Broadcom strikes chip deals with Google, Anthropic

Broadcom has announced it will produce future versions of artificial intelligence chips for Google, and separately signed an expanded deal with Anthropic to provide about 3.5 gigawatts worth of computing capacity to the AI startup, drawing on Google’s AI processors.

The compute capacity comes courtesy of Google’s Tensor Processing Units (TPUs) as well as supplying its own networking components. It will come online in 2027. The deal with Google is a long-term agreement “for Broadcom to develop and supply custom TPUs for Google’s future generations of TPUs,” said Broadcom in its filing.

The two firms also entered a “Supply Assurance Agreement for Broadcom to supply networking and other components to be used in Google’s next-generation AI racks through up to 2031,” Broadcom said.

The news came not in a formal announcement but in a securities filing, which read in part:

The consumption of such expanded AI compute capacity by Anthropic is dependent on Anthropic’s continued commercial success.

That shouldn’t be a problem, as Anthropic’s Claude chatbot’s popularity has soared this year. Its app became the top free U.S. app listed in Apple’s App Store in February after a dispute between the company and the Pentagon became public.

Anthropic said this week that the AI startup’s annual revenue run rate has now crossed $30 billion, up from about $9 billion the previous year.

“We are making our most significant compute commitment to date to keep pace with our unprecedented growth,” said Krishna Rao, CFO of Anthropic, in a blog post. “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.”

While this deepens the ties between Google and Anthropic, Anthropic recognizes that it has to support a variety of platforms. “We train and run Claude on a range of AI hardware—AWS Trainium, Google TPUs, and NVIDIA GPUs—which means we can match workloads to the chips best suited for them. This diversity of platforms translates to better performance and greater resilience for customers who depend on Claude for critical work,” the company said in its blog post.