Anthropic will expand its use of Google Cloud infrastructure, including up to one million Tensor Processing Units (TPUs), the two companies said Thursday. The deal is worth tens of billions of dollars and is expected to bring well over a gigawatt of compute capacity online in 2026. Google Cloud CEO Thomas Kurian said the move reflects the price-performance and efficiency Anthropic’s teams have seen with TPUs for several years.

The announcement comes as Anthropic reports more than 300,000 business customers and an annual revenue run rate approaching $7 billion. The number of large accounts—customers each representing more than $100,000 in run-rate revenue—has grown nearly sevenfold over the past year. Anthropic CFO Krishna Rao said the expanded capacity is meant to meet surging demand from Fortune 500 companies and AI-native startups while keeping Claude’s models at the frontier.

Anthropic continues to pursue a multi-chip strategy across Google TPUs, Amazon’s Trainium, and Nvidia GPUs. The company says it remains committed to its partnership with Amazon, its primary training partner and cloud provider, including the Project Rainier cluster with hundreds of thousands of AI chips across multiple U.S. data centers. Google, in January, agreed to a new $1 billion investment in Anthropic, adding to its previous $2 billion investment and 10% equity stake. Amazon has invested $8 billion to date.

The scale of the commitment is equivalent to a gigawatt-class buildout that industry estimates peg at around $50 billion, with roughly $35 billion for chips alone.