Anthropic Locks in AI Chip Supply Through New Partnership with Google and Broadcom

Google and Broadcom have inked an agreement with Anthropic, the company said on April 7th. The purpose of the deal is to assure the availability of next-generation tensor processing units (TPUs) with capacities of multiple gigawatts, which are expected to be operational in 2027. The company's technological solutions are in high demand from customers all throughout the world, and the TPUs will help train and deploy frontier Claude models to keep up with that demand.

In addition to the sale news, the AI business that is set to go public also announced that its run-rate revenue has topped $30 billion. This is an increase from $9 billion at the end of 2025, brought about by the continued acceleration of demand for Claude on a global scale. Anthropic is constructing the infrastructure to support its exponential expansion, according to a blog post by CFO Krishna Rao. As a result of the support of its clientele, the business has been able to help Claude establish himself as an industry leader in artificial intelligence.

Why it is a Game-Changer Agreement for All Three?

Broadcom announced in a separate filing with the exchange that it has engaged in a long-term arrangement to provide Google with TPUs for generations to come. Artificial intelligence in data centres is meant to be powered by these TPUs. Anthropic now has access to about 3.5 gigawatts of TPU-based compute power, thanks to Broadcom and Google's increased collaboration, as stated in a filing with the Securities and Exchange Commission (SEC) and confirmed in Anthropic's blog post.

More than 500 businesses were spending over $1 million each year on Anthropic's services, according to the company's February Series G funding. The number has doubled in less than two months, surpassing 1,000, it noted. According to the firm, most of the new computing will be located in the US. In light of the company's pledge to invest $50 billion in enhancing the American computing infrastructure by November 2025, this cooperation represents a significant expansion of that commitment.

Tug of War Between US Govt and Anthropic

A disagreement between Anthropic and the US government has persisted for some time. The problem began when Anthropic said categorically that widespread surveillance and completely autonomous weapon systems were not appropriate uses of their technology. In March, Anthropic was classified as a "supply-chain risk" by the Trump administration; the artificial intelligence firm subsequently contested this classification in court.

Multiple news outlets have reported that US military aircraft used Anthropic's AI to launch airstrikes against Iran. Hence, resulting in the deaths of several people, including Supreme Leader Ayatollah Ali Khamenei. In the upcoming initial public offering (IPO) race between Anthropic PBC and competitor OpenAI Inc., the AI company might make its stock market debut as soon as October. In early discussions with Wall Street institutions, the creator of the famous Claude chatbot has discussed lead positions in a possible listing.