Sony and Warner Chappell Sue Anthropic — Lawsuit No. 13 Brings All Three Majors' Publishers to Court

The legal escalation around Anthropic hit a symbolic milestone: on August 28, 2026, Sony Music Publishing and Warner Chappell Music jointly sued Anthropic — along with co-founders Dario Amodei and Benjamin Mann personally — in the U.S. District Court for the Northern District of California (case candce-26-09217, filed at 5:04 PM court time per the docket). With this filing, the publishing arms of all three major music companies are now litigating against the maker of Claude: Universal, Concord and ABKCO since 2023, BMG in March 2026, Round Hill on August 17, and now Sony and Warner Chappell.

What the Suit Contains

Four counts: direct infringement by torrenting against all three defendants; contributory infringement by torrenting against Amodei and Mann individually; direct infringement against Anthropic; and removal or alteration of copyright management information (CMI) against Anthropic. The publishers seek statutory damages of up to $150,000 per willfully infringed work and up to $25,000 per alleged CMI removal, plus a jury trial, destruction of infringing copies, and a full accounting of Claude’s training data.

The complaint describes “tens of thousands” of the publishers’ works as infringed — a characterization that puts theoretical statutory exposure in the multi-billion-dollar range, per Music Business Worldwide, which obtained and published the full 48-page complaint.

Where the Evidence Comes From

This case does not start from scratch: it leans heavily on material unsealed in Bartz v. Anthropic — the authors’ case that ended in a $1.5 billion settlement in September 2025 over the same conduct, in which a judge in the same district described Anthropic’s behavior as “straightforward piracy but at massive scale.” The filing states that Mann downloaded at least five million pirated books from Library Genesis via BitTorrent in June 2021, that employees torrented at least two million more from Pirate Library Mirror in July 2022, and quotes Mann’s unsealed description of LibGen as “sketchy AF” — alongside a 2024 internal planning document: “We don’t want it to be known that we are working on this.”

On the lyrics side, the publishers allege Anthropic scraped from licensed sites including MusixMatch and LyricFind, ran a “destructive scanning” operation on second-hand books, and used Common Crawl, The Pile and Books3. They further allege Claude reproduces lyrics verbatim in outputs, and that the guardrails added after the earlier litigation are “easily circumventable by simply ‘re-prompting’” the model.

The Symbolism, and the Economics

The symbolism is blunt: the company that built its brand as the “ethical AI company” is now the most-sued AI lab on copyright — this is lawsuit number thirteen against Anthropic by court-archive count. But the economics are what make the case more than headlines: the complaint argues the $1.5 billion settlement was “obviously not a large enough settlement to deter infringing conduct by a company that has parlayed such mass infringement into a staggering $2-trillion-dollar valuation” — citing an August 2026 Forbes report on a projected October IPO. If large settlements read as an operating cost, the publishers are now testing sharper instruments: per-work statutory damages, personal liability for founders, and compelled training-data disclosure.

What This Means for AI Teams

Training-data provenance is now a legal asset, not an engineering preference: this suit is built on provable download archives. Teams building or fine-tuning models need verifiable source records — not just for transparency, but as a defense.
The Bartz precedent is compounding: every new filing borrows the unsealed evidence of the last one, making each successive case cheaper to file and harder to defend. The legal calibration of training-data risk shifted after the settlement.
Licensing is the market’s chosen path: the complaint itself acknowledges publishers “have entered licenses permitting the authorized use” — the fight is over terms of access to data, not AI itself. Rights-clearance infrastructure is about to be the profitable layer.
Personal liability is a dangerous precedent: naming the founders for contributory infringement raises the stakes from corporate risk to personal exposure — a detail every lab’s data-governance process will now price in.

What Happens Next

Expect the case to be consolidated with the existing Anthropic copyright docket — trackers note it shares counsel and is likely headed to the same judge as Concord v. Anthropic II. The first signal to watch: whether this follows the Bartz path toward a large settlement, or produces a ruling that actually defines fair use boundaries for model training — the decision that will sort the market into licensing tracks and download risk.