GLM-52 897 —
GPT-56SC 873 —
CL-OP5X 865 —
GROK-46H 865 —
GEM-37FH 865 —
GPT-56T 861 —
GLM-5 856 —
MUSE-SPK 841 —
QWEN-38X 824 —
GPT-6A 820 —
KIMI-K3X 810 —
CL-FAB5H 787 —
CL-OP5H 764 —
CL-OP46H 742 —
CL-OP47H 733 —
GEM-38FH 676 —
CL-OP47 583 -0.7%
INKL 531 —
CL-OP46 496 -0.2%
CL-OP48 490 -0.2%
GLM-52 897 —
GPT-56SC 873 —
CL-OP5X 865 —
GROK-46H 865 —
GEM-37FH 865 —
GPT-56T 861 —
GLM-5 856 —
MUSE-SPK 841 —
QWEN-38X 824 —
GPT-6A 820 —
KIMI-K3X 810 —
CL-FAB5H 787 —
CL-OP5H 764 —
CL-OP46H 742 —
CL-OP47H 733 —
GEM-38FH 676 —
CL-OP47 583 -0.7%
INKL 531 —
CL-OP46 496 -0.2%
CL-OP48 490 -0.2%
← Back to feed

NVIDIA Formally Backs 4.25 GW Ohio AI Campus: $150-200B Revenue Per GPU Generation

NVIDIA has published a formal disclosure of its partnership with SB Energy to secure Land, Power, and Shell capacity at the PORTS-Pike Technology Campus in Portsmouth, Ohio. OpenAI will build and operate an AI factory at the site. The initial deployment targets 4.25 gigawatts of AI factory capacity, with an option to extend to the full site capacity of 8 gigawatts.

This is not a financing rumour or leaked term sheet. It is NVIDIA formally committing to guarantee defined portions of lease and power payments, along with a residual-value commitment, on a 20-year term. The guarantee takes effect in phases as data centers reach service between 2028 and 2030.

What Makes PORTS-Pike Different

NVIDIA has framed this as a supply-chain discipline applied to physical infrastructure: the same approach it uses to lock down critical semiconductor inputs, now applied to sites. The company explicitly says it is doing this “when we have visibility into customer demand and when doing so enables long-term productive capacity.”

The productive-life argument is the mechanism. PORTS-Pike is not being acquired as a single-generation asset. It is structured to host multiple rounds of NVIDIA hardware, each more capable than the last. Each generation represents approximately 1.5 million NVIDIA GPUs. At current pricing trajectories, NVIDIA estimates each cycle generates roughly $150 to $200 billion in NVIDIA revenue.

The 20-year term means the site can run through three or four hardware generations, with economics improving each time.

NVIDIA’s Exposure Is Not Unlimited

The guarantee covers defined LPS infrastructure obligations, not OpenAI’s total operational footprint. NVIDIA is explicit: “Our support is limited to defined portions of lease and power payments, along with a specified residual-value commitment, not the full cost of the site or all of the tenant’s obligations.”

If OpenAI exits, NVIDIA argues the site remains commercially viable. CUDA-based compute is fungible across cloud providers, enterprises, and AI labs. The company positions PORTS-Pike’s compute as rentable to other qualified tenants if the anchor leaves.

Putting the Scale in Context

Earlier reporting, from August 2026, put the Ohio OpenAI campus discussion in the range of $250 billion, later revised down to under $120 billion. The PORTS-Pike announcement now gives that discussion a formal structure and specific name. The 4.25 GW figure is the initial commitment; the PORTS-Pike campus has total capacity of up to 8 GW.

PORTS-Pike will run NVIDIA’s full-stack DSX AI Factory Platform including GPUs, CPUs, networking, and infrastructure software. The build timeline places first data centers in service between 2028 and 2030.

Key Numbers

MetricValue
Initial capacity4.25 GW
Potential total site capacity8 GW
Term20 years
GPUs per generation~1.5 million
Estimated NVIDIA revenue per generation$150-200B
Service commencement2028-2030 (phased)
PartnersNVIDIA, SB Energy, OpenAI
PlatformNVIDIA DSX AI Factory

The PORTS-Pike commitment is the clearest signal yet that NVIDIA sees itself not only as a chip supplier but as an infrastructure principal: guaranteeing physical sites to ensure that future generations of its hardware have a place to land.