Five weird details from Anthropic’s IPO filing

Anthropic has filed its S-1 prospectus with the SEC, a document of more than 300 pages that offers an unusually detailed look at its business. Among the most notable disclosures: sibling co-founders Dario and Daniela Amodei are the company's two highest-paid executives, with CEO Dario earning nearly $18 million in total compensation and President Daniela $16.4 million last year, mostly from stock and option awards. Anthropic also spent $2.75 million on personal security for Dario and $1.29 million on personal protection, home security, and transportation for Daniela, plus $97,055 for Dario's personal use of a private jet. The company says its success is highly dependent on the continued service of its senior team.

The filing also lists unconventional risk factors alongside standard warnings about competition and regulation. Anthropic cautions that AI systems could cause potentially "catastrophic" harm of a kind that the entire insurance industry could not underwrite, and raises the question of whether Claude or another AI system could possess or develop "forms of sentience" or behave in unintended, misaligned, or autonomous ways that could damage the company.

On compute, the commitments are staggering: roughly $3.1 billion per month to SpaceX for Nvidia GPU access once full capacity is delivered, totaling as much as $120.5 billion by 2029, plus $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 also plans to expand its infrastructure overseas amid US opposition to data center development. Its corporate structure is likewise unusual: as a public benefit corporation with a twist, a separate entity called "Founder LLC" will let the seven cofounders collectively control 50.1 percent of shareholder voting power on key matters after the IPO. Finally, despite a crowded AI market, Anthropic names only two direct competitors: OpenAI and Google.

(Source：The Verge)