Anthropic is locking up another giant block of AI computing power.
The Claude maker has reportedly agreed to spend $35 billion on cloud-computing capacity from Lambda, an Nvidia-backed cloud provider, at a data center being developed by Hut 8 in Nueces County, Texas, according to The Wall Street Journal.
The roughly 350-megawatt facility will use Nvidia chips supplied to Lambda. The unusual part is that Nvidia itself will hold the lease on the data center, according to the Journal. Lambda will then use the facility to provide computing capacity to Anthropic.
Hut 8 disclosed in July that a high-investment-grade tenant had signed two 15-year leases covering 704 MW at its Texas campus. The leases have a combined base-term contract value of $19.6 billion, although Hut 8 did not name the tenant.
Anthropic’s compute shopping spree
The agreement comes as Anthropic races to secure enough infrastructure to keep up with demand for its Claude products.
The company also reportedly agreed to spend $45 billion on computing capacity from Nscale at a West Virginia data center. Bloomberg also reported that Anthropic has signed $50 billion in cloud commitments with Fluidstack and another $45 billion deal with SpaceX.
The spending reflects a basic problem facing major AI companies: having a powerful model is not enough if there are not enough chips, data centers and electricity available to run it at scale. Anthropic’s push has also created opportunities for specialized cloud providers such as Lambda, which can win enormous contracts while relying on Nvidia’s capital and hardware ecosystem.
Nvidia is more than a chip supplier
The Texas deal shows how Nvidia’s influence is extending beyond selling GPUs. By taking the data center lease itself, Nvidia appears to be helping remove one of the biggest hurdles facing AI infrastructure developers: securing a financially strong tenant capable of supporting a multibillion-dollar project.
That structure could make it easier for Lambda to deliver capacity to Anthropic without having to secure the underlying facility independently. It also gives Nvidia another route to keep its chips at the heart of expanding AI infrastructure.
But the arrangement could also increase Nvidia’s exposure to the infrastructure supporting demand for its hardware, although the financial terms of its agreements with Hut 8 and Lambda have not been disclosed. If AI demand remains strong, the strategy could accelerate deployment and chip sales. If demand weakens, however, Nvidia could face greater exposure to underused infrastructure.
The most important part of the deal may be what it says about the AI industry’s bottleneck. AI companies are no longer competing only for the newest processors. They are competing for ready-to-power data center capacity, and that requires land, electricity, buildings, financing and long-term commitments.
Hut 8 is also developing other Texas facilities for Anthropic that will use Google’s competing tensor processing units, showing that Anthropic is spreading its infrastructure bets across multiple hardware ecosystems.
For Claude customers, the deal does not bring any immediate changes to pricing or product availability. Its longer-term importance is that greater capacity could help Anthropic meet growing demand, improve service reliability and reduce its dependence on any single cloud or chip provider.
Read more: Anthropic is also building an in-house chip team as it looks for more control over the hardware needed to support Claude’s growth.