On 20 July 2026 a federal judge in San Francisco signed the largest copyright settlement in US history. Anthropic will put $1.5 billion into a fund for authors and publishers whose books sat in a pirate library the company used to train Claude. Reuters called it the record. Three authors had filed the case. About half a million titles ended up in the class.
The same week, the New York Times was still in discovery against OpenAI and Microsoft. Variety, citing the paper's own figures, put the Times's legal spend above $28 million, $4.2 million of that in one quarter. No verdict. No cheque from the other side.
Filed is a complaint. Decided is a holding. Paid is cash you can see, or a licence with a number. Coverage mashes the three into "the lawsuits." That word flatters everyone.
The AI Lawsuit Tracker snapshot of 1 June 2026 counted 200 cases, 68 defendants, 130 of them copyright, and more than $6 billion in disclosed stakes. I did not treat the $6 billion as money that changed hands. Most of it is a demand in a complaint.
Bartz. The pirate library, not the training
Bartz v. Anthropic, N.D. Cal. 3:24-cv-05417, started with Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson. They did not need a court to say that training is always illegal. They needed a court to look at LibGen and Pirate Library Mirror.
In June 2025 Judge William Alsup split the case. Training on lawfully acquired books was "quintessentially transformative" and fair use. Building a central library from torrented files was not. That split is why the case settled instead of becoming a clean win for the labs.
Anthropic agreed in August 2025 to $1.5 billion, about $3,000 a title after fees, on a class of roughly 500,000 works. The Authors Guild published the installment plan: $300 million after preliminary approval, $300 million after final approval, $450 million by September 2026, more later. Claims closed 30 March 2026. Judge Araceli Martínez-Olguín granted final approval on 20 July.
The release covers past conduct through 25 August 2025. It does not license the next training run. It does not cover outputs. A lab that wants more books still has to buy them or take another complaint.
$1.5 billion is the record. It is also small next to Anthropic's last published valuation. Both can be true. The first sentence is about authors. The second is about whether the number changes how labs ingest text.
Kadrey. Same month, other judge, other pirate result
Kadrey v. Meta, N.D. Cal. 3:23-cv-03417. Thirteen authors, including Richard Kadrey. Llama, not Claude. Judge Vince Chhabria granted Meta summary judgment on 25 June 2025, two days after Alsup's split.
Chhabria also called the training transformative. He put the weight on factor four, market harm. The authors, he wrote, had not shown a real market for licensing books into training, and Llama did not spit their books back out. On that record he treated even the pirate-site copies as fair use. Then he wrote a long warning that a better-pleaded market-dilution case could go the other way. Debevoise and the Authors Alliance both read it as a win with an asterisk.
Two Northern District judges. Same month. Same subject. Alsup: the library is infringement. Chhabria: not on this evidence. Neither is the Ninth Circuit.
Ross. The first "no"
Thomson Reuters v. Ross Intelligence, D. Del. 1:20-cv-613, is not a chatbot case. Ross built a competing legal-research tool and trained on Westlaw headnotes. Judge Stephanos Bibas granted Thomson Reuters partial summary judgment on 11 February 2025. Ross copied 2,243 headnotes. The use was commercial. It was not transformative because the point was a substitute for Westlaw. Market harm went to Thomson Reuters.
That is the first federal holding that AI training is not fair use. It is also a headnote case, not a novel case. The Third Circuit took the appeal in June 2025 (No. 25-8018). Until Philadelphia rules, Bibas is one district judge and Alsup is another.
The Times. Money out, nothing back
The New York Times Company v. Microsoft and OpenAI, S.D.N.Y. 1:23-cv-11195, filed 27 December 2023. The complaint asked for billions. Judge Sidney Stein let the core copyright claims through in March 2025 and cut some DMCA counts. The Daily News, the Center for Investigative Reporting, the Intercept, and Ziff Davis sit in the same cluster. Summary judgment was on the 2026 calendar.
What has actually happened is discovery. In May 2025 the court ordered OpenAI to keep ChatGPT logs. In January 2026 Stein affirmed a 20 million log sample. In July 2026 the publishers moved for sanctions, saying OpenAI hid its ability to search training sets and output logs for two years.
The $28 million is paid. It is paid by the plaintiff, to lawyers. It is not a recovery. A company that has already written that cheque will not walk away cheap. It also has not produced a holding on whether training on news is fair use.
Getty lost the weights theory. GEMA won the lyric
Getty sued Stability AI in London. Training had happened outside the UK, so Getty dropped the primary infringement count. What remained: importing the model into Britain was importing an infringing copy.
Mrs Justice Joanna Smith decided Getty Images v Stability AI, [2025] EWHC 2863 (Ch), on 4 November 2025. The model is an "article." The weights, on the facts the parties agreed, do not store Getty's photographs. They are not an infringing copy under the CDPA. Getty kept a narrow trademark win on watermarks that leaked into outputs. That is a decided cell. It is not a US holding. It is not a win on training. It is a loss on the theory that a .safetensors file is a stack of stolen pictures.
Germany's collecting society GEMA beat OpenAI at LG München I in November 2025, case 42 O 14139/24. Memorized German lyrics in GPT-4 were reproduction. The TDM exception did not save the lab. The case is on appeal. A first-instance European holding is not the CJEU. Anyone training on music in the EU now has a judgment to brief.
Music closed a docket without a number
The major labels sued Suno and Udio in 2024. Warner settled with Udio in April 2026 as a "strategic partnership." Terms confidential. UMG and Sony stayed in against Suno. A confidential partnership can include cash, a licence, equity, or a product credit. From outside the room you cannot put a dollar in the paid column. You can only say the docket went quiet.
That is the same trick as calling a memorandum "investment." A closed case is not a priced licence.
Authors Guild v. OpenAI, filed September 2023, is still live. George R.R. Martin and David Baldacci are in the caption. Stein let output-infringement claims survive in October 2025. No class cheque. Reddit sued Perplexity in April 2026, S.D.N.Y. 1:25-cv-08713, on RAG outputs and database rights. A retrieval case, not a pretraining case. Filed is still cheap.
The money that moved sat in contracts
Subtract Bartz and the paid column for training-data copyright is almost empty. Warner-Udio is empty on purpose.
The cheques that did clear were private. The Wall Street Journal put OpenAI's May 2024 News Corp deal at more than $250 million over five years. That is a reported figure, not a filed close. It covers the Journal, Barron's, the Times of London, the New York Post, and Australian titles. Axel Springer signed in December 2023. The companies did not disclose terms. Nieman Lab later wrote the deal was "said to be worth about $10 million per year." AP was first, summer 2023, terms undisclosed. Reddit-Google was reported around $60 million a year for API access. A June 2026 census counted 48 news licences against 16 music and 12 image, with OpenAI far ahead on volume.
Those are paid. They are also private. They do not bind the Times, the Authors Guild, or GEMA. Mixing them into "the lawsuits" hides the fact that the people who sued still have not been paid, except the Bartz class.
Five holdings, one industry
Alsup: lawful books, fair use; pirate library, not. Chhabria: training, including from pirate sites, fair use on this record. Bibas: competing legal search, not fair use. Smith: weights, not copies. Munich: memorized lyrics, copies.
A lab can win London, lose Munich, settle San Francisco, and still face Stein in the same year. Anyone who says "the courts have decided AI training" is collapsing three cells and three countries.
The practical line after Bartz is narrower than the headlines. Do not torrent LibGen into a central library. Buying the book and training on it is, in one US district, fair use. Output that recites the book, or the lyric, is a separate count. Stein and Munich are not on the labs' side.
What would move the ledger: a Times judgment or a disclosed settlement with a number. A Ninth Circuit or Third Circuit opinion. A CJEU ruling on TDM and memorization. A music deal that publishes the fee. Until then the industry has one historic cheque, one English holding that helps defendants, one German holding that helps collecting societies, and a Times bill for $28 million that has not bought a verdict.