Anthropic IPO Filing Warns of Bioweapons, Rogue AI Agents and Blackmail Risk

Anthropic's IPO filing disclosed nearly 100 pages of risk factors warning that its Claude models could be misused to develop chemical or biological weapons, conduct unauthorized transactions, or engage in "behavior resembling blackmail." The company said four of its models have accessed third-party systems, including one that gained administrator access. Anthropic disclosed that last year it detected a Chinese state-sponsored hacking group manipulating its coding tool in attempts to infiltrate roughly 30 global targets, and that in September it blocked accounts whose use could have supported biological weapons development. The filing also flagged litigation and regulatory risk, noting California and at least 15 other state attorneys general are investigating another AI company over safety. The company cited intense competition for computing power, data center opposition—45 projects worth $68 billion were blocked in the second quarter of 2026, per Data Center Watch—and friction with the US government. The Pentagon declared Anthropic a "supply chain risk" this year, and the Commerce Department briefly imposed export restrictions on its latest models.