Google is expected to support a large Texas data center campus leased by Anthropic, with the first phase of the project estimated to cost more than $5 billion. The move ties one of the fastest-growing AI developers more closely to one of its biggest cloud partners as US companies compete for computing capacity.
The site is operated by Nexus Data Centers. According to the report, Google is expected to provide construction loans, while a group of banks is competing to arrange additional financing by mid-year. Anthropic recently signed a lease for the 2,800-acre campus, and construction is already underway. Early debt financing came from Eagle Point.
Initial capacity targets 500 megawatts with room to scale far higher
Under current plans, the campus is expected to deliver about 500 megawatts of capacity by late 2026, with the potential to expand to 7.7 gigawatts later. That scale shows how the AI infrastructure race now extends well beyond chip supply and cloud contracts. Power, land and financing are all becoming critical inputs.
The location is also near major gas pipelines, which could allow the operator to use on-site gas turbines instead of relying only on the public grid. For large AI training facilities, stable power delivery and expansion capacity matter as much as hardware access, especially as training and inference loads keep rising.
Google and Anthropic push their partnership deeper into infrastructure
The Texas buildout adds to an existing relationship between the two companies. Anthropic said in October 2025 that it would expand its use of Google Cloud TPUs and related services, with plans to access up to 1 million TPUs for training and serving its Claude models.
Google’s expected role in the Texas project shows that cloud providers and model developers are moving into heavier infrastructure commitments, not just service agreements. For Anthropic, this kind of buildout is tied directly to future model capacity. For Google, it extends its position across a larger part of the AI compute stack.
Pentagon dispute remains active as court blocks risk label for now
At the same time, Anthropic is still fighting a legal dispute with the Pentagon. A federal judge in San Francisco blocked the Pentagon from branding the company a “supply-chain risk” while the case proceeds. Judge Rita Lin said the government’s action appeared punitive rather than security-driven, and reporting cited her description of the process as “arbitrary.”
The ruling does not require the Pentagon to keep using Anthropic’s tools, but it does stop broader punitive measures for now. The conflict grew out of a dispute over military use of Anthropic’s AI after the company refused to loosen safeguards related to surveillance and autonomous weapons. US military units used Anthropic’s Claude AI during strikes on Iran, placing the company at the center of both the race for AI infrastructure and the debate over how governments use advanced AI systems.