Via pitchbook.com
The largest copyright settlement in US history gives authors a payout but hands AI companies a legal playbook for using purchased books without ongoing fees.
Anthropic just wrote a $1.5 billion check to authors whose books were used to train Claude. The literary world is calling it a win. Here’s the thing: the same ruling that punished Anthropic for pirating books also gave every AI company a clear legal path to use purchased texts for training, no licensing fees required, no ongoing royalties, no need to ask permission twice.
What the ruling actually says #
On July 20, 2026, US District Judge Araceli Martínez-Olguín granted final approval to the class-action settlement in Bartz v. Anthropic. The lawsuit, filed in 2024 and led by authors Andrea Bartz and Kirk Wallace Johnson, alleged that Anthropic used millions of copyrighted books sourced from shadow libraries to train its flagship AI model.
The core finding drew a bright line. Using pirated materials to train AI models is copyright infringement. But using lawfully acquired texts? That’s fair use.
The settlement itself is staggering by any measure. Roughly $1.5 billion will be distributed across an estimated 500,000 titles, working out to approximately $3,000 per book. That makes it the largest known copyright settlement in US history. Only about 350 authors opted out of the class.
Anthropic had agreed to the $1.5 billion minimum payout back in September 2025, following a summary judgment ruling in June of that year that found the company liable for its use of pirated source material.
Why authors aren’t popping champagne #
A $3,000 check for a book that took years to write is not exactly life-changing money. The ruling effectively creates a two-tier system. Piracy is punished. Legal acquisition is blessed. For authors, this means that once a publisher sells a copy of their book to an AI company, the text can be ingested, processed, and used to generate competing content, all without any recurring compensation.
The precedent doesn’t require AI firms to negotiate ongoing licensing deals. It doesn’t mandate revenue sharing. It doesn’t even require disclosure of which specific works were used in training.
What this means for investors #
The immediate market implication is straightforward. Anthropic absorbed a $1.5 billion hit and emerged with legal clarity that makes future training cheaper and more predictable. For AI companies, the cost of compliance just dropped significantly. Buy the books, train the models, move on.
For the broader AI sector, this precedent reduces legal risk around training data acquisition. That should accelerate partnerships between AI firms and publishers, potentially creating a new revenue stream for the publishing industry even as individual author compensation remains thin. There’s also the question of whether $3,000 per book becomes the implicit benchmark for future settlements. If so, the economics of AI training just got a price tag, and it’s remarkably cheap relative to the value these models generate.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our