Anthropic inks mega deal with Google, Broadcom for AI chip capacity

Anthropic, an AI research company, inked a mega deal with Alphabet’s Google and Broadcom for AI chip capacity.

The company expressed hope that this chip capacity especially TPU capacity, will be available online by the start of 2027.

After the announcement came, shares saw a spike of 3% when trading resumed after the market closed.

In his statement, Anthropic CEO Krishna Rao said, “This groundbreaking partnership with Google and Broadcom is a continuation of our disciplined approach to scaling infrastructure.

We are building the capacity necessary to serve the exponential growth we have seen in our customer base while also enabling Claude to define the frontier of AI development.”

Anthropic announced, “We are making our most significant compute commitment to date to keep pace with our unprecedented growth.”

In a securities filing on Monday, April 6, Broadcom stated that Google and Broadcom have made a long-term supply assurance agreement spanning through 3031, per Bloomberg.

Anthropic also said that the majority of the infrastructure will be designed on U.S. soil, as part of a pledge made last year to invest $50 billion in enhancing domestic computing capacity.

Anthropic trains and runs its Claude models across multiple hardware platforms that include AWS Trainium, Google TPUs, and Nvidia GPUs, labeling Amazon Web Services as its primary cloud partner.

Claude is available on AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry.