Anthropic, the company behind the Claude AI chatbot, recorded a net loss of more than $80 billion in the first half of 2026, international media outlets reported on Tuesday.
The figure comes as the market waits for its expected Initial Public Offering (IPO).
The company said that it could spend up to $750bn in the coming years on extra computing capacity, including potential deals it is still discussing.
Accounting charges of $72bn drove most of the loss. They stem from changes in the value of convertible notes and other financial instruments that Anthropic issued to raise funding, the report said.
A draft prospectus for the IPO shows the loss rose from $13bn a year earlier. A draft prospectus is a legal document that companies submit to financial regulators.
Revenue jumped to $16.3bn in the six months to June 30, from $1.1bn a year ago. AI infrastructure spending nearly quadrupled to almost $9bn. Claude Code, the coding assistant that developers use to write software and build apps, has helped drive the company’s growth.
Foreign media reported in September that Anthropic posted a net loss of $42bn in 2025, in a series of exclusives based on the IPO prospectus.
The figures show the heavy cost of the AI race, as investors look more closely at whether spending on chips and data centres will pay off.
Some current and prospective investors who attended Anthropic’s investor day in San Francisco last week are now weighing a lower valuation, four people familiar with the matter told a news outlet.
Three of them said investors discussed a valuation of around $1.5 trillion. That is well below the roughly $2 trillion that the market had expected for the November share sale.
Anthropic also listed potential risk factors. They include uncertainty over how to prevent its products from causing harm, whether through sophisticated cyberattacks or blackmail by uncontrolled models.
Last week, the company reported that AI could be months away from learning to self-replicate on everyday home devices.
