Anthropic files IPO papers with $385 billion in compute commitments

Anthropic committed to spending roughly $385 billion on computing power through 2033, according to its S-1 prospectus filed with the SEC. The document details the massive scale of infrastructure deals required to power the company’s AI models. These costs far exceed the typical operational expenses of other technology firms entering the public market.

The largest single commitment is with SpaceX. Anthropic plans to pay approximately $3.1 billion per month for access to Nvidia GPUs. This arrangement totals $120.5 billion by 2029. The company also locked in long-term deals with other major tech providers. It agreed to spend $44.9 billion with Microsoft through 2033. A separate contract with Google is valued at at least $114.5 billion through the same year. Amazon will receive $110 billion over the next decade.

Corporate governance remains concentrated among the company’s founders. A separate entity named Founder LLC will give the seven co-founders 50.1 percent of voting power. This structure allows them to retain control over key decisions after the IPO. The filing states the founders are distinctly equipped to be stewards of the company’s mission to benefit humanity through responsible AI development.

Executive compensation reflects the company’s growth and risk profile. CEO Dario Amodei earned nearly $18 million last year, primarily through stock and option awards. President Daniela Amodei received $16.4 million. The company also spent $2.75 million on personal security for Dario and $1.29 million on protection and transportation services for Daniela. Dario Amodei additionally used a private jet for personal trips, costing the company $97,055.

The prospectus includes risk factors that go beyond standard market concerns. Anthropic warns that AI systems could cause catastrophic harm that exceeds the capacity of the entire insurance industry to underwrite. It also notes the possibility that models like Claude could develop forms of sentience or behave in unintended, misaligned, or autonomous ways. These scenarios would present unusual business challenges for the company.

In a competitive landscape filled with numerous AI developers, Anthropic names only two direct rivals. The filing identifies OpenAI and Google as its primary competitors. The company plans to expand its own infrastructure footprint, including overseas, as it faces increasing opposition to data center development in the United States.