NVIDIA is reportedly set to become a major tenant at a one-gigawatt AI campus in Texas, with a lease potentially worth $50.2 billion if all renewal options are exercised. The lease agreement with Hut 8’s Beacon Point campus could mark a big shift in how NVIDIA participates in the data center market. This move signals a transformation in the company’s role from a traditional technology supplier to a key player in the development of large-scale AI infrastructure.
The Texas Lease Deal
The Financial Times identified NVIDIA as the tenant, though NVIDIA has not confirmed its role. Hut 8 has remained noncommittal about the details, and NVIDIA has declined to comment directly on the matter, citing ongoing partnerships and infrastructure deployment.
Each phase is estimated at around $9.8 billion. The lease will fully commercialize the campus and push Hut 8's total contracted AI portfolio to 949 megawatts. The campus is designed with high power density, enabling it to support hundreds of megawatts of AI workloads, particularly for GPU-heavy operations.
Hut 8 projects a $26.6 billion base-term contract value from its entire AI portfolio, with an average annual net operating income exceeding $1.75 billion. This suggests the project is not only about infrastructure but also about long-term financial planning and stability for all parties involved.
Nuclear-Powered Data Centers
Separately, Aalo Atomics and Crusoe announced a partnership to build a nuclear-powered AI data center in Idaho. The goal is to deploy commercial 50-megawatt nuclear-powered AI factories by 2029. This project is designed to show how nuclear energy can power AI operations. The Aalo-Crusoe project is starting small with a modular data center powered by a relatively small advanced reactor. The demonstration aims to address energy limitations and promote scalable, sustainable computing models.
By integrating nuclear energy with AI infrastructure, the partnership proposes a new kind of data center that operates independently of traditional grid constraints. Modular reactors could be deployed incrementally, allowing AI facilities to expand their power output in a controlled and predictable manner. The approach could serve as a prototype for future facilities that combine compute architecture with long-term energy planning and sustainability goals.
NVIDIA’s DSX platform aims to make AI infrastructure more standardized and efficient. The design integrates reference blueprints, hardware, software, and partner tools. This approach promises faster deployments and better performance per megawatt. DSX is not simply a hardware platform but a full-stack solution that includes system design, networking, cooling, and orchestration, all tailored for large-scale AI operations.
The lease structure may also allow NVIDIA to sublease unused capacity to cloud providers and developers. This model could give smaller companies access to large-scale resources without signing their own big leases. It could also help standardize AI deployments around NVIDIA's architecture. For example, a smaller cloud company could purchase capacity from NVIDIA, reducing upfront costs and allowing them to focus on application development rather than infrastructure.
Hut 8 sees the lease as a way to gain a long-term, investment-grade tenant and secure construction financing. Under this model, smaller cloud providers could tap into scarce power and data center resources through NVIDIA. The presence of a high-credit tenant like NVIDIA reduces financial risk and makes the facility more attractive to lenders and investors. This is a key strategy for scaling AI infrastructure across the U.S., particularly in energy-constrained markets.
The $19.6 billion base value is one thing, but the potential for $50.2 billion shows the long-term bet by NVIDIA. If this lease goes ahead, it will significantly expand NVIDIA’s presence in the data center market. The lease is not just about GPUs anymore. Instead, it reflects a broader strategy of integrating hardware, software, and infrastructure, positioning NVIDIA as a foundational element in the AI ecosystem.