Anthropic, the AI research company behind Claude, has signed a major long-term deal with Google and Broadcom to secure AI chip capacity. The partnership aims to ensure sufficient TPU and other AI chip resources, expected to be online by early 2027.
Shares in Anthropic spiked 3% after the announcement when markets reopened.
CEO Statement
Anthropic CEO Krishna Rao said the deal “continues our disciplined approach to scaling infrastructure,” supporting the company’s growing customer base and enabling Claude to push the boundaries of AI development.
The company described this as its “most significant compute commitment to date,” reinforcing its rapid expansion.
We’ve signed an agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity, coming online starting in 2027, to train and serve frontier Claude models.
— Anthropic (@AnthropicAI) April 6, 2026
Infrastructure & Investment
- Most of the infrastructure will be built in the U.S., part of a $50 billion pledge to boost domestic computing capacity.
- Claude models run across AWS Trainium, Google TPUs, and Nvidia GPUs, with AWS as the primary cloud partner.
- Claude is accessible via AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry.
Broadcom confirmed the deal through a securities filing, highlighting a long-term supply assurance agreement extending through 2031.
This collaboration strengthens Anthropic’s AI compute capabilities amid surging demand for advanced AI models.
