Anthropic has signed a deal with Google and Broadcom for about 3.5 gigawatts of computing capacity built on Google’s tensor processing units, with the capacity expected to come online starting in 2027, the company said.
The agreement expands an existing relationship. Broadcom CEO Hock Tan said during an earnings call last month that Broadcom would provide one gigawatt of Google TPU computing to Anthropic in 2026, according to CNBC. Broadcom helps Google design its TPUs. In a securities regulator filing on Monday, Broadcom said Google and Broadcom also signed a long-term supply guarantee agreement through 2031, according to Bloomberg.
Along with the IT news, Anthropic said its revenue rate has surpassed $30 billion on an annualized basis, more than three times the roughly $9 billion figure it posted at the close of 2025. Business traction has also accelerated: The number of customers committing at least $1 million a year has surpassed 1,000, a threshold Anthropic said is double what it was reporting at the time of its G Series announcement in February.
“We are making our largest IT commitment to date to keep pace with our unprecedented growth,” Anthropic Chief Financial Officer Krishna Rao said in a statement.
Anthropic said most of the new infrastructure will be built on U.S. soil, framing the commitment as a continuation of a promise made in November 2025 to commit $50 billion to national computing capacity. According to Monday’s securities filing, Broadcom noted that Anthropic’s ability to take advantage of the additional compute depends on its continued business performance and noted that discussions are underway with third-party operating and financial partners to support the launch.
Anthropic trains and runs its Claude models on multiple hardware platforms, including AWS Trainium, Google TPU, and Nvidia GPU, and describes Amazon Web Services as its primary cloud and training partner. Claude is available on AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry.
The revenue growth comes as Anthropic navigates a legal dispute with the Pentagon, which designated the company as a supply chain risk after a standoff over AI safety barriers. Anthropic has warned that the label could cost it billions in lost revenue. Still, the company’s annualized revenue has more than tripled in the months since the dispute became public, driven in part by demand for its Claude Code development tools and broader enterprise adoption.
Broadcom shares gained as much as 3.6% in after-hours trading after the presentation was released, according to Bloomberg. No financial conditions were attached to the agreement. A post-earnings research note from analysts at Mizuho puts Anthropic’s Broadcom AI revenue potential at $21 billion by 2026 and $42 billion by 2027, according to CNBC.