Anthropic's IPO Filing: Insurers Use Claude to Underwrite Risk, but Who Will Insure Claude?

On 26 October 2026 Anthropic, the company behind Claude, publicly filed its registration statement (Form S-1) with the US Securities and Exchange Commission for a planned listing on Nasdaq. An S-1 is the document in which a company explains its business and, at great length, everything that could go wrong with it. In EU terms it is essentially a prospectus: the S-1 contains the prospectus that investors will read before buying shares. I read the filing for one thing: what it says about insurance.

Insurance appears in two very different roles. Insurers are customers who trust Claude with core work. And insurers are the risk carriers Anthropic is not sure can cover it.

Insurers as customers

In the business section, Anthropic lists insurance as one of the industries where Claude is deployed, separately from financial services. When it gives examples of mission-critical production workflows, the first one it names is underwriting insurance policies, ahead of clinical reports, cybersecurity and portfolio analysis for asset managers.

Anthropic also says adoption is particularly deep in highly regulated sectors, where the cost of unreliable AI is too high and trust cannot be added afterwards. In other words, it sells safety and reliability to firms like insurers as part of the product.

Insurers as risk carriers

The risk factors tell the other side. Anthropic writes that its insurance may not be enough to cover the liabilities it faces, especially when one incident or a series of incidents causes disproportionate harm. It then goes a step further:

• Capacity. Given the scale of harm a catastrophic AI incident could cause, no single insurer, and possibly not the insurance industry in aggregate, may have the capacity or expertise to underwrite matching cover.
• A young market. Anthropic calls the market for AI-specific risks "nascent and evolving" and says it may not be able to buy cover for some AI liabilities on reasonable terms, or at all.
• Cyber. It holds cyber insurance, but warns that some cyber and data liabilities may become uninsurable and that claims may not be paid in full or on time.
• Litigation. In September 2025 Anthropic agreed to pay USD 1.5 billion to settle Bartz v. Anthropic, a class action by book authors over the use of their books to train its models. The filing notes that insurance did not fully cover that settlement, and may not cover similar claims in future.

The liability question behind it

Why is the exposure so hard to insure? Because nobody yet knows who is liable for what an AI agent does. The filing is candid about this. Anthropic says existing rules on agency, product liability and professional negligence were built for human decisions and have not been settled for autonomous AI systems. It is unclear when an agent's action binds the user or the company that deployed it, and whether Anthropic's own disclaimers and liability caps would hold. It also expects claims against it even where the customer configured and supervised the agent.

Anthropic's position is that it provides a service, not a product, for product liability purposes, while accepting that courts may disagree.

The S-1 is a legal document written to protect the company, and risk factors are by nature cautious. Still, it is rare to see an AI developer say so plainly that the insurance industry may not be big enough for the risk it is creating, while counting that same industry among its most important customers.

Sources

• Anthropic, PBC, Form S-1 Registration Statement, filed with the SEC on 26 October 2026 (Risk Factors and Business sections)
• Anthropic, Anthropic files S-1 with the SEC for proposed initial public offering, 26 October 2026

Frequently asked questions

What does Anthropic's S-1 say about insurance?

It names underwriting insurance policies as an example of a mission-critical use of Claude, and lists insurance among the industries using it. In its risk factors it warns that its insurance may not cover its AI liabilities, that a catastrophic AI incident could exceed the capacity of the insurance industry, and that the market for AI-specific risk is still young.

When did Anthropic file its S-1?

Anthropic confidentially submitted a draft on 1 June 2026 and filed the public S-1 on 26 October 2026. It plans to list Class A shares on the Nasdaq Global Select Market under the symbol ANT. The price range and number of shares had not been set at filing.

For weekly analysis on financial innovation and regulation in Europe, subscribe to my newsletter.