Anthropic Acknowledges AI "Existential" Risks in IPO Prospectus
Anthropic, the developer of the Claude artificial intelligence model, has warned that government stances towards the company and its technology could have widespread impacts on its business, including relationships with commercial customers and partners. This warning is contained within the initial public offering (IPO) prospectus reviewed by Reuters on Friday (2/10/2026).
Citing Reuters reports, companies commonly list risks related to policy changes, particularly those acting as government contractors. SpaceX, for instance, noted in its IPO documents that maintaining good relations with US government agencies is crucial to its business.
However, Anthropic’s disclosure is considered broader. The company suggests that government perception of the firm and its behaviour could spill over to customers, partners, and other commercial relationships, and even civilisation itself. The prospectus also mentions that advanced AI could pose catastrophic or existential risks to humanity. This is an extraordinary statement for a company preparing for an IPO with a valuation that could reach US$2 trillion.
Nevertheless, Anthropic noted that revenue from government agency contracts accounts for less than 1% of its total annual revenue.
In the prospectus, Anthropic listed several interactions with the US government over the past year as examples of actions that could harm its business:
February: The US President ordered federal agencies to stop using Anthropic models. The US Department of Defense also designated Anthropic as a supply chain risk to national security.
June: The US Department of Commerce imposed global export restrictions on the Fable 5 and Mythos 5 models. Anthropic deactivated both models for all customers to comply with the regulations.
These restrictions were subsequently lifted by the Department of Commerce, and Anthropic resumed operating the models. However, the company warned that similar actions could recur. Anthropic also noted that such measures could damage its reputation through negative press and public scrutiny, regardless of the final outcome.
The company added that sales to government agencies carry additional risks, including shifts in government views regarding Anthropic or its technology.
AI Security Spotlight
Anthropic is currently under scrutiny amidst concerns over the rapid development of AI, intense global competition, and a lack of oversight. AI safety issues have come to the fore after CEO Dario Amodei called for the industry to slow development, following repeated reports of hacking activities by “rogue” autonomous AI agents.
Amodei met with President Donald Trump at a dinner last week, as calls for AI regulation intensified. Trump largely rejected these calls. Meanwhile, the US Federal Trade Commission (FTC) is reportedly conducting cross-industry investigations into AI companies, including Anthropic.