Nvidia has agreed to provide a guarantee of up to $105 billion to support OpenAI’s lease of a massive data center campus in Ohio being developed by SoftBank-owned SB Energy, deepening the chipmaker’s role in financing the infrastructure required to sustain the artificial intelligence boom.
Nvidia said Monday it will also invest $1.5 billion in SB Energy, adding to a growing portfolio of infrastructure investments designed to expand the computing capacity available to customers using its chips.
The agreement is seen as the latest indication that Nvidia is moving beyond its traditional role as a supplier of AI processors and increasingly helping finance the data centers, power systems and other infrastructure needed to deploy them. It also raises fresh questions about the financial relationships developing across the AI industry, as chipmakers, cloud providers, model developers and infrastructure companies increasingly invest in one another.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
The Ohio project, in Pike County, will have as much as 8 gigawatts of computing capacity, making it one of the largest AI infrastructure developments announced in the United States. The first 800 megawatts is expected to come online in 2028, while OpenAI will lease the facility for 20 years.
Nvidia will be the site’s exclusive chip supplier.
Nvidia CEO Jensen Huang defended the financing arrangement, rejecting suggestions that it represents circular financing in which AI companies effectively use money from their suppliers and investors to purchase the suppliers’ products.
“We are securing long-lived infrastructure for Nvidia compute so OpenAI can deploy the most productive AI factories that can be upgraded repeatedly with each new generation delivering more intelligence and better economics,” Huang said.
Nvidia said its guarantee will cover only a portion of the project’s lease and power payments, as well as a commitment to maintain a minimum value for the site. It will not guarantee the entire cost of construction or all of OpenAI’s obligations.
Under the proposed structure, OpenAI would remain responsible for rent. If OpenAI defaults, Nvidia would cover the difference between the guaranteed minimum value and whatever SB Energy can recover by leasing or selling the facility.
The final financing structure has not yet been determined, according to people familiar with the matter cited by Reuters. The project is expected to combine equity and debt. Equity could include money raised through a potential SB Energy initial public offering as well as direct investment from SoftBank.
Once the equity component is established, the remaining financing could come from project-finance loans and potentially public debt such as bonds, the people said.
Nvidia’s Infrastructure Bet
The deal represents an unusually large commitment by Nvidia to infrastructure that will ultimately create additional demand for its own processors.
Nvidia has pursued this strategy as AI developers race to secure scarce computing capacity. The company is effectively helping ensure that the physical infrastructure needed to deploy successive generations of its chips is built ahead of demand.
Huang said the Ohio facility could generate as much as $200 billion in Nvidia revenue over time. Including a potential 3.75-gigawatt expansion, Nvidia expects to sell OpenAI a total of 16 gigawatts of computing capacity and could generate as much as $600 billion in revenue from OpenAI through 2030.
That creates a powerful commercial incentive for Nvidia to help solve one of the biggest constraints facing the AI industry: the availability of suitable data centers with enough electricity and grid capacity.
Nvidia’s latest commitment follows a separate initiative announced last week in which the company partnered with six major financial institutions, including BlackRock, to establish financing platforms targeting more than $500 billion in third-party capital for AI infrastructure.
The scale of those commitments has intensified scrutiny over whether the industry’s rapid expansion is being supported by complex financial arrangements among companies that ultimately depend on one another.
“Investors are right to be worried about what seems to be a never-ending loop of AI deals but realistically the field of players isn’t all that vast and there was always going to be a degree of circular financing,” said Danni Hewson, head of financial analysis at AJ Bell.
“The biggest test is whether these investments ultimately generate decent returns for all those laying out cash and that’s something that can only be figured out further down the line,” she said.
Power Emerges As The Next AI Bottleneck
The Ohio project also reveals a challenge that is becoming as important to AI development as chips themselves: electricity. The proposed campus will initially have 4.25 gigawatts of capacity, eventually rising to as much as 8 gigawatts. One gigawatt of computing power is roughly equivalent to the electricity consumption of about 750,000 U.S. homes on average.
To support the development, SoftBank and SB Energy plan to build at least 10 gigawatts of new power generation and spend $4.2 billion on regional grid infrastructure through a partnership with AEP Ohio.
The scale of that investment highlights why AI infrastructure is now being planned around access to electricity rather than simply the availability of land. The U.S. power grid is already facing constraints in some regions, while data center projects have encountered opposition from communities concerned about electricity prices, water consumption, and the broader impact of large-scale facilities on local infrastructure.
The Ohio project is expected to create about 35,000 construction jobs through 2032 and roughly 2,500 permanent operating positions, according to OpenAI.
People familiar with the project said the risk of significant construction delays from local opposition is relatively low because Ohio officials have supported the development because of its expected economic benefits. OpenAI and SoftBank have also committed $80 million to community projects.
The project involves federal land and has participation from the U.S. Departments of Commerce and Energy.
A Test of The AI Infrastructure Economy
The Ohio development is part of a broader race to build what the AI industry calls “AI factories”: large computing facilities designed to train and run increasingly sophisticated models.
While securing these sites can help Nvidia create a predictable market for successive generations of its processors, it is expected to grant OpenAI essential long-term access to computing capacity as it scales its models and AI services.
But the economics of the arrangement will ultimately depend on whether demand for AI computing grows quickly enough to justify the enormous capital expenditure required to build and operate these facilities.
The 20-year lease provides OpenAI with long-term access to capacity, while Nvidia’s guarantee shifts part of the financial risk of the project onto the chipmaker. That creates an unusual alignment of interests. Nvidia benefits when OpenAI expands its computing footprint because the company is the exclusive chip supplier to the Ohio site. OpenAI benefits by securing infrastructure for future model development. SoftBank and SB Energy gain a major long-term customer for a large infrastructure project.
The arrangement therefore ties together the fortunes of three major participants in the AI infrastructure buildout.