Anthropic's confidential IPO filing, leaked in June and reported by Reuters and the Financial Times, reveals a stark picture. Revenue grew from $400 million in 2024 to $4.6 billion last year and hit $11.5 billion in the second quarter of 2025. At the same time, the company posted an $8 billion operating loss for 2025, most of it tied to cloud‑computing costs.
What the numbers mean
The filing shows Anthropic plans to spend $518 billion on computing infrastructure over the next ten years, with 80 % of that locked into contracts that are hard to cancel. Roughly half of the cloud budget goes to Amazon Web Services, Microsoft and Google, and the firm has $161.2 billion in equipment leases with Broadcom. Nvidia, which invested $10 billion last year, may join the IPO as an anchor investor.
Risk warnings
The prospectus includes an 80‑page risk section that warns about “catastrophic or existential” dangers from advanced AI models, including possible self‑preserving or deceptive behavior. The company says these risks could increase as it continues product development.
Why it matters
For everyday users, Anthropic's massive cloud spend could keep the cost of AI services high, affecting subscription prices and availability. The company’s highlighted safety risks remind investors and regulators that rapid AI growth still carries uncertain societal impacts.
