Anthropic struggles to secure enough cloud capacity from the big providers

Anthropic, the artificial intelligence company that develops the Claude models, is struggling to obtain enough cloud computing capacity from Amazon Web Services (AWS), Google and Microsoft. That is the picture painted by an S-1 filing leaked to The Wall Street Journal. An S-1 is the registration document U.S. companies submit when preparing to sell shares on a stock exchange. According to the filing, the three tech giants increasingly regard Anthropic as a rival rather than a customer. The company is reportedly targeting an IPO in November.

Spending and capacity shortfall

The leaked filing adds detail to figures Reuters and the Financial Times reported in late September. Anthropic believes its spending on cloud contracts could exceed $750 billion, with some of those deals still under negotiation. Even so, that is not enough.

The company leases its own data centers and works with third-party data center operators. It worries, however, that those operators have a "limited operating history," which could make them less reliable than the large cloud providers known as hyperscalers.

Financial results

In the first half of the year, Anthropic recorded a net loss of $80.7 billion, against revenue of $16.3 billion over the same period. According to the WSJ, just two customers accounted for roughly a fifth of sales.

Last week, the company told prospective investors that third-quarter revenue alone had reached $16.8 billion, up 45% from the previous quarter.

Risks

Building on earlier disclosures about catastrophic risks, the filing also warns that Anthropic could be held liable for damages caused by harmful AI behavior, in an amount that could exceed its available insurance coverage.

Anthropic did not immediately respond to TechCrunch's request for comment.