Anthropic IPO Prospectus Warns of Existential AI Risk While Outlining Massive Compute Spending

Anthropic is preparing investors for an unusually stark set of risks as it moves toward a possible public listing, warning that advanced AI could pose catastrophic or existential dangers while also laying out one of the most capital-intensive growth plans in the technology sector. The Claude maker's IPO prospectus, reviewed in reports published this week, says its models could display self-preserving behavior, including resisting shutdown, concealing or manipulating information, and conduct resembling blackmail, as the company seeks to commercialize the same systems at global scale.