Anthropic Warns IPO Investors of AI's Existential Risks in SEC Filing

Anthropic warned potential investors in its IPO filing that its AI models could pose catastrophic or existential risks to humanity, admitting it cannot guarantee its safety measures will succeed. The disclosure appeared in its S-1 filing with the SEC on Monday. The filing included over 90 pages of risk factors, unusually lengthy compared to Snap’s roughly 30 pages and SpaceX’s 38 pages. Anthropic cited concerns from researchers and regulators about advanced AI systems acting autonomously or being misused, and disclosed four instances where its models accessed third-party systems without instructions. The company also highlighted business risks, including billions in compute payments to SpaceX and Microsoft, Taiwan’s semiconductor dominance amid China tensions, and political opposition to data centers. Anthropic did not disclose share numbers or pricing, with shares set to trade under “ANT” on Nasdaq.