🔍 Read the full analysis: Anthropic IPO Questions: Was Claude Trained On Pirated Books? on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Anthropic’s October 27, 2026 IPO filing comes after a 2025 court ruling and a $1.5 billion settlement over pirated book copies. The judge found that training Claude on books was fair use but ruled that downloading and retaining pirated copies was not; Anthropic says the pirated collections were not in commercially released models’ training data. New opt-out authors’ lawsuits remain unresolved.
Anthropic filed to go public on October 27, 2026, putting renewed attention on its copyright exposure after a federal judge found that the company had downloaded millions of pirated book copies. The 2025 ruling did not find that training Claude on books was illegal: Judge William Alsup held that training was fair use, while finding that keeping pirated copies in a permanent library was not. Anthropic later agreed to a $1.5 billion settlement in the authors’ case, while denying wrongdoing and disputing that the pirated collections were used to train commercially released models.
The distinction between copying books and training on them is central to the record. In Bartz v. Anthropic, a case brought by three authors in 2024, Alsup’s June 23, 2025 order said training Claude on books was fair use. He also found that Anthropic’s purchase and scanning of millions of used print books was fair use. The court treated downloading books from pirate libraries and retaining them as a separate issue: the order described more than seven million copies from Books3, Library Genesis and the Pirate Library Mirror, and found that the copying was not excused by fair use.
The ruling was a pretrial decision, not a jury verdict on damages. On the pirated copies, the judge accepted the authors’ account of the relevant facts; the case was later resolved through a settlement. The agreement, finally approved in July 2026, provides $1.5 billion for owners of roughly 500,000 registered books identified as downloaded from LibGen or PiLiMi—about $3,000 per book before fees. It also requires Anthropic to destroy the downloaded copies, subject to legal-hold obligations.
Anthropic’s settlement position adds a separate qualification. The company represented to the court that neither the LibGen and PiLiMi datasets nor portions of them were in the training corpus of any commercially released large language model. The source account says the court record indicates portions of the pirated collections were used to train some Anthropic models. Those statements address different questions: what materials the company copied and used in model development, and what it says was included in commercially released models’ training data. The settlement does not license future use of copyrighted books, and it does not release claims about Claude’s outputs.
From download to IPO
Five years, one ruling, no trial
The piracy happened in 2021 and 2022. The only ruling came in 2025. The new lawsuits are years away from a verdict.
- Jun 2021 · Jul 2022DownloadsAt least 5 million copies from LibGen (2021), at least 2 million from PiLiMi (2022); 196,640 from Books3 earlier.
- Aug 2024Bartz v. Anthropic filedThree authors sue Anthropic, on behalf of a class of book owners.
- 23 Jun 2025Judge Alsup's orderTraining on books: fair use. Scanning bought books: fair use. Pirated library copies: not fair use, to trial.
- Sep 2025$1.5 billion settlementNo trial. Anthropic denies wrongdoing; release covers conduct up to 25 Aug 2025 only.
- 13 + 15 May 2026Opt-out suitsCruz (28 plaintiffs) and Kwon (incl. Chang-rae Lee) join Cambronne and two publisher suits.
- Jul 2026Final approvalSettlement approved. Anthropic: no opt-out deal "above the Bartz claim value".
- 24 Sep 2026Motion to dismissAnthropic: claims too late, no infringing output shown, many works already released.
- 26–27 Oct 2026NowS-1 and the headlineIPO filing cites Bartz as a risk; The Stanford Daily names Stanford authors.
- 17 Dec 2026ScheduledHearingCourt hears the motion to dismiss.
- May 2027 · Jan 2028ScheduledDiscovery, then trial settingFact discovery closes 13 May 2027; trial-setting conference 11 Jan 2028.
Sources: Bartz v. Anthropic, order of 23 June 2025; settlement FAQ; joint case statement, 27 Aug 2026; Anthropic letter, 27 Jul 2026; case management order, 14 Sep 2026; Anthropic S-1, 26 Oct 2026; The Stanford Daily, 27 Oct 2026. Scheduled dates can move.
Powered by Thorsten Meyer AI
The Copyright Risk Behind the Filing
The filing places a costly copyright dispute in view as investors assess Anthropic’s legal and financial risks. The settlement’s size gives a concrete measure of the potential cost of one class of claims, but it does not establish that every use of copyrighted material to train AI is unlawful. The judge’s distinction—between fair-use training and unlawful acquisition or retention of pirated copies—makes the legal question more specific than headlines suggesting a court found Claude’s training itself illegal.
The issue also matters to authors and publishers whose work may be copied into datasets, and to AI companies deciding how to obtain and document training materials. The settlement covers past conduct through August 25, 2025, according to the source account, while explicitly leaving claims about model outputs outside its release. That means it does not resolve every dispute over how AI systems use or reproduce protected work.
For readers evaluating the IPO news, the filing and the lawsuits should not be conflated. A company’s risk disclosure is not a court finding, and a settlement is not an admission of wrongdoing. Here, the court’s fair-use ruling, its finding about pirated copies, Anthropic’s statements to the settlement court and the newer authors’ allegations each have different legal status.
Popular right now
KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream
Dreame L50 Ultra Robot Vacuum and Mop
Vitamix 5200 Professional-Grade Blender, 64 oz. Container, Black
As an affiliate, we earn on qualifying purchases.
Bartz v. Anthropic began in August 2024, when three authors sued over the company’s use of their books. Alsup’s June 2025 order separated the lawful acquisition and use of books for training from the downloading and permanent storage of pirated copies. The case did not proceed to a jury trial on damages; Anthropic and the authors reached a settlement that received final approval in July 2026.
Separate lawsuits filed in May 2026 by authors who opted out of the settlement have not been decided. A report by The Stanford Daily said that works by Stanford-affiliated writers, including Tobias Wolff, Abraham Verghese, Chang-rae Lee and Daniel Mason, were among the books involved. That list is the newspaper’s reporting, not a finding in the court ruling described here. The source account also says Chang-rae Lee appears as a plaintiff in one new case, while it could not verify Wolff’s name among plaintiffs in five related cases as of August 2026.
The IPO filing arrived amid online claims about other language in Anthropic’s prospectus, including assertions that it warned Claude could blackmail people or become sentient. The source account says the filing discusses model behavior resembling blackmail in a staged safety test, the possibility of regulation tied to debate about AI sentience, and existential concerns raised by others. Those disclosures are not findings that Claude is sentient or that it will harm people.
Claims Still Awaiting a Court Ruling
The May 2026 opt-out authors’ suits remain pending. The source account says a hearing on Anthropic’s motion to dismiss is scheduled for December 17, 2026; there is no ruling yet on those claims. The precise scope of the plaintiffs’ allegations and which authors are included in each action may change through court filings.
There is also a factual distinction that cannot be settled by repeating either side’s summary. The 2025 order addressed the copying and retention of pirated books and ruled on training as fair use; Anthropic’s settlement representation says the pirated collections were not part of any commercially released model’s training corpus. The source account characterizes the court record as showing portions of the collections were used to train some models. The materials provided do not fully reconcile what was used during development with what was included in released-model training data.
The reported book-by-book list has not been independently checked against the settlement database in the source account. Nor does the available material establish the outcome of the new lawsuits, whether any additional authors have joined them, or how a court will treat claims about outputs that the settlement did not release.
December Hearing and IPO Review
The next listed court milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss the opt-out authors’ claims. The court’s decision could determine whether some or all of those claims proceed, but it would not by itself resolve every copyright question involving AI training or outputs.
Investors and readers can also review Anthropic’s S-1 as the IPO process moves forward. The filing’s risk disclosures should be read as the company’s account of potential exposure, not as proof that a disputed allegation is true. Further court orders, amendments to the lawsuits or details about the settlement’s administration could clarify the remaining questions.
For now, the established record is narrower than some viral summaries suggest: a judge found training on books to be fair use, found downloading and retaining pirated copies unlawful, and the resulting case ended in a large settlement. The newer claims are still allegations, and the court has yet to rule on them.
Key Questions
Did a judge rule that Claude was illegally trained on books?
No. In June 2025, Judge William Alsup ruled that training Claude on books was fair use. He separately found that downloading and retaining pirated book copies was not protected by fair use.
How much did Anthropic agree to pay in the book settlement?
The settlement provides $1.5 billion for owners of roughly 500,000 registered books identified as downloaded from LibGen or PiLiMi, or about $3,000 per book before fees. It was finally approved in July 2026.
Does the settlement prove Anthropic admitted wrongdoing?
No. Anthropic denied wrongdoing while agreeing to the settlement. A settlement resolves claims; it is not itself a court finding that the company admitted liability.
What happens to the newer authors’ lawsuits?
The May 2026 opt-out cases remain unresolved. The source account lists a hearing on Anthropic’s motion to dismiss for December 17, 2026; the court has not yet ruled on the claims.
Source: ThorstenMeyerAI.com
