Anthropic S-1 filing shows record revenue alongside record losses

Anthropic’s long-awaited S-1 filing lays out a company whose financial picture is defined by sharp contradictions. The go-public document, from a company valued at $965 billion, reports $16.29 billion in revenue for the first half of 2026, reflecting strong demand for its Claude products. In the same period, the company recorded a net loss of $80.7 billion for the six months leading up to June 2026, a figure that could make it one of the largest net losses ever reported by a public or soon-to-be-public company.

The company attributes the bulk of that loss to an accounting charge tied to convertible notes, but the filing makes clear that Anthropic is still losing money. The document leans heavily on the cost of computing power. The word “compute” appears more than 580 times, and the company’s costs and commitments run through nearly every section. Anthropic spent about $9.7 billion on compute in the first half of 2026 for model research and development alone. The company also says it plans to spend as much as $567 billion on compute in total.

Among the filing’s most striking commitments is an arrangement to buy compute from SpaceX, the rocket and satellite company led by Elon Musk, for up to $120.5 billion through 2029. The figure far exceeds what was disclosed in SpaceX’s own S-1, despite the public disputes between Anthropic and Musk. The compute dependence, in the reporting’s framing, keeps the company tied to ever-larger infrastructure costs at every stage of its growth.

The filing also points to spending beyond infrastructure. Anthropic spent more than $3 billion on marketing in the first half of 2026, and the sales and marketing function receives more prominent attention in the S-1 than might be expected. The company warns that it may not build its sales organization quickly enough, that it has limited experience running a direct sales force, and that weak marketing could slow revenue growth or cause it to decline. The risk language suggests that demand for AI, however large, still has to be actively won in a competitive market.

The filing also underscores how early many of the company’s capabilities remain. The phrase “limited experience” recurs in discussions of management, commercialization and infrastructure operations. Anthropic describes AI as one of the most consequential technologies in human history, with the potential to usher in an “age of miracles,” while also acknowledging “existential risks to society” and the possibility that the technology may pose catastrophic risks. Those statements sit alongside the company’s financial disclosures in the same document.

Anthropic, like OpenAI, has drawn wide investment from across Silicon Valley. At least 100 venture capital firms back the company, though none appear to own more than 5 percent, according to the filing. Public markets reacted positively to news of the planned listing, and the offering is expected to test investor appetite for AI companies at a very large scale. Whether the company can convert its revenue growth into sustainable profitability will be the central question once its shares begin trading.

2Views

Author

Derek Weston

Sports Writer

Derek Weston covers public affairs, politics, business, culture and daily news for Core Memo. The role focuses on verification, context, and clear explanations for readers.

All articles by this author