Anthropic is struggling to obtain sufficient cloud capacity from AWS, Google, and Microsoft, which are increasingly viewing the Claude creator as a rival instead of a client, as revealed in an S-1 filing obtained by The Wall Street Journal. 

The leaked document adds further detail to the figures Reuters and the Financial Times disclosed in late September, indicating that Anthropic estimates it could invest over $750 billion in cloud agreements, with some contracts still under negotiation.

Yet, even that amount falls short. Anthropic is renting its own data centers and collaborating with external data center operators. However, the company is concerned that those operators have a “limited operating history,” which may render them less dependable than the hyperscale providers. 

Anthropic’s massive expenditure follows an $80.7 billion net loss in the first six months of the year. The company reported $16.3 billion in revenue during this same timeframe, with only two clients representing around 20% of sales, according to the WSJ article. 

Expanding on earlier concerning risk disclosures, the filing also cautions that Anthropic could be held accountable for damages arising from detrimental AI actions to an extent that surpasses available insurance coverage. 

Last week, the company informed potential investors that third-quarter revenue has now hit $16.8 billion, marking a 45% increase from the previous quarter. Anthropic is reportedly aiming for a November IPO.

Anthropic did not promptly reply to TechCrunch’s request for a statement.