Nvidia in Talks to Provide $250 Billion Guarantee for OpenAI Super Data Center Project, Total Scale Nears $500 Billion

The artificial intelligence infrastructure race is heating up on an unprecedented scale. Chip giant Nvidia is in advanced discussions with AI research company OpenAI to provide approximately $250 billion (approximately $250 billion) in financing guarantees for a massive data center project, according to The Wall Street Journal, citing people familiar with the matter. The guarantee arrangement would help OpenAI lease computing capacity from a sprawling computing hub being developed by SoftBank Group's energy subsidiary SB Energy in southern Ohio.
The data center's planned capacity reaches a staggering 10 gigawatts, far exceeding the scale of virtually all current data centers globally. SoftBank founder Masayoshi Son has previously stated that the total cost of the entire project could reach approximately $500 billion. Tech publication The Information also reported, citing sources, that at current chip, power, and construction prices, the campus would cost at least $500 billion to build. If Nvidia's guarantee agreement is ultimately finalized, it would become one of the largest financing arrangements in tech history and mark a significant shift for the chip giant—from a pure hardware supplier to a central financial architect of AI infrastructure.
Deal Structure: One Guarantee Links Three Parties
Sources familiar with the matter revealed that Nvidia's approximately $250 billion guarantee would serve as a credit enhancement instrument, simultaneously covering OpenAI's lease obligations and the developer's project financing, thereby reducing lenders' concerns about repayment. According to The Information, OpenAI would control the computing equipment within the campus under a 20-year lease, with rent payments commencing once the site becomes operational. The first phase is expected to come online in 2028. Nvidia is also expected to serve as the hardware supplier. This arrangement allows OpenAI to lock in the massive computing resources it needs for years to come—to support its growing model training and inference demands—without directly bearing the full construction costs.
| Deal Element | Details |
|---|---|
| Guarantor & Scale | Nvidia, approximately $250 billion |
| Guarantee Coverage | OpenAI's lease obligations, plus developer's project financing |
| Lessee | OpenAI, controlling computing equipment at the campus |
| Developer | SoftBank Group's SB Energy |
| Site | Southern Ohio, former U.S. Department of Energy uranium enrichment plant site (Portsmouth area, Pike County) |
| Planned Capacity | 10 gigawatts, with at least 9.2 GW from natural gas generation |
| Total Construction Cost | At least approximately $500 billion |
| Lease Term | 20 years, commencing after campus becomes operational |
| First Phase Launch | Expected 2028 |
| Supporting Grid | SB Energy and AEP Ohio jointly investing $4.2 billion to upgrade and build new transmission lines in southern Ohio |
From 2025 Partnership to Today's Guarantee
This guarantee is not an isolated event but an extension of Nvidia and OpenAI's existing collaboration. In September 2025, the two companies announced a partnership to deploy at least 10 gigawatts of Nvidia systems, with Nvidia also expressing intent to cumulatively invest up to $100 billion in OpenAI—funds to be deployed in tranches as each gigawatt of capacity comes online. The first phase of that collaboration is planned to utilize Nvidia's Vera Rubin platform. By that measure, the roughly $250 billion guarantee now under discussion is approximately 2.5 times the upper limit of that investment commitment.
Under the same collaboration framework, SoftBank Group's SB Energy committed to building 10 gigawatts of new power generation capacity, with at least 9.2 gigawatts from natural gas, alongside billions of dollars in new transmission infrastructure. In January 2026, OpenAI and SoftBank Group further announced a $1 billion investment in SB Energy to support data center construction for the Stargate project.
Site Background: From Uranium Enrichment Plant to Computing Hub
The site selected by SoftBank in southern Ohio carries strong industrial transformation symbolism—it is the former location of a uranium enrichment plant under the U.S. Department of Energy. U.S. Commerce Secretary Howard Lutnick described it at the project's announcement as "a $33 billion power project in Portsmouth," placing it within the framework of Japan's $550 billion investment commitment to the U.S. As part of the project, SB Energy and AEP Ohio are jointly investing $4.2 billion to upgrade and build new transmission lines in southern Ohio. The U.S. Department of Energy noted that the project aligns with commitments requiring data center companies to bear grid infrastructure costs themselves. AEP CEO Bill Fehrman stated that the partnership brings critical infrastructure to Ohio in the Appalachian region and "will not increase customer electricity rates."
Notably, when the U.S. Department of Energy initially announced the power project, it did not identify the lessee. If the current reports prove accurate, they fill in the missing piece of the puzzle for this 10-gigawatt campus.
Responses and Risk Controversies
Reuters confirmed The Wall Street Journal's reporting in its own coverage but stated it was unable to independently verify the report's contents. Neither Nvidia nor OpenAI immediately responded to requests for comment, and the lease itself remains under negotiation, with issues surrounding financing, permitting, and deployment timelines yet to be resolved.
If the deal is ultimately completed, its impact will extend far beyond the three companies involved. A $250 billion guarantee is unprecedented in the technology sector, reflecting not only the extreme hunger for computing power in AI but also the enormous capital mobilization capabilities of leading tech firms. The key question in the AI race is shifting from "who has the best chips" to "who can finance and power the factories that run those chips."
However, analysts have raised warnings about such structures. Neil Shah, Vice President of Research at Counterpoint Research, noted that such "symbiotic deals" are increasingly becoming the norm in AI infrastructure rollouts, but cautioned that enterprise customers should not accept bundled infrastructure wholesale. Instead, they should demand hybrid use of multiple computing providers to avoid placing all their bets on a single ecosystem. He also warned that a 10-gigawatt campus cannot be built overnight—full completion would take at least a decade. "Making long-term commitments based on such a timeline comes with enormous uncertainty," he said. Other analysts argue that the sharper risk lies on the financing side, which will ultimately be passed downstream in the form of minimum commitment levels, reservation tiers, and usage thresholds. "Scarcity doesn't disappear—it just gets transformed into contract terms."
Beyond this, a project of this magnitude still faces multiple uncertainties, including power supply stability, regulatory approvals, and the return-on-investment timeline for AI. With computing demand for products like OpenAI's ChatGPT continuing to surge and global AI chip supply remaining tight, these debates are unlikely to be settled anytime soon.
Once added, BigGo Finance appears first in Google Search Top Stories, so you get the broadest, most up-to-the-minute, and most comprehensive global financial news first.