OpenAI plans 10 GW Ohio data center, Nvidia may provide funding

Serge Bulaev

Serge Bulaev

OpenAI is in talks to lease land in Ohio for a huge data center that might use as much electricity as several million homes.

OpenAI plans 10 GW Ohio data center, Nvidia may provide funding

Reports confirm that the OpenAI plan for a 10 GW Ohio data center is advancing, with negotiations now underway for what could be the world's largest AI computing facility. According to reporting from The Information and Reuters, this ambitious project, potentially financed by Nvidia, would consume power equivalent to several million homes and redefine the financial and logistical scale of hyperscale infrastructure.

The proposed deal involves a 20-year lease on federal land in southern Ohio, where SoftBank-affiliated SB Energy would develop the site. OpenAI would manage the computing hardware, with Nvidia reportedly set to provide both GPUs and crucial financial support. Nvidia is expected to provide hardware and financial guarantees for the Ohio project, with the estimated construction cost of the campus reaching approximately $500 billion.

How a 10 GW campus compares

OpenAI is negotiating a 20-year lease for a 10-gigawatt (GW) data center campus in southern Ohio. This single site, potentially backed by Nvidia, would consume more electricity than many existing data center hubs and marks a significant shift toward massive, single-tenant AI infrastructure projects.

A 10-gigawatt power draw would dwarf today's largest cloud campuses. For perspective, Northern Virginia's data center market had an estimated capacity of over 10 GW (likely 12 - 15 GW) in 2025, making it the world's largest hub. If fully realized, the Ohio site alone would represent a substantial addition to global data center capacity. This project signals a strategic pivot from typical 100-300 MW hyperscale builds to immense, AI-optimized campuses that integrate compute, power, and land into a single package.

Grid and infrastructure hurdles

Accommodating a 10 GW load presents significant challenges for Ohio's power grid. Utilities have already expressed concerns over existing data center growth, with AEP Ohio pausing new contracts in 2024 to demand more customer investment in upgrades. The regional grid operator, PJM, has projected potential capacity shortages in the coming years. Industry reports suggest proposed solutions include building substantial new gas-fired power generation capacity and investing billions in transmission upgrades, though these plans are not yet approved.

Anticipated impacts on the region include:
- Construction of extensive high-voltage transmission lines.
- Expansion of natural gas pipeline capacity and interconnections.
- Potential for higher electricity bills if costs are socialized across the customer base.
- Substantial water requirements for cooling systems.

Financing model with Nvidia's imprint

Nvidia's potential dual role as both hardware supplier and financier is a pivotal aspect of the deal. This structure suggests a new, vertically integrated model: the chipmaker secures demand, the AI developer guarantees long-term compute access, and the energy provider receives crucial credit support. Analysts see this as a departure from traditional short-term leases, moving instead toward multi-decade contracts tied directly to power and performance milestones.

What remains uncertain

Despite the advanced stage of talks, the project is far from guaranteed. Key documents, including signed leases, environmental impact reviews, and binding grid-interconnection agreements, have not been finalized. Both Reuters and Network World caution that negotiations could still fail or the project's scope could be significantly reduced. While industry reports suggest the first phase could begin in the coming years, this timeline is subject to delays from permitting, supply chain issues, and potential local opposition. For now, the 10 GW Ohio campus stands as a landmark concept, highlighting how the relentless demand for AI is pushing the data center industry toward unprecedented scale.