AI research lab Anthropic announced the signing on Monday New agreement with Google and Broadcom Claude for increased processing and computing power to power AI models. This reworking of its computing deals comes at a time when demand for its AI models continues to grow.
The deals will expand Anthropic’s use of Google Cloud’s tensor processing units, or TPUs, the company’s advanced AI chips, and are an extension of an agreement between the companies. October 2025 for more than a gigawatt of computing power.
This new computing capability will come online in 2027, Anthropic said in a blog post. The company did not provide specifics for the computing expansion, but recently The deal includes 3.5 gigawatts of computing, Broadcom SEC filings show.
Most of this calculation will be located in the US and will be a continuation of the company A $50 billion commitment to invest in US computing infrastructureAnthropic said in the article.
“This groundbreaking partnership with Google and Broadcom is a continuation of our disciplined approach to infrastructure scaling: we’re building the capacity needed to serve the exponential growth we’re seeing in our customer base, while enabling Cloud to redefine the frontiers of AI development,” said Krishna Rao, CFO of Anthropic, in a press release. “We’re making our most significant computing commitment to date to keep up with our unprecedented growth.”
Anthropic did not respond to TechCrunch’s request for comment.
The company has seen an increase in demand for its Claude models in recent months, despite the US Department of Defense’s Anthropic label. supply chain risk. Anthropic also recently closed $30 billion Series G funding round which valued the company at $380 billion.
Techcrunch event
San Francisco, CA
|
October 13-15, 2026
The company’s run-rate revenue now stands at $30 billion, a sharp jump from the $9 billion the company noted at the end of 2025. Anthropic also has more than 1,000 business customers who spend more than $1 million on an annual basis.
