Anthropic expands Google and Broadcom partnership to secure more AI computing power

Anthropic said on April 6, 2026, that it is expanding its partnerships with Google and Broadcom to add multiple gigawatts of next-generation TPU capacity for future Claude models, a move that underscores how central compute access has become in the artificial intelligence race. The company said the arrangement is intended to support its next wave of model development as demand for frontier AI systems continues to rise.

The announcement comes as major AI developers compete not only on model quality and product features, but also on the ability to secure enough chips, cloud capacity and power to train and run increasingly large systems. Anthropic has positioned Claude as a direct rival to other leading AI assistants, and the latest infrastructure deal suggests the company is planning for a much larger operating footprint.

More capacity for future Claude models

In its statement, Anthropic said the expanded partnership will provide access to multiple gigawatts of next-generation TPU capacity. The company did not disclose financial terms in the announcement, and it did not say exactly how the capacity would be allocated across training, inference or other uses.

Google has been one of Anthropic’s key infrastructure partners, and Broadcom has been involved in the hardware side of the arrangement. The companies said the deal is designed to support future Claude models, though they did not provide a launch date for any specific product release.

Compute remains a strategic bottleneck

The agreement reflects a broader reality in artificial intelligence: the companies building the most advanced models need long-term access to specialized computing resources well before products reach the market. That demand has pushed AI developers into larger and more complex partnerships with cloud providers and chipmakers.

Anthropic’s latest move also comes amid intense competition for high-end accelerators and data center capacity across the industry. For model developers, securing compute early can shape how quickly they can train new systems, expand usage and respond to rivals.

What the deal means for Anthropic

Anthropic has been working to scale Claude while maintaining a strong position in enterprise AI. The company has not said how the new capacity will affect pricing, availability or the pace of future launches, and it is not clear how quickly the additional TPU resources will come online.

Still, the announcement is a clear sign that the company is planning for higher demand and larger model workloads. In a market where access to infrastructure can be as important as model design, the deal gives Anthropic more room to compete as the AI race continues to intensify.

What to Watch

Investors and customers will be watching for any further details on when the new TPU capacity becomes available and whether Anthropic uses it to accelerate Claude releases, expand enterprise offerings or lower latency for existing products. The broader question is whether other leading AI companies will respond with similarly large infrastructure commitments in the months ahead.