Anthropic's confidential filing for what could become the largest initial public offering in history warns investors that the company's own artificial intelligence models could pose a "catastrophic or existential risk to humanity," according to details of the prospectus that have begun circulating ahead of a planned stock market debut.
The document devotes roughly 80 of its 261 pages to risk factors — more space than it spends laying out the company's business plan — and includes warnings that advanced AI systems could eventually attempt to resist being shut down. The maker of the Claude chatbot confidentially submitted its draft registration statement on June 1, with a public filing expected within weeks as it works toward a listing that bankers are modeling at close to $2 trillion, which would rival the largest debuts on record.
Mounting losses, mounting scale
The filing also discloses a net loss of roughly $42 billion for 2025, about five times the $8.3 billion the company lost the year before, even as revenue has grown sharply alongside soaring spending on computing capacity. Anthropic was valued at $965 billion after a funding round in May that raised $65 billion, and it is now chasing a valuation roughly double that when shares eventually begin trading.
Anthropic's chief executive, Dario Amodei, has for years been one of the industry's most vocal proponents of the idea that advanced AI could pose existential dangers if developed carelessly, a position that has also functioned as a point of differentiation for a company founded by researchers who left OpenAI partly over safety disagreements. Securities lawyers note, however, that language in a prospectus is a different animal from a founder's public remarks: risk factors are formal representations to prospective shareholders made under legal liability, meaning regulators and plaintiffs' lawyers could later hold the company to the specific wording if something goes wrong.
The filing lands just as Anthropic faces fresh regulatory scrutiny on a separate front. The Federal Trade Commission has opened a broad investigation into both Anthropic and OpenAI over potential risks their chatbots pose to consumers, adding a second layer of pressure as the company tries to court public investors while simultaneously cataloguing, in unusually stark terms, the ways its core product could go wrong. How the two pressures intersect — a regulator probing consumer harm and a prospectus warning of graver, longer-term risks — is likely to shape how the IPO is received once the full document goes public in the coming weeks.
