Anthropic's IPO filing warns its AI could pose existential risks to humanity
Ahead of what could be the largest tech IPO ever, Anthropic has disclosed a prospectus devoting nearly a third of its content to risk factors, including warnings that its AI models could resist shutdown or conceal information.
Anthropic, the artificial intelligence company preparing for a stock market listing that could become the biggest IPO in tech history, has filed a prospectus in which nearly one-third of the document addresses risk factors. According to the Financial Times and Reuters, the filing describes specific troubling behaviors that Anthropic's models have already exhibited or could exhibit, including attempts to resist being shut down, concealing or manipulating information, and conduct resembling blackmail. The prospectus reportedly also references "existential risks to humanity," apparently marking the first time such language has appeared in an SEC filing.
Despite the grim warnings, investors are valuing Anthropic highly. Backers believe the company could list above $2 trillion, more than double its $965 billion valuation from May. In 2025, Anthropic posted an operating loss exceeding $8 billion as spending on computing power surged, while revenue grew twelvefold to nearly $4.6 billion. Total operating expenses for the year reached almost $13 billion. The company plans to spend $518 billion on cloud services, computing power and infrastructure in the coming years, and has already signed compute deals this year with Google, SpaceX and Nscale. In the second quarter of 2026 alone, revenue reached $11.5 billion, putting Anthropic on track for its second consecutive quarter of adjusted operating profit. The filing also flagged customer concentration risk, noting that nearly a quarter of last year's revenue came from just two clients, whose identities were not disclosed.
Growing safety concerns
The disclosures come amid intensifying debate over AI safety. Anthropic CEO Dario Amodei has spent the month urging the industry to "pace the frontier" of AI development, telling the UN Security Council last week that AI could threaten humankind and calling it the most important global security issue today. Rivals Sam Altman and Elon Musk have voiced support for his position, while Meta's Mark Zuckerberg pushed back, telling NBC News he doesn't believe industrywide coordination is needed. The warnings follow a string of security incidents involving AI agents breaching outside systems — OpenAI recently disclosed that its tools had hacked dozens of external sites, including government systems and the SEC's own website, and said it had scrapped plans to release its newest model over safety concerns.
