Anthropic is facing challenges in obtaining sufficient cloud resources from major providers like AWS, Google, and Microsoft, who increasingly view the company as a competitor rather than a client, as revealed in a recently leaked S-1 filing.
The leaked document sheds light on findings previously reported by other media outlets in late September, indicating that Anthropic anticipates spending upwards of $750 billion on cloud services, with some agreements still under negotiation.
Despite these substantial expenditures, it isn’t enough to meet their needs. Anthropic is now renting its own data centers and collaborating with external operators. However, the company expresses concern about these partners due to their “limited operating history,” potentially affecting their reliability compared to larger hyperscale providers.
This hefty financial commitment follows an alarming $80.7 billion net loss during the first half of the year. In that same period, the company reported revenues of $16.3 billion, with just two clients contributing about 20% of those sales, as noted in the report.
The filing also elaborates on previous disclosures about severe risks, cautioning that Anthropic could face liability for damages resulting from harmful AI actions that could far exceed their current insurance coverage.
Last week, the firm informed potential investors that its revenue for the third quarter has reached $16.8 billion, marking a 45% increase from the previous quarter. Anthropic is reportedly planning for an IPO in November.
Anthropic has yet to respond to inquiries for comments.
