Nvidia is pulling back from the most aggressive version of its proposed financial support for OpenAI’s planned Ohio data center, shifting the structure from a potential $250 billion guarantee to less than $120 billion for the project’s initial phase. The change does not indicate that OpenAI has abandoned discussions over the broader 10-gigawatt campus, with the company still discussing a binding lease covering the full development. Instead, the restructuring separates the financing of the first stage from the capital requirements of later expansion, giving Nvidia a narrower financial position while keeping the larger infrastructure ambition intact. The revised arrangement would reportedly cover about 5 gigawatts of the planned capacity rather than the entire campus from the outset. Reuters reported Aug. 14 that Nvidia had reduced its planned support from the previously discussed $250 billion level to less than $120 billion, citing the Wall Street Journal’s reporting.

Nvidia OpenAI Financing Moves To A Phased Model

The revised structure changes the financial architecture around one of the most ambitious AI infrastructure projects now under discussion in the US. Nvidia had previously been considering a guarantee of roughly $250 billion that could support OpenAI’s lease arrangements and the development financing associated with the Ohio campus. That figure immediately drew attention because it would have placed an extraordinary amount of potential financial exposure behind a single AI infrastructure relationship. The new proposal instead places less than $120 billion behind the first portion of the development, with financing for subsequent capacity to be addressed through separate arrangements. That creates a clearer boundary between Nvidia’s commitment to near-term infrastructure and the capital required to build the campus toward its eventual scale. Under the revised structure, financing for later capacity would be considered separately after Nvidia’s initial support for the first phase.

The broader project remains planned at 10 gigawatts, although OpenAI is still discussing a binding lease covering the full development. OpenAI continues discussions around a binding lease covering the full 10-gigawatt development in southern Ohio, while SB Energy remains responsible for developing the campus. The project was first reported as a potential $500 billion-plus undertaking when Nvidia’s proposed guarantee emerged in July, with the overall figure reflecting both infrastructure and computing requirements. Earlier reporting also indicated that Nvidia was considering a separate financing arrangement for processors needed by OpenAI at the site, with chip purchases potentially reaching $350 billion. Those figures demonstrate why the guarantee became central to the financing discussion rather than functioning as a conventional equipment-support arrangement. The project requires capital across power generation, transmission, buildings, cooling systems, networking and accelerated computing, creating a financing problem that extends well beyond the purchase of GPUs.

Nvidia Builds A Broader AI Capital Network

The Ohio development arrives as Nvidia pursues a broader strategy to bring outside capital into the AI infrastructure economy. Earlier in August, Nvidia announced partnerships involving Apollo Global Management, BlackRock, Blackstone, Brookfield Asset Management, Goldman Sachs and KKR aimed at mobilizing more than $500 billion in third-party capital over time for AI infrastructure. The initiative reflects an important change in Nvidia’s role as AI infrastructure moves beyond hardware procurement into power, facilities and long-term financing. Rather than relying primarily on Nvidia’s own balance sheet, the company can connect customers and projects with institutional investors capable of supplying larger pools of capital. That approach allows Nvidia to support demand for its technology without necessarily carrying the entire financial burden of every infrastructure project. The strategy also gives private capital a more direct route into the physical infrastructure required to support AI workloads. 

Meanwhile, the Ohio project provides a practical test for that financing philosophy. Nvidia’s revised arrangement would support the initial phase, while financing for later capacity would be considered separately. The model effectively separates the commercial demand for Nvidia systems from the financing responsibility for the entire campus. That distinction could become increasingly important as AI infrastructure projects move from individual buildings toward multi-gigawatt developments. Institutional investors can finance infrastructure, private equity can provide development capital, banks can arrange debt and technology companies can support equipment procurement. Such a structure distributes risk across participants rather than placing the entire burden on one corporate balance sheet. The Ohio campus therefore becomes a useful case study in how AI infrastructure financing could evolve beyond traditional corporate and project-finance models.

The $350 Billion Chip Question

Nvidia’s potential exposure does not necessarily end with the infrastructure guarantee. Previous reports indicated that Nvidia was also considering financing for processors required by OpenAI at the Ohio campus, with the associated chip purchases potentially reaching as much as $350 billion. That possibility would create another layer of financial interdependence between Nvidia and OpenAI. The infrastructure guarantee would support the environment in which the computing systems operate, while processor financing would support the equipment generating demand for Nvidia’s products. Together, those structures could create an unusually close relationship between a supplier, a tenant and a large infrastructure developer. The scale makes the distinction between strategic partnership and financial exposure especially important for investors evaluating Nvidia’s long-term risk profile.

The combined economics explain why the Ohio project has attracted such scrutiny. A campus approaching $500 billion in potential overall investment would require financing mechanisms capable of supporting multiple layers of infrastructure and equipment expenditure. The reported project cost includes the data center infrastructure and the Nvidia computing equipment required for the planned development. OpenAI’s lease obligation would also need to align with the timing of construction and the eventual availability of operational capacity. That creates a long financing horizon in which assumptions about AI demand must remain sufficiently durable to support the underlying investment. Nvidia’s decision to limit its initial guarantee therefore gives the company more room to reassess those assumptions before supporting the entire development.

OpenAI Gains Infrastructure Control

For OpenAI, the Ohio project remains strategically significant because it would provide greater control over dedicated computing infrastructure. OpenAI has relied on external infrastructure providers including Microsoft and Oracle for access to large-scale computing capacity. A long-term lease covering a dedicated 10-gigawatt development would represent a different infrastructure model, giving OpenAI greater influence over capacity planning, equipment deployment and the physical environment supporting its models. The project could also reduce dependence on capacity that must be shared across broader cloud infrastructure portfolios. OpenAI’s proposed lease remains subject to financing arrangements, with Nvidia’s guarantee under discussion as part of the broader financing structure. Nvidia’s support can help bridge that financing challenge without requiring OpenAI to fund the entire infrastructure build from its own balance sheet.

The structure also highlights the growing importance of long-term leases in AI infrastructure. A conventional cloud contract allows customers to scale usage without taking direct responsibility for an entire physical campus. A dedicated lease reverses much of that relationship by placing the customer closer to the infrastructure itself. OpenAI is discussing a binding lease covering the full 10-gigawatt development, while Nvidia’s revised financing support would initially cover roughly half of the planned capacity. That arrangement makes the creditworthiness of the tenant an important consideration for developers and lenders. Nvidia’s willingness to provide financial support can therefore influence whether the project reaches financial close on commercially viable terms. The revised guarantee keeps that support in place while reducing the amount Nvidia would initially put behind the full development.

A New Boundary For AI Infrastructure Finance

Nvidia’s Ohio decision ultimately points to a broader transformation in how AI infrastructure gets financed. The Ohio project involves multiple parties, including Nvidia, OpenAI, SB Energy and financial institutions involved in the financing discussions. No single participant necessarily needs to fund every layer of the build, particularly when projects combine long construction periods with uncertain future demand. Nvidia has announced partnerships with Apollo Global Management, BlackRock, Blackstone, Brookfield Asset Management, Goldman Sachs and KKR to establish compute-financing platforms intended to mobilize more than $500 billion of third-party capital. The Ohio arrangement fits that model by limiting the company’s initial guarantee while preserving its strategic connection to OpenAI’s deployment plans. The larger question is whether similar structures can support the next generation of AI campuses without creating excessive financial interdependence between technology suppliers and their customers.

Still, the revised commitment should not be interpreted as Nvidia stepping away from OpenAI or from the broader infrastructure expansion. Nvidia and OpenAI already announced a strategic partnership in 2025 focused on deploying at least 10 gigawatts of Nvidia systems, with Nvidia saying it intended to invest up to $100 billion in OpenAI progressively as systems were deployed. The Ohio financing discussions follow the companies’ September 2025 announcement of a proposed partnership covering at least 10 gigawatts of Nvidia systems and up to $100 billion of Nvidia investment in OpenAI. The newer structure reflects a more disciplined approach to the financial mechanics surrounding that relationship as infrastructure requirements become larger.

Ohio Could Set The Pattern For Future AI Campuses

The significance of the Ohio deal extends beyond one campus because future AI infrastructure projects may face the same capital constraints. The Ohio project combines planned data center capacity with power infrastructure and long-term computing requirements.. Hardware suppliers will want to accelerate deployments, but they will also need to manage exposure if customers or projects encounter delays. Financial institutions will increasingly evaluate not only the physical assets but also the credit strength of AI companies occupying them. That combination could make structured guarantees and multi-party financing increasingly common across the sector. Nvidia’s decision to support the first stage rather than guarantee the entire Ohio project provides an early example of how those competing interests can be balanced. 

The next test will be whether the revised structure can move from negotiation to execution without undermining the larger 10-gigawatt plan. Reports indicate Nvidia and OpenAI could finalize an agreement soon, although the terms remained subject to negotiations in the latest reporting. OpenAI continues to pursue the full lease, while Nvidia seeks a financing position that better reflects the risks of supporting such a large project. SB Energy remains central to the physical development, while outside financial institutions could play a greater role in funding later phases. The outcome will provide a clearer signal about how much financial responsibility technology suppliers are willing to assume as AI infrastructure reaches unprecedented scale.