintiffs File Motion for Summary Judgment v. OpenAI and Microsoft — The Authors Guild — https://authorsguild.org/news/plaintiffs-file-motion-for-summary-judgment-v-openai-and-microsoft/ — 2026-09-05 — A
36. Authors Guild artificial intelligence advocacy — The Authors Guild — https://authorsguild.org/advocacy/artificial-intelligence/ — 2026-09-15 — B
37. Bartz v. Anthropic PBC settlement administration site (N.D. Cal. 3:24-cv-05417) — Court-appointed settlement administrator — https://www.anthropiccopyrightsettlement.com/ — 2026-07-20 — A
38. Google users are less likely to click on links when an AI summary appears in the results — Pew Research Center — https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/ — 2025-07-22 — A
39. The Infinite Dial 2026 — Edison Research and SSRS — https://ssrs.com/insights/the-infinite-dial-2026/ — 2026-03-12 — A
40. General-Purpose AI Code of Practice and Article 53(1)(d) training-data summary template — European Commission, DG CNECT / AI Office — https://digital-strategy.ec.europa.eu/en/policies/ai-code-practice — 2025-08-01 — A
41. Copyright and Artificial Intelligence — consultation — UK DSIT / Intellectual Property Office — https://www.gov.uk/government/consultations/copyright-and-artificial-intelligence — 2026-03-19 — A
42. News/Media Alliance artificial intelligence advocacy and AI Licensing Program — News/Media Alliance — https://www.newsmediaalliance.org/issues/artificial-intelligence/ — 2026-09-15 — B
43. Copyright Royalty Board — final rules for digital performance of sound recordings, 2026–2030 — Copyright Royalty Board — https://www.crb.gov/ — 2026-03-13 — A
44. Bartz v. Anthropic Copyright Settlement — FAQ for authors; corporate announcements — Penguin Random House — https://www.penguinrandomhouse.com/bartz-v-anthropic-copyright-settlement-faq-for-authors/ — 2026-09-15 — A
45. WEBTOON Entertainment Form 8-K, 2026 annual meeting results — WEBTOON / SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1997859/000199785926000082/wbtn-20260603.htm — 2026-06-04 — A

Tier summary: 42 Tier A, 3 Tier B, 0 Tier C. No content-farm or vendor market-report source is cited anywhere in this dossier, and no market-size figure originates from one.

More research in this sector · Explore trend records