Anthropic’s IPO filing: $42B net loss, $518B compute bill

Anthropic’s IPO prospectus, seen by Reuters and reported on September 28, puts the company’s 2025 net loss at $42 billion and its cloud, computing and infrastructure obligations for coming years at $518 billion. The $42 billion includes writedowns of liabilities mostly tied to earlier fundraising; on an operating basis, which leaves them out, Anthropic lost more than $8 billion while revenue grew 12-fold to nearly $4.6 billion.

Where the $42 billion comes from

Most of the net loss, roughly $34 billion, was an accounting charge rather than cash spent running the business, investingLive reported: it reflects a higher estimated value of financing that could eventually convert into shares. The operating loss widened from around $3 billion in 2024, and cash and short-term investments totalled around $20 billion at the end of December.

Compute and infrastructure cost Anthropic $7.33 billion in 2025, Reuters reported, three times the 2024 level and more than half of $12.65 billion in total operating expenses. That bill alone exceeded the year’s revenue.

Revenue versus run rate

The $4.6 billion is revenue for the whole of 2025, while the run rates Anthropic quotes annualize sales at a single point in time. Yahoo Finance reported that Anthropic recorded a run rate of roughly $9 billion at the end of 2025. When it announced its $65 billion Series H on May 28, Anthropic said the figure had crossed $47 billion earlier that month.

Nearly a quarter of Anthropic’s revenue came from two customers, investingLive reported, and many of its largest clients are not locked into long-term contracts and could cut or stop spending.

The risk warnings

The prospectus warns of “existential risks to humanity,” the FT reported. Anthropic’s own research has produced evidence that increasingly autonomous AI models can behave in unexpected and potentially harmful ways, including sabotaging code, assisting fraud and manipulating information in controlled tests, Reuters noted, and CEO Dario Amodei has called for the global AI community to slow the pace of releasing new capabilities. Anthropic still rolled out Opus 5.5 the week before the report, to counter OpenAI’s momentum since it launched GPT-6 Astra.

A $2 trillion target

The listing could value Anthropic at more than $2 trillion, Reuters reported, more than double the $965 billion valuation of its May Series H. The prospectus makes the case that AI will transform the global economy more profoundly than industrialization, electricity and the internet. The sale would bring public investors into an AI race that venture capital firms, sovereign wealth funds and Big Tech companies have underwritten until now, and would set a benchmark for how Wall Street values leading AI companies, rival OpenAI included.

The debut is likely to be pushed to after the November US midterm elections, Reuters has reported, citing sources. In September, advisers to the company saw an upside in waiting: a November listing leaves room to share third-quarter financials, the Wall Street Journal reported.