Anthropic ups compute deal with Google and Broadcom amid skyrocketing demand

Anthropic, the AI research powerhouse, revealed on Monday that it has entered into a new agreement with Google and Broadcom to enhance its processing and computing capabilities for its Claude AI models. This strategic move comes as the demand for Anthropic's AI solutions continues to surge. Under the new arrangement, Anthropic will significantly expand its utilization of Google Cloud's advanced tensor processing units (TPUs). This follows an earlier agreement from October 2025, which covered over a gigawatt of compute capacity. The latest compute resources are set to become operational by 2027, according to a blog post from Anthropic. While specific details regarding the compute expansion were not disclosed, a recent filing by Broadcom with the SEC indicates that the deal encompasses 3.5 gigawatts of computing power. Most of this infrastructure will be established in the U.S. and aligns with Anthropic's $50 billion commitment to bolster U.S. computing infrastructure. Krishna Rao, CFO of Anthropic, emphasized the significance of this partnership, stating, "This groundbreaking collaboration with Google and Broadcom continues our disciplined approach to scaling infrastructure. We are building the necessary capacity to cater to the exponential growth in our customer base while enabling Claude to pioneer advancements in AI development." The company has experienced a remarkable increase in demand for its Claude models, driven by enterprise clients, despite concerns raised by the U.S. Department of Defense regarding supply chain risks associated with Anthropic. Recently, the company also finalized a $30 billion Series G funding round, which valued it at an impressive $380 billion. Anthropic's current run rate revenue stands at $30 billion, a significant leap from the $9 billion reported at the end of 2025. The company boasts over 1,000 business clients, each investing more than $1 million annually, further highlighting its rapid growth.