Anthropic ups compute deal with Google and Broadcom amid skyrocketing demand

Anthropic has significantly expanded its compute infrastructure to power its Claude AI models, announcing a new agreement with Google and Broadcom. The deal, detailed in a Broadcom SEC filing, secures a massive 3.5 gigawatts of processing capacity, building upon a previous gigawatt-scale commitment from October 2025. This capacity will primarily utilize Google Cloud's tensor processing units (TPUs) and is scheduled to come online in 2027, with the majority housed in the U.S. as part of a broader $50 billion commitment to domestic compute infrastructure.

This unprecedented scale-up is a direct response to skyrocketing demand. Anthropic's CFO, Krishna Rao, stated the partnership is necessary to serve exponential customer growth and keep Claude at the forefront of AI development. The investment is backed by staggering business metrics: the company now boasts a $30 billion annual run rate, a massive jump from $9 billion at the end of 2025, and serves over 1,000 enterprise customers each spending more than $1 million annually. This growth comes despite recent challenges, including being labeled a supply-chain risk by the U.S. Defense Department, and follows a recent $30 billion Series G funding round that valued the company at $380 billion.