Analysis
Sony Music Publishing and Warner Chappell Music sued Anthropic in federal court in Northern California, alleging the company illegally downloaded, scraped and torrented "thousands upon thousands" of copyrighted songs to train its Claude models, Fortune reported. The suit names CEO Dario Amodei and co-founder Benjamin Mann as individual defendants alongside the company itself -- a more aggressive framing than a corporate-only suit, since it puts named executives' own conduct directly at issue.
The complaint alleges Anthropic obtained lyrics and sheet music through pirate sources including Library Genesis and the Pirate Library Mirror, and separately scraped licensed lyric websites such as Musixmatch and LyricFind without authorization. Named works in the suit include the Beatles' "I Am the Walrus" and Leonard Cohen's "Hallelujah," among what the complaint characterizes as thousands of copyrighted compositions. Sony and Warner Chappell are seeking a jury trial and statutory damages of up to $150,000 per infringed work, plus $25,000 for each instance in which Anthropic allegedly removed copyright management information -- a damages structure that, applied across "thousands upon thousands" of works, could theoretically total in the billions of dollars, Variety reported.
Part of a pattern, not an isolated suit
This is not the first major music-industry action against Anthropic this year -- Pulse has tracked Anthropic's mounting legal and safety disclosures throughout 2026. BMG filed a similar suit in March, alleging Anthropic used lyrics from Justin Bieber, Bruno Mars, Ariana Grande and the Rolling Stones to train Claude, and claiming Anthropic never responded to a December 2025 cease-and-desist letter or engaged in licensing discussions. BMG's suit also seeks disclosure of Anthropic's training data and methods, not just damages -- a request that, if granted in either case, could force Anthropic to reveal training-data specifics it has resisted disclosing publicly.
The competitive and legal landscape
AI labs across the industry face a growing stack of copyright litigation over training data, and music rights holders have proven especially aggressive litigants given how clearly a specific song's use can be identified compared to, say, a scraped news article. Anthropic isn't alone in this fight -- OpenAI, Stability AI and others have faced comparable suits from publishers, authors and image-rights holders -- but the combination of two major music suits in six months, with named individual executive defendants in the newer one, signals plaintiffs' lawyers see a strengthening legal theory and are moving to lock in claims before any settlement framework emerges industry-wide.
Counterweight
A $150,000-per-work statutory maximum multiplied across "thousands upon thousands" of alleged infringements is a theoretical ceiling, not a likely outcome -- courts rarely award maximum statutory damages across every claimed instance in cases this large, and Anthropic has not yet filed its formal response. The company has previously argued that training on copyrighted material can qualify as fair use, a defense that has had mixed results across different AI copyright cases so far and remains genuinely unsettled law.
What to watch next is whether Anthropic's formal response follows the same fair-use defense strategy it and other labs have used in prior suits, and whether the individual-executive-defendant structure in this complaint becomes a template other rights holders adopt against AI labs going forward.
