Nvidia in Talks for $250 Billion Financing Guarantee for OpenAI Data Center

Nvidia is in advanced talks to provide roughly $250 billion in financing guarantees for OpenAI, backing a 10-gigawatt data center project in southern Ohio developed by SoftBank's SB Energy, the Wall Street Journal reported Sunday. The entire project, including chips, is expected to cost more than $500 billion, marking OpenAI's first move to own its infrastructure instead of renting from Microsoft, Amazon, and Oracle. The deal would lock in demand for Nvidia's next-generation GPUs and could reshape AI infrastructure financing, though analysts warn of risks if AI demand plateaus.

July 27, 2026 , Inside AI — Nvidia is in advanced discussions to provide roughly $250 billion in financing guarantees for OpenAI, backing a colossal data center project in southern Ohio that would mark a pivotal shift in AI infrastructure ownership, the Wall Street Journal reported Sunday. The guarantees would help OpenAI lease a 10-gigawatt project developed by SoftBank’s energy subsidiary, SB Energy, according to people familiar with the matter. The entire project, including the chips inside the data center, is expected to cost more than $500 billion . The deal would allow OpenAI to control its own infrastructure for the first time, moving away from renting capacity from Microsoft, Amazon, and Oracle. For Nvidia, it would lock in chip demand for years, securing a customer for its next-generation GPUs at an unprecedented scale. The $250 billion backstop covers the data center lease and debt financing but not the Nvidia chips themselves. Nvidia is also separately discussing financing OpenAI’s chip purchases, potentially worth up to $350 billion , the Journal reported. These discussions involve creating financing vehicles to reassure lenders about the project’s viability. “The backstop from Nvidia would help the ChatGPT maker lease a 10-gigawatt project that SoftBank's energy subsidiary is developing in southern Ohio,” the Journal stated, citing sources. The first phase, delivering around 800 megawatts of power, is expected to be completed in 2028 . The power is controlled by the U.S. government and funded separately by Japan under a recent trade deal tied to Tokyo’s pledge to invest $33 billion in a natural gas plant. U.S. Commerce Secretary Howard Lutnick is involved in deciding who gets access, the report said. OpenAI has been in advanced talks for several weeks to lease the site and is among the companies showing the strongest interest. Anthropic, Microsoft, and Google have also spoken to Lutnick recently, signaling a broader race for dedicated AI infrastructure. The financing structure reflects a growing trend: tech giants are increasingly tapping debt and equity markets to fund AI infrastructure, with spending set to top $700 billion this year. Nvidia’s role as a guarantor is unusual, effectively making it a financial backer for a customer, which could reshape how chipmakers and cloud providers negotiate large-scale deals. This project’s scale dwarfs existing data centers. For context, the largest current U.S. data center campus, operated by Switch in Nevada, has a planned capacity of 1.2 gigawatts . A 10-gigawatt facility would consume as much power as a small country, raising questions about grid impact and regulatory hurdles. The Ohio site’s reliance on a natural gas plant funded by Japan also ties AI expansion to international energy diplomacy. OpenAI’s move to own infrastructure mirrors Amazon’s early strategy with AWS, where controlling the underlying hardware became a competitive moat. However, the financial risk is immense: if AI demand plateaus, OpenAI could be saddled with underutilized capacity, and Nvidia’s guarantees could strain its balance sheet. Analysts at McKinsey https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier have estimated that generative AI could add up to $4.4 trillion annually to the global economy, but such projections depend on sustained adoption. Competing viewpoints highlight the gamble. Some industry observers see the deal as a necessary step to overcome the compute bottleneck for training next-generation models like GPT-5. Others warn of a potential overbuild, reminiscent of the telecom bubble, where massive infrastructure investments outpaced actual demand. A recent research paper https://arxiv.org/abs/2303.08774 on scaling laws suggests that model performance improves predictably with compute, but the economic returns diminish as costs skyrocket. Reuters could not immediately verify the report. Nvidia, OpenAI, and the U.S. Commerce Department did not respond to requests for comment outside regular business hours. The Ohio project’s power arrangement adds another layer of complexity. Under the U.S.-Japan trade deal, the natural gas plant funding is tied to broader economic cooperation, potentially giving Tokyo leverage over AI infrastructure access. This intertwines commercial AI ambitions with geopolitics, as nations vie for control over critical compute resources. If finalized, the deal would be one of the largest infrastructure financing arrangements in history, surpassing even the $100 billion Stargate AI project announced earlier this year. It signals that the AI arms race is moving from software to physical assets, with Nvidia positioning itself as both a supplier and a kingmaker.