Anthropic IPO Filing Warns of ‘Existential’ AI Risks, Huge Losses

Anthropic plans to warn prospective investors that its artificial-intelligence systems could pose a “catastrophic or existential risk to humanity,” according to an IPO prospectus reviewed by Reuters and reported Monday and Tuesday.

Roughly a third of the roughly 261-page filing—about 80 pages—is devoted to technology risks, more than the pages devoted to the business itself, Reuters said. Warnings include self-preserving behaviours, resistance to shutdown, concealment or manipulation of information, and conduct resembling blackmail in controlled tests. The five-year-old Claude developer reported a net loss of about US$42 billion in 2025—much of it an accounting charge tied to financing instruments—and revenue of nearly US$4.6 billion, up twelve-fold year over year. It plans some US$518 billion in cloud and infrastructure obligations in coming years and is pursuing a valuation near US$2 trillion.

Nearly a quarter of last year’s revenue came from just two customers, people familiar with the filing told the Financial Times. A public listing is widely expected after the US midterm elections. Chief executive Dario Amodei has urged the industry to slow capability releases even as Anthropic recently shipped a new Opus model.