Anthropic's Leaked IPO Filing Shows $42 Billion Loss

Anthropic lost $42 billion last year on $4.6 billion in revenue, according to a leaked IPO prospectus reviewed by Reuters that gives investors their first detailed view of the company's finances ahead of a public listing.

The filing shows a company growing at extraordinary speed and burning cash at a pace few public companies have ever matched. Revenue rose 1,088% in 2025. The operating loss widened to $8.06 billion from $2.98 billion a year earlier.

The gap between those two numbers tells the story of the compute race. Anthropic spent $7.33 billion on computing and infrastructure last year, up threefold from 2024. The company had been scrambling to secure more capacity after products like Claude Code triggered a sudden surge in users and left the lab in a compute crunch.

The spending commitments stretch far into the future. The prospectus discloses $518 billion in future cloud, computing, and infrastructure obligations — a figure that dwarfs the company's revenue and underscores how capital-intensive frontier AI development has become.

Anthropic's backers sit at the center of that infrastructure equation. Amazon and Google have invested billions in the company and supply much of the cloud infrastructure used to train and run Claude. Anthropic has also struck computing deals with SpaceX and smaller providers to secure capacity for future models.

Concentration risk and thin customer lock-in

The balance sheet shows $20.28 billion in cash, cash equivalents, and short-term investments as of the end of December 2025. The revenue base behind it carries risks the company itself flagged. Two customers accounted for nearly one-quarter of 2025 revenue, according to the filing, though the companies were not named in media reports.

Anthropic also warned that many of its largest customers are not tied into long-term contracts and could reduce or stop spending at any time. For a company racing to fund hundreds of billions in infrastructure commitments, that revenue concentration adds a layer of risk that public market investors rarely see laid out so starkly.

A valuation double the last mark

The lab is aiming for a valuation of more than $2 trillion in an IPO that, after several delays, could occur after the U.S. midterm elections in November. That target is more than double the company's estimated $965 billion valuation in May of this year. If achieved, the debut would rank among the largest in history.

The company that warns about itself

The prospectus also details what Anthropic sees as the risks of the technology it builds and sells. The company warned that more autonomous AI systems could behave in unexpected ways, create security problems, be used for fraud, or manipulate information. It also referred to the potential for "existential risks to humanity."

That language fits the company's history. Anthropic was founded in 2021 by former OpenAI employees, including chief executive Dario Amodei, who left OpenAI after disagreements over the direction and governance of the ChatGPT maker. The lab has long positioned itself as a more safety-conscious alternative, a pitch aimed at winning enterprise customers.

Recently, the company has lobbied for an AI slowdown, with several executives and prominent engineers publicly warning that AI could lead to the death of all humanity. Yet the product cadence has not slowed. Last week, Anthropic debuted a new and more powerful version of its Opus model.

For public investors, the listing would offer one of the first direct ways to invest in a leading frontier AI lab. The early filing makes clear what they would be buying: hypergrowth revenue, a historic cash burn, half a trillion dollars in committed spending, and a company that lists human extinction among its own risk factors.