Google and Anthropic Ink Multi-Billion Dollar Cloud Agreement

Anthropic and Google have formalized their cloud partnership, a multi-billion dollar deal granting the AI firm access to a substantial amount of Google’s Tensor Processing Units (TPUs). This strategic alliance underscores the escalating demand for AI compute power and the evolving dynamics of the cloud infrastructure landscape.

The agreement, valued in the tens of billions of dollars, marks Google’s most significant TPU commitment to date and is projected to deliver over a gigawatt of AI compute capacity by 2026. This positions Anthropic to further scale its AI model development and deployment.

While estimates suggest that constructing a 1-gigawatt data center could cost approximately $50 billion, with a significant portion allocated to specialized chips, Anthropic is taking a measured approach to its AI infrastructure build-out. This contrasts with some competitors’ more grandiose pronouncements, emphasizing pragmatic execution over sheer scale.

Founded by alumni of OpenAI, Anthropic has consciously adopted a more disciplined and diversified strategy, placing a strong emphasis on efficiency and targeting the enterprise AI market. This focused approach is proving to be a key differentiator in a crowded and rapidly evolving field.

A core tenet of Anthropic’s infrastructure strategy is its multi-cloud architecture. By distributing its Claude family of language models across Google’s TPUs, Amazon’s Trainium chips, and Nvidia’s GPUs, Anthropic can optimize for specific workloads such as training, inference, and research.

Google highlights the “strong price-performance and efficiency” offered by its TPUs for Anthropic’s needs. This collaboration allows Anthropic to harness the power of Google’s advanced hardware for computationally intensive AI tasks.

“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.

By diversifying its compute resources, Anthropic gains greater flexibility to optimize for cost, performance, and power efficiency. This strategic advantage allows the company to achieve better results for each dollar invested in compute infrastructure compared to a single-vendor approach.

Google views this partnership as a crucial endorsement of its cloud AI capabilities. “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,” stated Google Cloud CEO Thomas Kurian, highlighting the company’s seventh-generation “Ironwood” accelerator.

Anthropic’s increasing demand for compute resources is a direct result of its rapid business growth. The company’s annual revenue run rate is nearing $7 billion, and Claude is now being utilized by over 300,000 businesses— an impressive 300-fold increase in the last two years. Notably, the number of large customers contributing over $100,000 in run-rate revenue has grown nearly sevenfold within the past year.

The company’s agentic coding assistant, Claude Code, quickly became a significant revenue driver, generating $500 million in annualized revenue in just two months after its launch. This rapid growth demonstrates the strong market demand for enterprise-grade AI solutions.

While Google is powering a significant portion of Anthropic’s compute expansion, Amazon remains a crucial strategic partner. Amazon has invested $8 billion in Anthropic to date, more than double Google’s confirmed equity investment.

Crucially, AWS is considered Anthropic’s primary cloud provider, suggesting that its influence is both structural and financial. Anthropic’s supercomputer powered by Amazon’s Trainium 2 chips, known as Project Rainier, offers compelling performance and cost advantages over other chipsets. The decision to utilize Trainium reflects a strategy of optimizing for both performance and cost-effectiveness.

Wall Street analysts have already started to factor in the benefits of this partnership. Estimates suggest that Anthropic contributed significantly to AWS’s growth in the previous two quarters and is expected to continue to boost AWS revenue growth in the latter half of 2025.

Analysts predict that as Claude becomes a standard AI tool for enterprise developers, the corresponding usage will directly translate into revenue for AWS, fueling AWS’s expansion for years to come.

Google remains a key investor and strategic partner, having added to its previous investment with a new $1 billion commitment in January, bringing its total investment to $3 billion and securing a 10% equity stake.

Notably, Anthropic’s multi-cloud architecture provided resilience during a recent AWS outage. Due to its diversified infrastructure, the outage did not affect Claude’s availability and capabilities.

Despite its close ties to both Google and Amazon, Anthropic has carefully maintained a neutral position, retaining control over model weights, pricing, and customer data. It avoids exclusivity with any single cloud provider. This level of independence could prove to be a valuable asset as competition intensifies in the hyperscale cloud market, enabling Anthropic to optimize its resources and strategic direction based on the evolving landscape.