Anthropic to pay $1.5B in biggest US copyright deal

A US federal court has approved Anthropic’s $1.5 billion payout in a copyright case over training data for its AI models, marking the largest copyright settlement in US history. The money will be distributed across roughly 500,000 works, which works out to about $3,000 per book on average.

Writers Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson sued the company in 2024, accusing Anthropic of using copyrighted books without permission to train Claude. According to the authors, the company’s business model relied in part on the “mass theft of copyrighted works.”

The court found that Anthropic had illegally obtained millions of books through the shadow libraries LibGen and PiLiMi. But it also ruled separately that using books to train AI models qualifies as fair use under US copyright law.

This case was the first class action against a company in the AI copyright dispute to receive that status.

Recommended reading

Musk says Grok Imagine could make an Odyssey film in 2026

Anthropic agreed to the settlement in August 2025, a month after arguing that such an outcome could be a “death blow” to the company. The court issued preliminary approval at the end of September 2025.

Under the approved terms, compensation will go to owners of works found in LibGen and PiLiMi. The court said the roughly $3,000 payout per work is about four times the statutory damages level for willful copyright infringement.

Aparna Sridhar, Anthropic’s deputy general counsel, said the company was pleased to put the matter behind it. She added that more than 91% of authors and publishers covered by the settlement have already said they want to claim their share.

The financial hit is still relatively modest for Anthropic. After its $13 billion Series F round, the company was valued at $183 billion, and it is now preparing for an IPO at a valuation of about $1 trillion.