Anthropic's possible $2 trillion IPO prices in a future its $8 billion operating loss hasn't earned yet

Anthropic, the company behind the Claude models, has published its IPO prospectus. Markets could value it at about $2 trillion, according to an analysis by Stefon Walters of The Motley Fool. That would put it ahead of SpaceX as the largest IPO in history. The filing gives the first detailed look at the business. Revenue grew 1,050% to $4.6 billion in 2025. Net losses reached $42 billion in the same year.

The $42 billion figure overstates the problem. According to the same analysis, about $34 billion of it is an accounting charge. The operating loss, which counts only the costs of running the business, was around $8 billion. That is still a lot. Anthropic's total operating expenses were $12.65 billion, so it spent roughly $2.75 for every dollar it brought in. Compute, meaning the processing power used to train and run Claude, plus related infrastructure, accounts for 58% of that spending, or $7.3 billion. That bill was $400 million in 2023 and $2.5 billion in 2024, so it grew more than 18-fold in two years.

The numbers confirm something this industry often hides behind the hype: a frontier model is first a heavy-industry business. Each new generation needs a much bigger compute budget than the last. Revenue growing tenfold is impressive. It does not yet show that the cost curve bends. At $2 trillion, the market would be paying about 435 times 2025 revenue. A multiple like that is not based on current results. It assumes Anthropic becomes basic infrastructure for the economy. The source itself says that 'all burn, no profit' gets only so much patience when the price tag is this high.

The risk section is worth reading too. According to the analysis, Anthropic lists keeping autonomous AI models contained among its current challenges. It also mentions a wider possibility that advanced AI could have 'catastrophic' consequences. It is unusual for a company to tell future shareholders that its own product could be dangerous. It fits the safety-first identity Anthropic has built, and it will now face the discipline of securities law: what is written in a prospectus can be used against the company later.

Our reading: on the technology, Anthropic has earned its spot. Its Fable 5 leads our Zendoric Quality index with 90 points, ahead of OpenAI's GPT-5.6 at 79. Technical leadership and a durable business model are different things, though, and the market tends to merge them at the peak of a cycle. In the short term we expect volatility, and pressure to turn 'safety' into recurring revenue faster than the research timeline would advise. The more interesting effect is governance. A listed Anthropic will have to publish quarterly numbers, risks and compute spending. That transparency is more useful to the public than any valuation. In the long run, if the compute bill buys models that speed up medicine and science, $8 billion a year will look cheap. That outcome is an aspiration, not something the prospectus demonstrates, and investors should price that difference.