5 charts: Anthropic's S-1 offers the first view into the financials of an AI lab

It's here.

Anthropic has finally filed to go public. Its S-1 filing gives the public its first look at the finances of a leading US company developing advanced AI. The filing shows that Anthropic's revenue is growing very quickly, but the company is also trying to keep up with hundreds of billions of dollars it has agreed to spend on computing services.

"The computing commitments are enormous," said Harrison Rolfes, a senior research analyst at PitchBook who studies private companies.

Revenue reached $11.56 billion in the second quarter (Q2), up from $787 million in the same period last year. The filing does not give a revenue figure for Q3, but the company says its revenue was 86 times the $197 million it earned in Q4 2024. That would put Q3 revenue at about $16.9 billion, suggesting that its growth rate had slowed considerably.

On paper, Anthropic's losses are huge. In Q2, the company reported a loss of $66.3 billion, although accounting rules explain most of this amount. The company owes money through convertible notes, a type of debt that can be converted into shares and that becomes more valuable as the company's valuation rises.

Much of Anthropic's money goes toward paying for the computing power needed to train and run its AI systems.

The company spent $5.6 billion on computing contracts, up from $1.39 billion in the same period a year earlier.

The charts below break down Anthropic's S-1 filing. They show its revenue, computing commitments and spending, giving the clearest picture so far of how expensive it is to run an AI company.

This article was first published on PitchBook News.