Anthropic’s IPO filing devotes about 80 of its 261 pages to risk factors, warning that its AI could pose catastrophic or even existential risks. The scale of the warning has surprised many observers.

Partly, this is legal practice. Companies going public must disclose worst-case risks to investors. But the detail here goes beyond boilerplate, describing ways advanced AI could be misused or act unpredictably.

The warning matters because it comes from a builder of frontier AI, not a critic. It suggests the company’s own safety teams see real, if uncertain, dangers in the technology they are selling to businesses and consumers.

Investors must now price both Anthropic’s growth and its stated risks. The filing is one of the most watched technology listings of 2026, with the IPO expected to test how seriously markets take AI risk.

The disclosures will likely feed into regulation debates in the US and beyond. Governments are already weighing how to oversee AI development without stifling it.