Anthropic, the company that builds Claude, has reported a massive loss of more than $80 billion in the first half of 2026. This follows as it prepares for a possible stock market debut.

The figure may sound alarming, but there is a catch: most of the loss comes from accounting changes rather than actual cash leaving the company. At the same time, Anthropic's revenue is growing rapidly as businesses and developers turn to its AI tools.

The numbers offer a glimpse into the expensive race to build better AI models and the risks that come with it.According to a report by The Wall Street Journal, Anthropic's draft IPO prospectus shows a net loss of $80.7 billion between January and June 2026. The document, which companies file before selling shares to the public, also reveals plans for massive spending on computing power.

Anthropic could spend up to $750 billion in the coming years, including deals that are still being discussed.Before assuming that Anthropic burned through $80 billion in cash, it is worth looking at the fine print. Around $72 billion of the reported loss came from accounting charges linked to convertible notes and other financial instruments the company used to raise money.Convertible notes are loans that can later turn into company shares.

When a company's valuation rises, the value of these financial instruments can change, creating an accounting loss even when the company does not spend that amount in cash. That does not mean Anthropic's AI ambitions come cheap, though.

The company spent nearly $9 billion on computing infrastructure in the first half of 2026 alone, almost four times the amount spent during the same period last year. Its revenue, meanwhile, jumped to $16.3 billion from $1.1 billion a year earlier.Anthropic has benefited from the popularity of Claude Code, its AI coding assistant that helps developers write software and build applications.

However, the company needs enormous computing capacity to train and run its AI models, making access to chips and data centres a major expense.Anthropic's IPO faces questions over valuation and AI risks The financial disclosures could complicate Anthropic's planned IPO, which is reportedly targeted for November. According to Reuters, some investors who attended the company's investor day recently discussed a valuation of around $1.5 trillion, below the roughly $2 trillion valuation previously expected.Anthropic is also looking at long-term computing commitments worth hundreds of billions of dollars.

These investments could help the company keep up with demand, but they also raise questions about how quickly revenue can grow enough to justify the spending.There are other risks, too. The draft prospectus reportedly warns that the company's AI systems could cause harm, including through cyberattacks or blackmail by uncontrolled models.

It also highlights the possibility of severe security incidents and the company's dependence on a small number of major customers.- Ends