Anthropic said in a document submitted on Tuesday that in the past month, the company reached new procurement commitments with Microsoft and Google, and cloud server rental expenses will increase by about $10 billion from now until 2033 compared to the previously disclosed value. Meanwhile, Anthropic said that the $34.9 billion server lease agreement signed with the smaller cloud service provider Lambda may be cancelled because Lambda failed to build the data center within the time agreed by the two parties. However, Anthropic said that although the contract gave it the right to cancel the agreement, the company “has no intention of doing so at this time.” According to the latest disclosure, Anthropic continues to seek computing power for model training and inference, and is also carefully evaluating the delivery progress of small and medium-sized cloud service providers such as Lambda. This also reflects the fact that Lambda, like many of its peers, is difficult to quickly launch sufficient computing power to meet customer needs.

Anthropic said in a document submitted on Tuesday that in the past month, the company reached new procurement commitments with Microsoft and Google, and cloud server rental expenses will increase by about $10 billion from now until 2033 compared to the previously disclosed value. Meanwhile, Anthropic said that the $34.9 billion server lease agreement signed with the smaller cloud service provider Lambda may be cancelled because Lambda failed to build the data center within the time agreed by the two parties. However, Anthropic said that although the contract gave it the right to cancel the agreement, the company “has no intention of doing so at this time.” According to the latest disclosure, Anthropic continues to seek computing power for model training and inference, while also carefully evaluating the delivery progress of small and medium-sized cloud service providers such as Lambda. This also reflects the fact that Lambda, like many of its peers, is difficult to quickly launch sufficient computing power to meet customer needs.