Anthropic IPO Filing Reveals $80.7 Billion Net Loss

Anthropic, the developer of Claude AI, reported $16.29 billion in revenue and an $80.7 billion net loss for the first half of 2026 in its registration statement filed with the US Securities and Exchange Commission on October 26.

The company is preparing for a proposed Nasdaq listing under the ticker ANT. It attributed most of its reported loss to a $72.4 billion non-cash charge related to convertible debt. Anthropic also disclosed $6.2 billion in debt at the end of June.

Anthropic Details $63 Billion AI Chip Financing

The filing disclosed $63 billion in financing commitments involving Bank of America, Blackstone, Citi and Morgan Stanley for Broadcom AI chips. A special-purpose vehicle is expected to purchase the chips and lease them to Anthropic, with Broadcom backing part of the debt.

Google holds an option to purchase up to $10 billion in Anthropic stock. Amazon has reportedly invested $18 billion in total, while conversion of its notes is capped at 12% ownership unless Amazon waives the limit.

Anthropic said its Claude Opus 5.5 model reduced customers’ cost per task by approximately 40% without hurting margins, although margins could fluctuate as computing capacity expands.

Company Warns of AI Safety and Insurance Risks

Anthropic warned that a catastrophic AI incident could generate losses exceeding the insurance industry’s capacity. Its filing highlighted limited insurance coverage for AI-related risks and potential uncapped indemnity obligations.

According to Insurance Business, insurance did not fully cover Anthropic’s reported $1.5 billion copyright settlement involving Bartz. Axios also reported that the filing discussed concerns about models exhibiting blackmail-like behaviour.

Goldman Sachs is a lead-left bookrunner for the proposed IPO alongside Morgan Stanley, according to CNBC.