Nvidia is in discussions to provide a financing backstop of up to $250 billion to help OpenAI lease a massive 10-gigawatt data center campus in southern Ohio. The project, overseen by SoftBank’s SB Energy, could cost more than $500 billion in total including hardware.
The proposed facility in Pike County, Ohio, represents a strategic pivot for OpenAI as it seeks to control more of its own physical infrastructure rather than relying solely on rentals from cloud giants like Microsoft, Amazon, and Oracle. The site, which once functioned as a uranium-enrichment plant, is envisioned as one of the largest data centers in the world by power capacity.
The $250 Billion Credit Backstop and Hardware Financing
The core of the negotiations isn’t a simple purchase agreement, but a balance sheet maneuver. Nvidia is discussing a guarantee that would allow OpenAI to raise debt for the campus lease and construction on the strength of Nvidia’s own credit. According to CNBC, this $250 billion backstop would specifically cover the lease and construction debt, not the chips themselves.
The hardware costs are being handled through a separate stream of discussions. The Wall Street Journal reports that Nvidia is discussing separate financing support for OpenAI chip purchases that may total $350 billion. This distinction separates a product transaction from a long-term financial commitment.
This structure creates a circular financing loop: Nvidia provides the financial strength for OpenAI to build the infrastructure, which in turn creates guaranteed, long-term demand for Nvidia’s own systems. Gary Tan, a portfolio manager at Allspring Global Investments, noted the tension this creates for the market.
SoftBank and the 10-Gigawatt Ohio Campus
The project is being developed by SB Energy, a subsidiary of SoftBank Group Corp. To put the scale of the 10-gigawatt target into perspective, CNBC analysis of Energy Information Administration data suggests that 10 gigawatts is roughly equivalent to the annual power consumption of 8 million U.S. households.
The site’s viability depends heavily on government cooperation and international funding. The campus is located on federally owned land, and the power is controlled by the U.S. government. Under a recent trade deal, Japan has agreed to fund $33 billion for a natural gas plant to support the site. The U.S. and Japan will share power sale proceeds until Japan recoups its investment, after which the U.S. will hold 90% control.
- First Phase: Expected completion by 2028, providing approximately 800 megawatts of power.
- Full Scale: A total capacity of 10 gigawatts, dwarfing other projects like OpenAI’s recently announced $20 billion, 3.2-gigawatt project in Effingham County, Georgia.
Access to this power is currently a political decision. Commerce Secretary Howard Lutnick is reportedly deciding which firms get access, and Anthropic, Microsoft, and Google have all expressed interest in the site.
Nvidia’s Broader Infrastructure Strategy
This potential deal is part of a wider pattern where Jensen Huang is investing across the AI supply chain to remove bottlenecks. Beyond the Ohio project, Nvidia is investing $1 billion in Naver Corp. for an AI data center and collaborating with SK Group to build over 2 gigawatts of capacity on the Korean Peninsula.
The relationship between Nvidia and OpenAI has been volatile. In September 2025, the two signed a letter of intent to deploy 10 gigawatts of systems, with Nvidia intending to invest up to $100 billion progressively. However, that investment never materialized, though Nvidia did contribute $30 billion to OpenAI’s funding round in March.
As OpenAI moves toward a potential public offering—having confidentially filed with the SEC in June—Nvidia’s role as a financier may be reaching a limit. Jensen Huang stated that this latest move might be the last time
the company invests in OpenAI before it goes public.
Whether this $500 billion gamble pays off depends on whether the demand for frontier AI models continues to outpace the available power and chips. While the Ohio project could lock in years of hardware demand for Nvidia, it also ties the chipmaker’s balance sheet to OpenAI’s ability to monetize its models in an increasingly competitive market featuring open-weight alternatives from China.
The ultimate question remains whether the U.S. government and the Commerce Department will grant OpenAI the primary power allocations needed to make the 10-gigawatt vision a reality, or if the site will be split among the other tech giants currently vying for the land.
