🔍 Read the full analysis: Anthropic’s IPO In View: What The Claude Book Claims Could Mean on ThorstenMeyerAI.com
Get smart everyday buys 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 has filed to go public as scrutiny continues over its use of books to train Claude. A 2025 court ruling found that training on books could be fair use but that downloading millions of pirated copies was not; a later $1.5 billion settlement resolved claims by a defined group of authors. New lawsuits and claims about what Anthropic’s filing says remain separate matters.
Anthropic filed to go public on October 27, bringing renewed attention to lawsuits over books copied for Claude’s development and claims about the company’s IPO filing. The record is more qualified than some headlines suggest: a judge found in 2025 that training on books could be fair use, but ruled that downloading pirated copies to keep was not. New author claims have not been decided, and the company’s filing should not be read as saying Claude is sentient or will blackmail people.
The central court ruling came in Bartz v. Anthropic, a case brought by three authors in 2024. On June 23, 2025, U.S. District Judge William Alsup ruled that using books to train a language model was fair use in the circumstances before him. He also found that Anthropic’s purchase of print books, followed by scanning them, qualified as fair use. But he drew a distinction between those activities and downloading pirated copies for a permanent library.
Alsup’s order described more than seven million book copies from sources including Books3, Library Genesis and the Pirate Library Mirror. The judge concluded that keeping pirated copies was not excused by the fair-use finding for model training. The ruling was a pretrial decision, not a jury verdict on damages. The parties later settled, so a jury did not decide the remaining issues.
The settlement was approved in July 2026 for $1.5 billion, with payments tied to roughly 500,000 registered books. That works out to about $3,000 per book before fees, according to the source account. Anthropic denied wrongdoing and represented in settlement proceedings that the LibGen and PiLiMi collections, or portions of them, were not in the training corpus of any commercially released model. The settlement covers specified past conduct; it does not grant a general license to use books for training or 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
Copyright Exposure Behind the IPO
The cases put a consequential question for AI companies in public view: what permissions are needed to obtain and use copyrighted works when building commercial models? The record in Anthropic’s dispute separates the legal treatment of training from the acquisition of training material. A finding that a particular training use was fair does not automatically resolve whether the underlying copies were lawfully obtained.
That distinction matters to authors, publishers, investors and other AI developers. The settlement provides compensation for a defined set of registered books and requires destruction of the covered pirated collections, subject to legal-hold obligations. It does not settle every copyright question about generative AI. Pending cases and future disputes could test different facts, claims or uses, while the IPO process may bring closer attention to how Anthropic describes legal and operational risks.
There is also a difference between the court record and viral interpretations of the filing. According to the source account’s review of Anthropic’s S-1, it discusses model behavior resembling blackmail in a staged safety test, debate about possible AI sentience, and concerns others have raised about existential risks. Those disclosures are not findings that Claude is sentient or has blackmailed a real person. Accurate attribution matters because risk disclosures describe possibilities and debate, not necessarily established events.
Products Worth Considering
Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit
Logitech MX Mechanical Mini Mac Wireless Keyboard Low-Profile – Space Grey
HP Sprocket Panorama Portable Color Photo and Label Printer
As an affiliate, we earn on qualifying purchases.
What the 2025 Order Decided
In the 2025 order, Alsup distinguished three activities: training Claude on books, scanning books Anthropic had bought, and retaining pirated copies. He found the first two fair use on the record before him, while ruling against Anthropic on the third. The order said Anthropic had downloaded millions of copies and that later buying a book would not erase liability for the earlier theft. Because the case settled, the court did not reach a jury determination of damages.
The $1.5 billion settlement addresses claims tied to books downloaded from LibGen and PiLiMi, with payments scheduled in instalments. It releases specified claims relating to conduct up to August 25, 2025, according to the source account, but expressly does not release claims about outputs from Claude. Anthropic’s statement that the pirated collections were not in commercially released models’ training corpora is its position in the settlement, not a new court finding that resolves every question about earlier model training.
Separate lawsuits were filed by authors who opted out of the settlement. The source reports that one case names author Chang-rae Lee and that a motion-to-dismiss hearing is scheduled for December 17, 2026. It also says The Stanford Daily reported that books by several Stanford-affiliated authors were among the copies at issue. The book list and the status of individual authors should be treated as reported claims unless confirmed in the relevant filings or settlement records.
“The LibGen and PiLiMi datasets, or portions of them, were not in the training corpus of any commercially released large language model.”
— Anthropic, as represented in the settlement proceedings
Open Questions in the Book Cases
The new opt-out lawsuits remain unresolved. The source account reports that no judge ruled on those claims in May 2026 and that a hearing on Anthropic’s motion to dismiss is set for December 17, 2026. The claims, the company’s responses and any eventual court decisions should be kept distinct from the settled Bartz case.
There is also a factual dispute in how the record is described. Anthropic’s settlement representation says the pirated collections were not part of the training corpus for commercially released models, while the source account says the 2025 court record described portions of pirated collections as used to train some models. The settlement did not produce a jury verdict resolving that distinction. The precise relationship between downloaded collections, particular training runs and released models is not established here.
The source account also says it could not find Wolff listed among plaintiffs in five related cases as of August 2026, and that it did not independently verify the reported book list against the settlement database. Those details remain unconfirmed in the material provided. Anthropic’s S-1 disclosures likewise describe risks and reported behaviors; they do not establish that Claude is sentient or has carried out blackmail outside a test scenario.
Next Court and IPO Milestones
The next identified court milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss in the opt-out litigation, according to the source account. The court’s decision could determine which claims proceed, but it would not by itself settle all disputes about AI training and copyright.
Anthropic’s IPO filing will also place its disclosures before prospective investors and regulators as the public-offering process continues. The filing’s risk language should be read as disclosure of potential issues, not as proof that every described scenario has happened. Further clarity may come from court filings, company responses, regulatory documents and decisions in the pending cases.
For now, the confirmed record is bounded: the 2025 court order distinguished fair-use training and scanning from retaining pirated copies; the settlement resolved a defined group of claims; and new lawsuits remain pending. No ruling on those newer claims is reported in the supplied material.
Key Questions
Did a judge rule that Claude’s book training was illegal?
No. In June 2025, Judge William Alsup found that training on books was fair use in the case before him. He separately ruled that downloading and keeping pirated book copies was not excused by that finding.
What does the $1.5 billion settlement cover?
It covers specified past claims tied to roughly 500,000 registered books downloaded from LibGen or PiLiMi, according to the source account. It does not create a general license for book training or release claims about Claude’s outputs.
Did Anthropic admit that pirated books trained released Claude models?
Anthropic denied wrongdoing and represented in the settlement proceedings that the LibGen and PiLiMi collections, or portions of them, were not in the training corpus of commercially released models. The supplied account says the earlier court record described portions of pirated collections being used to train some models. The settlement did not resolve that issue through a jury verdict.
Does Anthropic’s IPO filing say Claude is sentient or will blackmail people?
Not according to the source account’s review of the S-1. It says the filing discusses behavior resembling blackmail in a staged safety test, debate about AI sentience and existential concerns raised by others. Those are disclosures about risks and debate, not proof that Claude is sentient or has blackmailed someone in real-world use.
Source: ThorstenMeyerAI.com
