OpenAI Makes Its Move in Ohio: 10GW Super Data Center Finalized, with Nvidia Backing It with a $1.5 Billion Investment

OpenAI has locked in a major new computing footprint in the United States, agreeing to a 20-year lease for a 10-gigawatt data center campus in Ohio. The project, developed by SoftBank-backed SB Energy, is positioned to become one of the largest AI computing sites ever planned, underscoring how aggressively top AI companies are securing long-term infrastructure for training and inference.

The deal goes beyond a standard lease. Nvidia is also participating directly, investing $1.5 billion in SB Energy in exchange for equity. In addition, the chipmaker is providing a form of value backstop for the project’s assets, a structure designed to make it easier for the developer to raise debt financing for such a large-scale buildout.

A massive Ohio campus built around long-term AI demand

At the center of the agreement is a data center park with a planned total capacity of 10 gigawatts. OpenAI will serve as the long-term tenant under a lease running two decades, giving the project a durable demand anchor and offering OpenAI a stable base of computing power for future AI workloads.

In practical terms, the scale alone sets this development apart. A campus of this size would rank among the biggest AI infrastructure projects disclosed so far, reflecting a broader shift in the industry: leading model developers are no longer relying only on short-term capacity arrangements and are instead moving to secure deep, dedicated access to power and compute over many years.

Nvidia adds capital and a financing safety layer

Nvidia’s role in the transaction is notable because it extends beyond supplying chips. By purchasing an equity stake in SB Energy with a $1.5 billion investment, Nvidia is helping support the project from the developer side as well.

More importantly for lenders, Nvidia is also backing a portion of the asset value. The initial support applies to roughly 5 gigawatts of phase-one capacity, provided Nvidia becomes the exclusive chip supplier for the first half of the campus. The company also retains the option to expand that support to the full 10 gigawatts later on.

This structure is meant to strengthen confidence among financing partners. Even in a scenario where OpenAI does not renew its lease in the future, the completed facilities are still expected to retain meaningful value. That expectation can improve the project’s financing profile, reduce borrowing costs, and make debt fundraising more achievable for SB Energy.

The risk-sharing terms are tightly defined

While Nvidia’s involvement is substantial, the arrangement does not amount to an open-ended guarantee. The disclosed framework places clear limits on the company’s exposure and ties support only to facilities that have already been completed.

The protection also works through a staged process rather than an immediate payout. If OpenAI were to leave, SB Energy would first need to try to re-lease the space at the original price. If that failed, the next step would be to sell the relevant assets. Only if the final sale value came in below the project’s assessed value would Nvidia cover the shortfall.

For the first phase, the cap on that backstop is around $105 billion, defining the outer boundary of Nvidia’s commitment. In other words, Nvidia is helping de-risk the project for financing purposes, but within a framework that sharply limits how far its obligations can extend.

A sign of intensifying competition for AI infrastructure

The agreement is the latest example of increasingly sophisticated financing structures emerging around AI infrastructure. As demand for large-scale compute accelerates, companies across the stack are using their balance sheets in more creative ways to secure access, support development, and reduce execution risk.

In recent years, major chip and technology companies including Nvidia and Google have repeatedly helped underpin projects linked to leading AI firms such as OpenAI and Anthropic. These arrangements often serve two purposes at once: they give developers a stronger foundation for raising money against long-term leases, while also allowing strategic backers to keep their own risks within defined limits.

According to the disclosed timeline, negotiations continued for several weeks before the parties finalized the agreement on Monday. The three sides also adjusted the size of the backstop during those talks, responding to concerns about how much exposure Nvidia would ultimately take on.

Taken together, the Ohio project highlights how the AI race is no longer just about models and products. It is increasingly about who can secure the land, power, hardware, and financing needed to build computing capacity at unprecedented scale.