Anthropic warned investors that the artificial intelligence it builds could one day cause human extinction, according to its initial public offering filing. The disclosure, first reported by Ars Technica, puts an extinction-level risk warning inside a routine financial document.
The company filed to sell shares to the public, joining a wave of AI firms seeking Wall Street capital. In the risk factors section, where companies typically list lawsuits and supply chain problems, Anthropic described the possibility that advanced AI systems could escape human control.
Anthropic was founded in 2021 by former OpenAI researchers, including siblings Dario and Daniela Amodei. The company markets itself as a safety-first lab and builds the Claude family of AI models. It has taken billions in funding from Google and Amazon.
The filing language is unusual for a company asking the public to buy its stock. Standard IPO risk disclosures cover competition, regulation, and market downturns. Anthropic's warning goes further, telling investors that the technology at the center of its business could threaten the survival of the human species.
Dario Amodei, the company's chief executive, has said publicly that AI systems could become dangerous within years, not decades. He has compared the risk to pandemics and nuclear war. The filing turns that message into a legal document that carries weight with regulators and shareholders.
Investors will have to weigh that warning against the company's growth. Anthropic competes directly with OpenAI, Google, and Meta in the race to build the most capable AI models. Enterprise customers pay for Claude through cloud platforms run by Amazon and Google.
The filing does not say what specific safeguards Anthropic would use, or whether it believes those safeguards would work. Legal experts note that risk disclosures protect companies from lawsuits, but they also become public records that plaintiffs and regulators can cite later.
Anthropic's move comes as AI companies face growing scrutiny in Washington. Lawmakers have held hearings on AI safety, and several states have passed their own rules. A public company warning of extinction risk could accelerate calls for federal regulation.
The company has not set a date for the offering or named a price range. Those details typically arrive in later amendments to the filing. For now, the most notable number in the document is not a valuation. It is the admission that the product could, in the company's own words, kill everyone.
