Leading AI developers, including OpenAI, Google, Meta, and Anthropic, currently lack insurance coverage for catastrophic risks associated with their technology. Traditional insurers are unwilling to underwrite these scenarios due to the potential for massive societal disruption and losses that exceed standard balance sheets.
Companies are currently managing these risks through self-insurance, captive insurance entities, or reliance on investor capital. OpenAI reportedly holds minimal coverage for emerging AI risks, with estimates around $300 million, though this figure remains subject to debate.
In its IPO prospectus released in late September 2026, Anthropic dedicated approximately 80 out of 261 pages to existential risks linked to advanced AI. The company is reportedly discussing a potential valuation of up to $2 trillion while disclosing that its technology could cause harm that remains uninsured.
Experts, including researcher Yoshua Bengio, suggest implementing mandatory liability insurance for AI firms, similar to the regulatory framework used in the nuclear power industry. Other proposed solutions include the creation of a mutual insurance pool for developers or the issuance of catastrophe bonds.
