Under the agreement, Google Cloud will provide over a gigawatt of AI computing power using its in-house Tensor Processing Units (TPUs) starting next year - infrastructure Anthropic says will be critical for training and running its next generation of Claude AI models.

Krishna Rao, Anthropic’s Chief Financial Officer, said in a statement: “Anthropic and Google have a long-standing partnership, and this latest expansion will help us continue to grow the compute we need to define the frontier of AI.

“This expanded capacity ensures we can meet our exponentially growing demand while keeping our models at the cutting edge of the industry.”

While the companies declined to specify the total value, the deal is reportedly worth tens of billions of dollars, strengthening Google’s position as both an investor and infrastructure provider.

Google has already invested more than $3 billion in Anthropic, which is valued at $183 billion following a $13 billion funding round in September.

The partnership intensifies competition among tech giants Amazon, Nvidia, and Google, all vying to supply the computing power fuelling the AI boom.

Google Cloud CEO Thomas Kurian said Anthropic’s expanded use of TPUs reflects “the strong price-performance and efficiency” its teams have achieved.

The move highlights an escalating “AI arms race,” as developers scramble to secure the chips and energy resources needed to train ever-larger models - a contest now worth trillions in potential market value.