Anthropic Secures Multi-Gigawatt TPU Expansion with Google and Broadcom to Meet Surging AI Demand

Anthropic, a leading AI research lab, has announced a significant expansion of its compute infrastructure through a new partnership with Google and Broadcom. The agreement will bring multiple gigawatts of next-generation Tensor Processing Unit (TPU) capacity online starting in 2027, aimed at powering Anthropic’s advanced Claude models and addressing the exponential growth in customer demand.

Why it Matters

This strategic alliance underscores the rapid expansion of AI capabilities and the growing importance of robust compute infrastructure. According to Krishna Rao, CFO of Anthropic, this partnership is a continuation of the company's disciplined approach to scaling its infrastructure. The new compute capacity will enable Anthropic to serve an increasingly demanding customer base while pushing the boundaries of AI development.

Key Metrics

• Revenue Growth: Anthropic’s run-rate revenue has surged to over $30 billion in 2026, up from approximately $9 billion at the end of 2025.
• Customer Base: The number of business customers spending over $1 million annually on Claude services has doubled in less than two months, exceeding 1,000 clients.
• Investment Commitment: This partnership builds on Anthropic's November 2025 commitment to invest $50 billion in strengthening American computing infrastructure.

The Opportunity

The expansion of compute capacity is critical for Anthropic as it continues to develop and deploy Claude, one of the most advanced AI models available. By leveraging next-generation TPUs, Anthropic aims to enhance performance and resilience, ensuring that customers can rely on Claude for mission-critical tasks. This partnership also solidifies Anthropic's position as a key player in the global AI landscape, with Claude being the only frontier AI model available on Amazon Web Services (Bedrock), Google Cloud (Vertex AI), and Microsoft Azure (Foundry).

Diverse Compute Platforms

Anthropic trains and runs Claude on a variety of AI hardware, including AWS Trainium, Google TPUs, and NVIDIA GPUs. This diversity allows Anthropic to optimize workloads for the most suitable platforms, resulting in better performance and reliability. Despite this expansion with Google Cloud, Amazon remains Anthropic's primary cloud provider and training partner, continuing to collaborate on Project Rainier.

Geographic Focus

The vast majority of the new compute capacity will be located in the United States, aligning with Anthropic's commitment to strengthening American computing infrastructure. This move is expected to create significant economic opportunities and contribute to the nation's technological leadership in AI.

Key Risks

• Supply Chain Disruptions: The global supply chain for advanced hardware remains vulnerable to disruptions, which could impact the timely deployment of new compute resources.
• Competitive Landscape: As other tech giants invest heavily in AI infrastructure, Anthropic must maintain its innovation pace and customer satisfaction to stay ahead.
• Regulatory Scrutiny: Increased regulatory focus on AI development and data privacy could pose challenges for Anthropic's expansion plans.

Conclusion

Anthropic's strategic partnership with Google and Broadcom marks a significant milestone in the company's journey to meet the growing demand for advanced AI solutions. With this expanded compute capacity, Anthropic is well-positioned to continue driving innovation and delivering value to its expanding customer base.