Nvidia Provides $105 Billion Credit Support for OpenAI Data Center, Invests $1.5 Billion in SB Energy

TradingKey - Nvidia (NVDA) announced that it will provide up to $105 billion in credit support for a large-scale AI data center campus being built for OpenAI in Pike County, Ohio, and invest $1.5 billion in project developer SB Energy. The project will be built, owned, and operated by SoftBank-backed SB Energy, with OpenAI serving as the primary tenant under a 20-year lease.

According to regulatory filings submitted by Nvidia, OpenAI plans to lease up to approximately 8 gigawatts of computing capacity from the campus, with the first 800 megawatts expected to come online in 2028. The initial phase of the project corresponds to about 4.25 gigawatts of IT load, and Nvidia has the option to further extend its support by approximately 3.75 gigawatts. The campus will adopt Nvidia's full-stack AI factory platform, becoming a key project for its large-scale deployment of AI chips and networking equipment.

It should be noted that the $105 billion is neither a one-time cash investment by Nvidia nor equivalent to new on-balance-sheet debt. The amount primarily represents a cap on credit support for certain lease payments, power costs, and residual value of data center assets. Nvidia would only potentially absorb any shortfall if OpenAI defaults and SB Energy is still unable to cover losses through subleasing or asset sales. Nvidia stated that the support will take effect gradually as the project comes online in phases between 2028 and 2030, and will decrease as OpenAI fulfills its lease payments.

Nvidia had previously discussed an even larger support package. According to media reports, guarantee amounts in early negotiations reached up to $250 billion, though the final disclosed terms scaled back to $105 billion, primarily covering the project's first-phase capacity. The company stated that this arrangement is not "circular financing," emphasizing that OpenAI will bear the lease payment obligations on its own.

The project reflects that competition in AI infrastructure has expanded beyond chip supply to land, power, financing, and data center operational capabilities. For Nvidia, direct investment and credit support help lock in long-term compute deployment and equipment demand; however, as it provides more financing backstops for customers and project developers, the market will also remain focused on the associated off-balance-sheet exposure, customer performance capabilities, and the sustainability of AI compute demand.