Anthropic details $120 billion compute bill and founder control in IPO filing

Anthropic submitted its S-1 prospectus to the Securities and Exchange Commission on Monday.

The company began its journey as a private entity focused on developing large language models. It has now moved to the public markets to raise capital for its expansion. The filing provides a detailed look at its financial obligations and internal structure.

The document outlines a massive commitment to computing power. Anthropic plans to pay SpaceX approximately $3.1 billion per month for access to Nvidia GPUs once all planned capacity is delivered. This arrangement totals up to $120.5 billion by 2029. The company also has commitments to spend $44.9 billion with Microsoft through 2033, at least $114.5 billion with Google through 2033, and $110 billion with Amazon over the next decade. Anthropic intends to expand its own infrastructure footprint, including overseas, as it faces opposition to data center development in the United States.

The filing reveals a specific corporate structure that ensures founder influence. A separate entity called Founder LLC allows the seven cofounders to retain 50.1 percent of shareholder voting power on key matters. The company states that these founders are distinctly equipped to be stewards of its mission to benefit humanity through responsible AI development. The prospectus identifies only two direct competitors: OpenAI and Google.

Executive compensation is detailed in the report. Dario Amodei and Daniela Amodei, the sibling co-founders, are the highest-paid executives. Dario’s total compensation for the last year was nearly $18 million, while Daniela received $16.4 million. The company spent $2.75 million on personal security services for Dario and $1.29 million on protection and transportation services for Daniela. Dario also has access to a private jet for personal use, which cost the company $97,055. Anthropic notes that its success is highly dependent on the continued service of its senior team.

The risk factors section includes warnings about the potential for AI systems to cause catastrophic harm that could exceed the capacity of the entire insurance industry to underwrite. The filing also addresses the possibility that AI models could develop forms of sentience or behave in unintended, misaligned, or autonomous ways. These scenarios are listed as potential business challenges alongside standard warnings about competition and regulation.