NVIDIA is in talks to provide a financing guarantee of up to $250 billion that would help OpenAI lease capacity at a massive AI data center campus planned for southern Ohio, a proposal that highlights the escalating scale of AI infrastructure investment and the evolving approach to financing that buildout.

The project, developed by SoftBank subsidiary SB Energy, is expected to become one of the world’s largest AI data centers. The proposed campus is designed to deliver 10 gigawatts of electrical capacity, with total investment potentially exceeding $500 billion when construction costs and AI hardware are included.

NVIDIA’s role would not be to directly finance construction. Instead, the chipmaker would provide a financial guarantee that strengthens OpenAI’s ability to secure financing for long-term leasing arrangements. Because OpenAI does not hold an investment-grade credit rating, backing from NVIDIA could lower borrowing costs for lenders financing the project and improve access to capital.

Additionally, NVIDIA is reportedly discussing financing that would help OpenAI purchase the NVIDIA GPUs planned for the Ohio facility. Those chip purchases could ultimately total about $350 billion, a vivid example of the unprecedented investment driving today’s AI sector.

For OpenAI, the Ohio campus would be an important strategic shift. Rather than relying primarily on hyperscalers like Microsoft and Amazon for computing resources, the company would lease dedicated infrastructure directly, providing greater control over compute capacity as demand for model training and inference keeps rising.

Located on federal land at the former Portsmouth Gaseous Diffusion Plant near Piketon, the campus has access to government-backed power supported by new natural gas generation and transmission infrastructure. The first construction phase is expected to provide roughly 800 megawatts of capacity by 2028 before expanding toward the planned 10-gigawatt buildout.

Concerns about Circular Financing

The proposed financing structure demonstrates how AI project funding is changing. Large tech companies are more often using their own balance sheets to support financing for customers and partners, an approach sometimes referred to as a “credit wrapper.” The model allows infrastructure developers to secure financing on better terms while helping tech vendors boost demand for their own products and services.

Google has previously employed similar arrangements to support AI infrastructure partners, while NVIDIA has expanded investments across the AI ecosystem through equity stakes and financing initiatives designed to increase use of its GPU platforms.

However, the financing model has drawn criticism from investors concerned about so-called circular capital flows within the AI sector. Critics argue that tech companies financing infrastructure that ultimately purchases their own products would amplify financial exposure if large projects fail to generate expected returns.

“NVIDIA has done a number of deals to support AI datacenters, most of which use their chips,” Jack Gold, President of J. Gold Associates, told Techstrong.it. “This is another example of circular financing. NVIDIA has lots of cash on hand and it’s easy for them to throw some of that cash toward companies looking to expand buildouts, which then helps NVIDIA sell more chips at very high margins.”