Google Deploys Gigawatt-Scale TPU Power for Anthropic in 'Over $10 Billion' AI Deal

TradingKey - The hardware race in artificial intelligence has reached a heavyweight transaction.

Alphabet Inc.'s Google Cloud has announced it will provide up to 1 million of its custom-designed Tensor Processing Unit (TPU) chips to rapidly rising AI startup Anthropic.

This multi-billion dollar deal ranks among the largest single transactions ever reported in the AI hardware sector and represents Anthropic's most significant expansion of TPU usage to date, substantially deepening the strategic partnership between Google and the company widely regarded as OpenAI's primary competitor.

According to Google's statement, these TPU chips — specifically designed to accelerate machine learning workloads — are expected to be deployed by 2026. Once operational, they will rapidly deliver more than 1 gigawatt of AI computing capacity.

“Anthropic and Google have a longstanding partnership and this latest expansion will help us continue to grow the compute we need to define the frontier of AI,” said Anthropic CFO Krishna Rao in a release.

“Anthropic’s choice to significantly expand its usage of TPUs reflects the strong price-performance and efficiency its teams have seen with TPUs for several years,” Google Cloud CEO Thomas Kurian added. “We are continuing to innovate and drive further efficiencies and increased capacity of our TPUs, building on our already mature AI accelerator portfolio, including our seventh generation TPU, Ironwood.”

Analysts note that Google's TPU employs a vertical power delivery design that is more efficient than NVIDIA's horizontal power delivery chips, making it a unique asset in the AI accelerator market that can significantly reduce energy costs.

Notably, Anthropic has adopted a distinctive multi-cloud architecture strategy. Its Claude model runs simultaneously on Google's TPUs, Amazon AWS's custom Trainium chips, and NVIDIA's GPUs, optimizing allocation across suppliers based on the characteristics and cost-effectiveness of different workloads (such as training, inference, and research).

Insiders report that this strategy enables the company to achieve higher computational efficiency per dollar compared to peers relying on a single cloud provider.

When AWS experienced a major outage on Monday, Anthropic's Claude service remained unaffected thanks to the distributed nature of its architecture, highlighting the advantages of this approach.

Anthropic is one of Silicon Valley's hottest AI startups. Founded in 2021 by former OpenAI employees, it has achieved 10x user growth since May, driven by Claude's exceptional performance in coding, generating approximately $500 million in annualized revenue.

Its valuation has skyrocketed as well, reaching $183 billion following a recent massive $13 billion funding round led by Iconiq Capital with participation from Fidelity and Lightspeed Venture Partners — a figure nearly triple its value at the beginning of the year.