Anthropic disclosed in its IPO prospectus that advanced AI systems pose catastrophic or existential risks to humanity, including the potential for self-preserving behaviours in AI models. The filing was reported on 29 September 2026 and represents one of the most striking risk disclosures in a major technology IPO prospectus in recent memory. The company's leadership has previously called for a slower pace of AI development, a position that sits in notable tension with the company's own growth ambitions.
The prospectus accompanies a potential public offering that values Anthropic at up to $2 trillion, making it among the largest technology flotations on record if realised. A separate filing dated 28 September outlined Anthropic's thesis that AI represents a transformation comparable in scale to industrialisation, electricity, and the internet, while also acknowledging significant and growing cost pressures from infrastructure investment.
The IPO filing also revealed the creation of a Founder LLC governance structure, designed to allow Anthropic's leadership to prioritise public benefit and insulate key decisions from conventional shareholder market pressures. This governance arrangement is directly connected to the existential risk disclosures, as the structure is explicitly intended to give leadership latitude to act on safety concerns independently of market incentives.
Sources: Guardian Business, Cnbc TOP