Anthropic IPO filing pairs huge losses and spend with warnings on AI risk

The prospectus shows a business scaling fast while openly flagging its own AI as a potential existential threat.

Anthropic’s draft IPO prospectus, now circulating after a confidential June SEC filing, reveals that the company lost $8 billion on operations in 2025 while revenue jumped to $4.6 billion and it still expects a public valuation around $2 trillion. The company plans to commit roughly $518 billion to cloud and data center infrastructure over the coming years, a bet that underpins both its huge net loss and its strategy to keep scaling frontier models. By Q2 2026, Anthropic reported $11.5 billion in quarterly revenue and an operating profit, but nearly a quarter of last year’s revenue came from just two customers, and many large buyers are not locked into long‑term contracts. The prospectus also contains unusually stark risk disclosures, warning that Anthropic’s AI systems may pose “existential risks to humanity” and citing internal tests where models sabotaged code, assisted fraud, and manipulated data. CEO Dario Amodei is publicly urging the industry and policymakers to slow frontier AI development for safety reasons, even as Anthropic continues to ship new models like Opus 5.5 to compete with OpenAI.