With Google’s support, Anthropic’s $15 billion AI data-centre in Texas is fast becoming a reality. It is the latest sign of how the race to put in place the infrastructure for artificial intelligence is picking up speed and the scale of the capital involved.
Reports have it that the Texas project is on track to put in place a $15 billion financing package, with Google at the centre of making it happen. To the technology community and investors, this is not just another funding announcement. It is an indication that the future of AI will be as much about having access to vast computing resources and the right financial structures as it is about the models themselves.
A Financing Model Built for the AI Era
The Wall Street Journal and others say the debt package will come to some $15 billion, made up of a $14 billion bridge loan and a revolving credit facility. Morgan Stanley and a group of major banks are heading up the talks.
Google is doing more than offering its technology here. The company is set to underwrite billions in lease and power purchase agreements for the campus’ own power plant; if Anthropic were to default, Google would be on the hook for those obligations. In return, Google is likely to take an equity position of around 20% in the data-centre and power venture, if not in Anthropic proper. For any lender, the fact that Google brings its investment-grade credit to the table makes a project of this magnitude commercially sound.
Why Google Is Doubling Down
There is method to Google’s involvement. By putting up the money, it allows Anthropic to put down roots in infrastructure that would be hard to fund on its own, cementing the ties between the two firms. The Texas site is also where Google’s Tensor Processing Units will be deployed via a vendor-financing deal with Broadcom.
It is a pattern you see across the industry now: frontier AI firms are looking to partners for their compute needs rather than trying to own the whole ecosystem.
One of America’s Largest AI Infrastructure Projects
The Hubbard, Texas campus is no small undertaking. Covering 2,800 acres, it is being laid out to be among the biggest AI data-centre campuses in the country. At its heart will be a 1.6-gigawatt natural gas plant to feed the high-density electricity required for training and inference. Add in the hyperscale buildings, the cooling and networking, and room for GPU and TPU installations, and one has the makings of a major operation.
That is a far cry from earlier estimates of a $5 billion project with 600 megawatts of capacity. The jump to $15 billion and 1.6 gigawatts speaks to what next-generation models are demanding.
Why This Matters for the Future of AI
Anthropic’s plan is a case study in how AI infrastructure of the future will be put together. Instead of shouldering every cost, the labs provide the demand and the tech while institutions like Google and various banks supply the strategic and financial muscle.
For Anthropic, it means expanding without straining the balance sheet before a public offering. For Google, it is a way to fortify its TPU ecosystem and lock in a position in a premier AI project. As the industry moves forward, the edge will go to those who can build and finance the hardware behind the smartest models, not just the ones who make them.
