Anthropic, an AI-focused company, has just announced a major expansion of its partnership with Google Cloud. It plans to use up to one million specialized TPU chips (tensor processing units), representing an investment worth tens of billions of dollars. The agreement is expected to bring more than a gigawatt of computing capacity online as early as 2026. According to Anthropic's official statement, this will help push the boundaries of AI product research and development.
Thomas Kurian, CEO of Google Cloud, emphasized in the statement that Anthropic's decision to significantly expand its use of TPUs reflects its long-standing experience with their cost-effectiveness and efficiency. Google continues to innovate, including with its seventh-generation TPU called Ironwood, which is part of its mature portfolio of AI accelerators. Krishna Rao, Anthropic's CFO, added that this expansion will help meet growing customer demand, from large Fortune 500 companies to AI-focused startups. They all rely on the Claude model for their critical work.
Deal details
Anthropic now serves more than 300,000 business customers, and the number of large accounts—those generating more than $100,000 in annual revenue—has increased nearly sevenfold over the past year. This expansion with Google Cloud is intended to ensure sufficient computing power for rigorous testing, research, and responsible deployment at massive scale. Industry estimates put the cost of a gigawatt-scale data center at around $50 billion, with approximately $35 billion going toward chips.
Anthropic's computing strategy is unique in that it relies on a diversified approach involving three platforms: Google's TPUs, Amazon's Trainium, and Nvidia's GPUs. This enables efficient use of resources while continuing to advance the capabilities of the Claude model. The company remains committed to its primary partner, Amazon, with which it is collaborating on Project Rainier—a massive computing cluster with hundreds of thousands of AI chips distributed across several data centers in the US.
Anthropic's growth and rising market position
Anthropic's annual revenue is approaching $7 billion, reflecting rapid growth. The Claude model powers more than 300,000 businesses, a 300-fold increase over the past two years. The number of large customers generating more than $100,000 in annual revenue has grown nearly sevenfold. The Claude Code coding assistant reached $500 million in annualized revenue within just two months of launch, making it what Anthropic calls the fastest-growing product in history.
Amazon remains a key investor with an $8 billion investment, more than double Google's $3 billion investment, which gives it a 10% stake. Rothschild & Co Redburn analyst Alex Haissl estimates that Anthropic added 1 to 2 percentage points to AWS growth in the fourth quarter of last year and the first quarter of this year, with an expected contribution of more than 5 points in the second half of 2025. Scott Devitt of Wedbush noted that once Claude becomes a standard tool for enterprise developers, it will directly support AWS revenue for many years to come.
Thanks to its multi-cloud architecture, Claude was unaffected by the recent AWS outage because workloads are distributed among providers. Anthropic retains control over model weights, pricing, and customer data, without exclusivity to any provider. This neutral stance could prove decisive amid growing competition among major cloud providers.
Anthropic plans further investments in computing capacity to keep its models at the forefront. This agreement with Google underscores how the company is choosing a solid, efficient approach over grandiose plans such as OpenAI's 33-gigawatt Stargate project. Customers can therefore expect more reliable and powerful AI tools in the near future.
Sources: cnbc.com
