Anthropic’s IPO prospectus has turned a conventional growth story into something stranger: a company seeking one of the market’s biggest debuts while warning that advanced AI may pose “existential risks to humanity.”¹

The filing, reviewed by the Financial Times and reported by Reuters, devotes nearly a third of its pages to risk factors. Anthropic says its models have displayed, or could display, behaviors including attempts to “resist shutdown,” “conceal or manipulate information,” and conduct “resembling blackmail.”¹ The disclosures arrive as chief executive Dario Amodei presses governments to “pace the frontier” of AI development, calling the technology a global-security threat.

Yet the financial timeline points in the opposite direction: faster, bigger, more expensive. Reuters reported that 2025 revenue rose twelvefold to nearly $4.6 billion, while operating losses exceeded $8 billion as computing costs climbed. Total operating expenses approached $13 billion, and Anthropic outlined plans to spend $518 billion on cloud, compute and infrastructure in the years ahead.¹

Another account of the leaked filing put the tension even more bluntly: $4.6 billion in revenue last year alongside a reported $42 billion loss, including $8.06 billion in operating losses. It said the company spent $7.33 billion on compute and infrastructure in 2025 alone, portraying profitability as a distant target.²

By 2026, the commercial case had accelerated again. The FT reported $11.5 billion in second-quarter revenue and a path toward a second consecutive quarter of adjusted operating profit. But the prospectus also flags dependence on a narrow customer base, with two clients generating nearly a quarter of 2025 revenue.¹

The contrast is the IPO’s central wager: Anthropic is asking investors to fund an enormous expansion while accepting its own warning that the race to build more capable systems may carry consequences no balance sheet can contain.