Nvidia backs OpenAI Ohio data-center leases with reported $105 billion guarantee tied to 4.25 GW of capacity

Nvidia is reported to have agreed to guarantee up to $105 billion in lease obligations tied to OpenAI’s planned AI data-center buildout in Ohio, according to a market report published by Yahoo Finance and carried by BeinCrypto. The figure is described as covering commitments across 4.25 gigawatts of data-center capacity, a scale that underscores how the next phase of artificial intelligence infrastructure is increasingly financed through long-term power and real-estate commitments.

The report frames the guarantee as a backstop for lease obligations rather than as an outright purchase or direct operating-cost promise by Nvidia. In other words, the guarantee is presented as a credit-enhancement mechanism intended to reduce the risk borne by the leasing side if the project’s lease payments are not fully met. That distinction matters because it changes what portion of the overall economic burden sits with Nvidia versus project sponsors and other counterparties.

Capacity, in this context, is measured in gigawatts, which reflects the electrical power that facilities can draw to run servers and supporting equipment. For AI data centers, power availability is often a gating factor because large-scale model training and inference require dense compute plus power-hungry cooling, networking, and power distribution systems.

The report’s headline number is the reported $105 billion guarantee, which is described as linked to the lease commitments associated with the Ohio buildout. It is also described as spanning 4.25 GW of capacity, suggesting a multi-site or campus-wide plan rather than a single facility. The Ohio location would be expected to require major grid interconnection and utility coordination, which can take time and capital, making long-term lease structures a common way to match financing with construction timelines.

Nvidia, for its part, is the supplier most associated with the hardware layer of modern AI systems, particularly the graphics processing units (GPUs) and accelerated computing components used to train and run large machine-learning models. As AI demand has surged, Nvidia has benefited from the expansion of data-center capacity, and financing arrangements that make new campuses bankable can indirectly support the broader demand environment for AI infrastructure.

That said, the market report provides limited detail on the contractual mechanics and does not, in the available excerpt, specify the exact entities that are receiving the guarantee, the trigger conditions, the duration of the leases, or the performance milestones that would activate or limit the guarantee. It also does not clarify whether the guarantee is partial and stepwise, capped against particular loss scenarios, or structured through multiple credit instruments.

From a business perspective, a reported guarantee of this size, if accurate and fully representative of the final deal terms, indicates that strategic AI infrastructure is being underwritten at increasingly large scale. It also highlights how hardware suppliers are not only selling equipment but, in some cases, participating in the financing ecosystem that enables capacity buildouts, especially when the projects depend on complex power and long-term leasing.

What to watch next is whether any parties involved publish additional documentation, regulatory disclosures, or investor commentary that confirms the guarantee amount, the counterparties, and the timing. Market-sensitive terms like guarantee duration, exposure limits, and whether incentives are tied to capacity delivery are often the key variables that determine how much financial risk is truly transferred, and whether the arrangement evolves into further commitments.