Broadcom has agreed to develop and supply future generations of Google’s tensor processing units (TPUs) and expanded its deal with Anthropic, which will give the AI company access from 2027 to around 3.5GW of computing capacity based on Google’s AI processors, according to a securities filing on Monday (6 April).
The filing also mentioned that Broadcom and Google signed a supply assurance agreement under which Broadcom will provide networking and other components for Google’s next-generation AI racks through to 2031.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
Broadcom shares rose 3% in extended trading after the disclosure.
Under the expanded collaboration, Anthropic will access the capacity through Broadcom as part of the multiple gigawatts of next-generation TPU-based computing power it has committed to use.
The filing said Anthropic’s use of the expanded capacity depends on its continued commercial performance.
The companies are also discussing the deployment with some operational and financial partners.
Anthropic said the additional TPU capacity is expected to come online from 2027 through Google Cloud services and access to Google-built TPUs supplied through Broadcom. Most of the new infrastructure will be in the US.
On a March earnings call, Broadcom CEO Hock Tan said the company was “off to a very good start in 2026” supplying Anthropic with 1GW of computing capacity using Google’s in-house TPUs. Broadcom supports Google in producing its TPUs.
The agreements reflect demand for infrastructure used to run generative AI models.
Anthropic said in a blog post that its annualised revenue had passed $30bn, up from about $9bn at the end of last year. The company has more than 1,000 business customers spending more than $1m a year, double the number from two months earlier.
Anthropic finance chief Krishna Rao, in the blog post, said: “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.”
Anthropic is also expanding its use of Google Cloud services, including BigQuery, Cloud Run, and AlloyDB, to support model development, data, and applications.
The company said thousands of customers access Claude models through Google Cloud, including Coinbase, Palo Alto Networks, Cursor, Shopify, and Replit.
Broadcom is also working with Anthropic rival OpenAI on custom AI silicon.
OpenAI and Anthropic both rely heavily on Nvidia graphics processing units through cloud providers including Amazon, Google and Microsoft.
OpenAI has also committed to using 6GW of AMD GPU capacity, with the first gigawatt expected to come online in the second half of this year.
