Sony and Warner Take Anthropic to Court Over AI Music Copyright
Two of the world's biggest music publishers want billions from Anthropic, alleging its models trained on their songs without a licence. India's own AI-copyright reckoning is already under way.
The News
Sony Music Publishing and Warner Chappell Music, two of the largest music publishers in the world, have filed a copyright suit against Anthropic in the US District Court for the Northern District of California. The publishers allege that the AI company trained its models on "tens of thousands" of copyrighted songs without securing a licence.
The financial exposure is severe. The publishers are seeking up to $150,000 for each infringed work, the statutory ceiling for willful infringement under US law. On top of that, they want as much as $25,000 for every instance in which identifiable copyright management information was stripped from a track. Multiplied across tens of thousands of works, the potential damages run to several billion dollars if a court awards the maximum.
Anthropic, the developer of the Claude family of models, has fielded accusations of this kind before. The new complaint is a fresh escalation in the long-running dispute between the recorded-music business and the generative-AI industry over who owns the material that trains these systems.
Why It Matters
This is not the first time the music industry has taken aim at Anthropic. Back in 2023, Concord, Universal Music Publishing and ABKCO sued the company in Tennessee over Claude reproducing copyrighted song lyrics, one of the earliest music-copyright challenges aimed at a frontier AI lab. What has changed is the ambition of the claim and the size of the potential cheque.
The timing also matters. In September 2025 Anthropic agreed to pay roughly $1.5 billion to settle a class action brought by authors whose books had been used in training. That settlement signalled that AI developers can no longer assume training data is free, and it handed rights-holders a template for pressing similar demands. A win for Sony and Warner would harden that precedent and push the whole industry towards paid licensing rather than scrape-first, argue-later.
Indian Angle
India is running its own version of this fight. In the Delhi High Court, news agency ANI Media sued OpenAI in late 2024 over the alleged use of its content for training, and book publishers and news bodies have since sought to join the proceedings. The Sony-Warner action gives Indian rights-holders fresh ammunition and a richer set of arguments to borrow.
The legal footing here is narrower than in the United States. India's Copyright Act, 1957 offers only a limited "fair dealing" exception under Section 52, not the broad American "fair use" doctrine that AI firms lean on. That makes an unlicensed-training defence harder to sustain in an Indian courtroom, a point that labels such as Saregama, T-Series and Sony Music India, along with the Indian Music Industry body, will be watching closely.
There is a builder's angle too. Home-grown model developers like Sarvam AI and Krutrim need large volumes of Indian-language text and audio. If global litigation entrenches licensed data as the norm, Indian startups will have to budget for content deals from day one, opening a market for local rights-holders to sell what they own.
FAQ
Who is suing Anthropic?
Sony Music Publishing and Warner Chappell Music, two of the world's biggest music publishers, filed the suit in the US District Court for the Northern District of California. They accuse Anthropic of training its AI models on tens of thousands of copyrighted songs without permission.
How much money is at stake?
The publishers are seeking up to $150,000 per infringed work plus up to $25,000 for each removal of copyright management information. Across tens of thousands of works, the maximum award could reach several billion dollars, though courts rarely grant the full statutory ceiling.
Has Anthropic been sued over content before?
Yes. Music publishers including Concord and Universal sued it over song lyrics in 2023, and in 2025 the company agreed to pay about $1.5 billion to settle an authors' class action over books used in training. This latest suit continues that pattern.
What does this mean for Indian AI startups?
Indian model builders such as Sarvam and Krutrim may increasingly need licensed training data. With India's narrower "fair dealing" law and the ongoing ANI v OpenAI case in the Delhi High Court, the cost of unlicensed scraping is rising for domestic developers too.
Where can I read the original report?
The full coverage is published by The Verge, linked in the source paragraph below.
This story was reported by The Verge. Read the full original coverage at The Verge.