NVIDIA

NVIDIA plans to spend as much as $105 billion to secure land, power and shell capacity for the PORTS-Pike Technology Campus in Piketon, Ohio, a massive data center campus being developed on the site of a defunct federal uranium enrichment facility near Columbus, Ohio. NVIDIA is investing to up to $105 billion to secure land, power and shell (LPS) capacity for the massive PORTS-Pike Technology Campus planned on the site of the former Portsmouth Gaseous Diffusion Plant, a long-shuttered federal uranium enrichment facility near Columbus, Ohio. 2 GW of which would come from natural gas.

As ENR reported in March, Bechtel and Kiewit Corp. have already been tapped to provide engineering, procurement, construction and other services for the project. Under the arrangement, SB Energy, a subsidiary of SoftBank Group , will develop, own and operate the data center campus, including the power generation and delivery systems as well as the physical buildings. NVIDIA will supply its computing technology that the tenant, OpenAI, will use to power large-scale AI workloads. 5 billion in SB Energy to support its growth. LPS capacity refers to the infrastructure required to launch high-performance data centers.

Land refers to the physical site; powers refers to the large-scale electrical supply and grid connectivity required to operate AI systems; and shell describes the constructed data center buildings furnished with core mechanical and electrical systems, ready for installation of computing equipment. "Our role in the project is really the development," SB Energy CEO Rich Hossfeld told CNBC. "We're securing the land. We are doing the power solution and we are going to build the shells—the turnkey centers ... if you think of a maybe $500-billion project, our role is 30 to 40 percent of that."

"AI factories require a full stack of critical resources: advanced chips, packaging, memory and networking—as well as land, power and shell," NVIDIA CEO Jensen Huang said in a blog post . "Just as NVIDIA has used its scale, long-term visibility and supply-chain partnerships to secure critical semiconductor resources, we are now applying that same discipline to secure LPS capacity exclusively for NVIDIA AI factories." 25 gigawatts of AI factory capacity. 5 million NVIDIA GPUs, or [about] $150 billion to $200 billion in NVIDIA revenue. Over 20 years, the site can support multiple upgrade cycles."

The agreement reflects a growing industry trend in which access to land and power is a primary constraint on data center development, particularly for AI applications that require much higher energy capacity than traditional computing.