Nvidia defends OpenAI Ohio deal as $600B compute opportunity
Nvidia CEO Jensen Huang defended the company's financial support for OpenAI's Ohio AI campus against circular financing claims. The deal includes a $1.5 billion investment in SB Energy and secures up to 8 IT-GW of capacity. Huang sees a $600 billion compute opportunity with OpenAI through 2030.

*this image is generated using AI for illustrative purposes only.
Nvidia (NASDAQ: NVDA) CEO Jensen Huang pushed back against characterizations of "circular financing" surrounding Nvidia's partnership with SB Energy to build an AI campus in Ohio for OpenAI. Huang clarified that OpenAI will pay the lease directly, framing Nvidia's support as a mechanism to secure scarce infrastructure rather than finance demand.
Huang noted that frontier AI labs are growing faster than their balance sheets and long-term credit profiles can support. He estimated that OpenAI's existing and planned deployments could represent roughly $600 billion in Nvidia compute through 2030. This figure assumes repeat business across multiple hardware upgrade cycles at the Ohio site, with each generation potentially involving about 1.5 million Nvidia GPUs.
Deal Structure And Financial Commitments
The project involves SB Energy building, owning, and operating the data center at the PORTS-Pike Technology Campus in Ohio under a 20-year lease agreement. Nvidia has invested $1.5 billion in SB Energy immediately to support the development. The company provides credit support for the initial 4.25 IT-GW of capacity and holds an option for the remaining 3.75 IT-GW, bringing the total potential capacity to 8 IT-GW.
According to reports, Nvidia initially considered guaranteeing OpenAI's full data-center commitments upfront, potentially around $250 billion, before scaling back the structure amid investor concerns over risk exposure. Under the current arrangement, if OpenAI leaves the site, SB Energy would seek another tenant or buyer, with Nvidia covering any remaining loss in value up to $105 billion.
| Deal Parameter | Details |
|---|---|
| Initial Capacity Secured | 4.25 IT-GW |
| Option Capacity | 3.75 IT-GW |
| Total Potential Capacity | 8 IT-GW |
| Primary Customer | OpenAI |
| Lease Duration | 20 years |
| Nvidia Investment In SB Energy | $1.5 billion |
| Community Benefits Fund | $80 million |
| Max Loss Coverage By Nvidia | $105 billion |
Revenue Potential And Market Context
Huang stated that each hardware generation at the Ohio site could generate $150 billion to $200 billion in revenue for Nvidia. The campus will exclusively host Nvidia GPUs, CPUs, networking, and software. OpenAI expects to fund its lease commitments through revenue, cash flow, and investor capital. Nvidia also agreed in February to invest $30 billion in OpenAI directly.
Polymarket traders currently assign OpenAI about a 20% chance of completing an IPO this year, suggesting private capital may remain central to the buildout. The $600 billion figure represents Nvidia's estimate of the broader OpenAI opportunity, not a single signed contract.
What the Numbers Show
The structure of the partnership highlights a shift towards hyperscalers providing direct financial backing for physical infrastructure. By providing credit support for land, power, and shell buildout rather than merely leasing finished space, Nvidia is de-risking the capital expenditure for SB Energy. The immediate $1.5 billion investment into SB Energy further aligns the two companies' interests, ensuring that the financial health of the data center operator is directly supported by the chipmaker's commitment to securing exclusive compute capacity for its key customer, OpenAI.
How might Nvidia's $105 billion loss coverage exposure impact its balance sheet stability and credit ratings if OpenAI fails to meet lease obligations?
Will other hyperscalers like Microsoft or Google adopt similar infrastructure financing models, potentially shifting the industry standard from leasing to direct capital backing?
What are the implications for Nvidia's exclusivity clause if competitors develop comparable AI infrastructure solutions that bypass the Ohio campus?

































