Anthropic’s IPO prospectus, made public on September 29, devotes roughly 80 of its 261 pages to risk, including what it calls “existential risks to humanity.”

The filing commits the company to a staggering $518 billion in future compute spending. It is one of the most detailed public disclosures yet of how a frontier AI lab views the dangers of its own technology.

The prospectus landed in a dramatic week for AI. Within 72 hours, the three frontier labs converged on an identical price: Claude Sonnet 5.5, GPT-6.1 Sol, and Gemini 4 Argon all list at $2 per million input tokens and $10 per million output. A commodity price floor has arrived at the top tier of AI.

At the same time, scrutiny is rising. The FTC opened a consumer-protection investigation into OpenAI, Anthropic, and METR over autonomous agents. The Bank of England warned that AI-linked debt issuance has doubled in a year.

Why it matters: the company warning about existential risk is also the one spending half a trillion dollars to build more capable systems. The tension between AI’s promise and its perils is now written into a legal filing.