Anthropic secures 3.5 gigawatts of new compute capacity

On Monday, AI research company Anthropic announced a new agreement with Google and Broadcom to secure increased processing and compute capacity for its Claude AI models. The deal expands Anthropic’s use of Google Cloud’s tensor processing units, which are Google’s custom AI chips. This agreement represents an expansion of a previous deal the companies struck in October 2025, which provided more than a gigawatt of compute capacity. According to a Broadcom SEC filing, the newly signed agreement includes 3.5 gigawatts of compute capacity. Anthropic stated that this new capacity is scheduled to come online in 2027, with the majority of the compute infrastructure housed in the U.S.

The infrastructure expansion follows a period of rapid scaling for Anthropic, supported by a $30 billion Series G funding round—a late-stage round of venture capital financing—that valued the company at $380 billion. According to CFO Krishna Rao, the partnership with Google and Broadcom is a continuation of the company’s disciplined approach to scaling infrastructure. Rao noted that the agreement builds the capacity necessary to serve the growth in its customer base while enabling its Claude models to define the frontier of AI development.

To support this expansion, Anthropic disclosed several key financial and operational metrics:

• Valuation: The company is valued at $380 billion following its $30 billion Series G funding round.
• Run rate revenue: Anthropic’s projected annual revenue has reached $30 billion, representing a significant increase from the $9 billion recorded at the end of 2025.
• Customer base: The company now has more than 1,000 business customers spending more than $1 million on an annualized basis.

Rao emphasized the scale of the expansion, stating: “We are making our most significant compute commitment to date to keep pace with our unprecedented growth.”

This massive infrastructure expansion is part of Anthropic’s broader $50 billion commitment to invest in U.S. compute infrastructure. However, the company’s rapid scaling occurs alongside ongoing regulatory scrutiny. Specifically, the U.S. Defense Department has labeled Anthropic a supply-chain risk.

Why it matters

Anthropic’s massive infrastructure expansion highlights the escalating compute arms race in AI, even as the company navigates regulatory scrutiny regarding its supply chain.