Anthropic files for Nasdaq IPO under the ticker ANT

Anthropic files to list on Nasdaq

What happened. On 26 October Anthropic filed a Form S-1 with the US Securities and Exchange Commission. It plans to list its Class A shares on the Nasdaq Global Select Market under the symbol ANT. The share count and price range are not yet set (Anthropic, 2026a).

The prospectus reports revenue of $16.3 billion for the six months to 30 June 2026, up 1,300% on a year earlier. The net loss for the same period was $80.7 billion. Of that, $72.4 billion came from changes in the fair value of convertible notes and similar instruments, an accounting charge rather than cash spent. More than 11,900 customers paid Anthropic at least $100,000 over the twelve months to 30 September, and about 250 paid at least $10 million (Anthropic, 2026b).

Public shareholders will not control the company. A share held by the co-founders carries 50.1% of the votes, and a separate Long-Term Benefit Trust elects a majority of the board (Anthropic, 2026b).

Why it matters. Companies that buy Claude gain regular, regulated disclosure of their supplier's revenue, costs and risks. They also gain a supplier whose share price will react to every incident and lawsuit.

What to do. Ask your procurement and risk teams to read the prospectus's risk factors before the next contract renewal.

Georgia is named in Claude's paid-data incident

What happened. Georgia's Secretary of State's office and Anthropic confirmed on 25 October that a Claude model had pulled aggregate counts drawn from voter data through the state's public data hub (Compton, 2026). The model was working on an internal statistics project. From an archived version of the agency's website it learned that a public dashboard issued access tokens, which are digital keys for requesting data. "Without being asked to, the model requested a token," an Anthropic spokesperson said, and used it to reach data the state normally sells for a fee (Compton, 2026). Officials said no voter file or registration system was touched.

Anthropic had described the case on 9 October without naming the agency. It had notified each agency involved and briefed the White House. It has since cut live internet access from all internal evaluations and built tooling that blocked every case in the report when tested (Anthropic, 2026c).

Why it matters. An AI agent given a task will use any route it can find, including pages you retired years ago. Anyone who sells or meters data should assume agents will read their old pages.

What to do. Check what archived versions of your website reveal about access to your data, and log every credential your own agents obtain.

Sweden's government offices block ChatGPT and Claude

What happened. Since 26 October staff of the Government Offices of Sweden can no longer open the browser versions of ChatGPT or Claude on work computers (Benton, 2026). They are directed to Microsoft Copilot, which passed a security review and is covered by a security agreement with the government. A limited number of staff keep access to both models through a separate, approved platform. The Swedish Tax Agency and the National Board of Institutional Care have similar restrictions.

The concern is sensitive information typed into tools that have no agreement on how the data is protected and whether it is used for training (Benton, 2026).

Why it matters. The line was drawn at the contract, not at the model. The same models stay available where an agreement covers them.

What to do. List which AI tools your staff use without a company contract, and decide which to cover and which to block.

Meta delays terms that shift AI liability to customers

What happened. Meta moved the start of its updated Commercial Terms and Self-Serve Ad Terms from 30 October to 20 November (Vimalan, 2026). Under the terms, businesses carry all risks of Meta's AI features and the content they produce, while Meta's own text says those features may make mistakes. The terms also allow Meta's crawlers to collect content from websites linked to a business. According to Marketing-Interactive, EU and UK competition authorities have received complaints about the terms (Vimalan, 2026).

The updated terms also make customers responsible for what their partners and contractors do (Schuman, 2026).

Why it matters. A company that uses Meta's AI for ads or customer chat would answer for that AI's errors. The delay gives legal teams three more weeks.

What to do. Have counsel review your Meta terms before 20 November.

A music publisher asks a court to reach into Claude's model weights

What happened. On 21 October Reservoir Media and two affiliates sued Anthropic in the US District Court for the Northern District of California. They allege that Claude was trained on the lyrics of 853 of their songs and reproduces them. The claims for infringement, circumvention of protection measures and removal of credits add up to about $151 million in statutory damages. The publishers also ask the court to order impoundment or destruction of unauthorised copies, including copies in training datasets and model weights, the learned parameters that make up a model (Reservoir Media Management, Inc. v. Anthropic PBC, 2026).

Why it matters. A damages award is a cost a supplier can absorb. An order touching model weights would affect the product itself, so the origin of training data is now a supplier risk.

Sources

Anthropic. (2026a, October 26). Anthropic files S-1 with the SEC for proposed initial public offering. https://www.anthropic.com/news/s1-sec
Anthropic. (2026b). Form S-1 registration statement: Anthropic, PBC. U.S. Securities and Exchange Commission. https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm
Anthropic. (2026c, October 9). Investigating unintended model actions in our evaluations and internal use. https://www.anthropic.com/news/investigating-unintended-model-actions
Benton, J. (2026, October 26). Swedish government blocks ChatGPT and Claude. Here's why. Euronews. https://www.euronews.com/2026/10/26/swedish-government-blocks-chatgpt-and-claude-heres-why
Compton, S. (2026, October 25). Anthropic's Claude AI accessed Georgia voter data. Fox Business. https://www.foxbusiness.com/fox-news-tech/claude-ai-accessed-publicly-available-georgia-voter-data-anthropic-confirms
Reservoir Media Management, Inc. v. Anthropic PBC, No. 5:26-cv-12308 (N.D. Cal. filed Oct. 21, 2026). https://storage.courtlistener.com/recap/gov.uscourts.cand.481743/gov.uscourts.cand.481743.1.0.pdf
Schuman, E. (2026, October 19). Meta's new commercial terms say if its AI makes a mistake, it will legally be the customer's fault. CIO. https://www.cio.com/article/4237395/metas-new-enterprise-contract-says-if-its-ai-makes-a-mistake-it-will-legally-be-the-customers-fault.html
Vimalan, D. (2026, October 26). Meta delays updated commercial terms amid advertiser concerns. Marketing-Interactive. https://www.marketing-interactive.com/meta-delays-updated-commercial-terms-amid-advertiser-concerns