NVIDIA is scaling back the financial backing it had planned for a massive OpenAI data center project in Ohio, according to reports, reducing its proposed guarantee from about $250 billion to less than $120 billion. The revised arrangement would initially cover only the first phase of the development, easing NVIDIA’s exposure to one of the most ambitious AI infrastructure projects under consideration in the United States.
The change comes as investors increasingly scrutinize the enormous capital requirements behind the artificial intelligence boom. The Ohio facility is expected to eventually reach 10 gigawatts of power capacity and could require hundreds of billions of dollars in infrastructure and computing investment. NVIDIA’s decision to reduce its initial financial commitment highlights the growing importance of managing risk as AI companies and their infrastructure partners attempt to build data-center capacity at unprecedented scale.
NVIDIA Scales Back Ohio Data Center Guarantee
NVIDIA had been discussing a financial guarantee of roughly $250 billion to support OpenAI’s lease of a proposed data-center campus in Ohio.
Under the revised plan, NVIDIA is expected to provide less than $120 billion in initial backing, according to reports citing people familiar with the discussions.
The guarantee would help support leasing and construction-related debt for the project. Rather than backing the entire planned facility immediately, NVIDIA would initially support the first phase, with decisions about later phases expected to be made separately.
| Key Detail | Information |
|---|---|
| Companies involved | NVIDIA and OpenAI |
| Location | Ohio, United States |
| Planned capacity | Up to 10 gigawatts |
| Earlier NVIDIA guarantee | About $250 billion |
| Revised initial backing | Less than $120 billion |
| Initial phase | About 5 gigawatts |
| Developer | SB Energy |
| Parent company | SoftBank |
| Expected use | AI computing infrastructure |
| Status | Deal reportedly nearing finalisation |
The revised structure is intended to reduce NVIDIA’s financial exposure while allowing the first stage of the project to move forward.
First Phase Could Cover 5 Gigawatts
The Ohio project is planned as a 10-gigawatt data-center campus, making it one of the largest AI infrastructure developments ever proposed.
Under the revised arrangement, NVIDIA would initially provide financial backing for roughly half of that planned capacity, or about 5 gigawatts.
The remaining capacity could be addressed later as OpenAI and its infrastructure partners gain greater visibility into demand, financing and construction requirements.
Planned Development
10-gigawatt campus
↓
First phase
↓
Approximately 5 gigawatts
↓
NVIDIA financial backing
↓
OpenAI computing capacity
↓
Future phases
↓
Remaining capacity
This phased approach could help reduce the financial risks associated with committing capital to the entire project before all of its requirements are known.
OpenAI Is Still Expected to Lease the Full Capacity
Despite NVIDIA reducing its initial guarantee, OpenAI is reportedly still expected to sign a binding lease covering the project’s full 10-gigawatt capacity.
The distinction is important.
NVIDIA’s financial guarantee and OpenAI’s planned lease represent separate elements of the overall arrangement.
OpenAI would commit to the infrastructure, while NVIDIA’s backing would help support the financing required to develop it.
Project Structure
SB Energy
↓
Develops data center
↓
OpenAI
↓
Leases computing capacity
↓
NVIDIA
↓
Provides financial backing
↓
Banks and lenders
↓
Finance construction and infrastructure
This structure would allow the project to move forward without requiring NVIDIA to finance the entire development directly.
Why NVIDIA Is Reducing Its Commitment
The decision comes amid growing investor concern about NVIDIA’s expanding financial exposure to the AI infrastructure ecosystem.
NVIDIA has increasingly moved beyond selling GPUs and into financing and investing in the infrastructure required to deploy those chips.
While these investments can strengthen demand for NVIDIA hardware, they also expose the company to risks associated with customers’ ability to finance and utilise increasingly expensive data centers.
A smaller guarantee could therefore allow NVIDIA to maintain a strategic relationship with OpenAI while limiting the amount of capital it puts at risk.
The AI Infrastructure Boom Requires Massive Capital
AI companies are rapidly increasing their computing requirements.
Training and operating advanced models requires enormous quantities of GPUs, electricity, cooling equipment, networking hardware and physical data-center space.
As models become more capable, AI companies are seeking increasingly large computing campuses.
AI Infrastructure Chain
AI models
↓
More computing demand
↓
More GPUs
↓
More data centers
↓
More electricity
↓
More financing
↓
Greater financial risk
The Ohio project illustrates how the AI boom is increasingly becoming an infrastructure and financing challenge rather than simply a software-development race.
The Ohio Project Could Cost Hundreds of Billions
The full Ohio campus could become part of a development costing hundreds of billions of dollars when construction, power infrastructure and AI chips are included.
The project’s planned 10-gigawatt capacity makes it particularly capital-intensive.
For comparison, a data center of this scale requires substantially more infrastructure than conventional facilities designed for ordinary cloud computing.
Scale of the Project
10 GW
↓
Large-scale power infrastructure
+
Thousands of AI servers
+
NVIDIA GPUs
+
Networking systems
+
Cooling infrastructure
+
Buildings
↓
Hundreds of billions in potential investment
The enormous scale explains why financing arrangements are becoming an increasingly important part of AI infrastructure development.
NVIDIA Is Becoming More Involved in AI Financing
NVIDIA’s role in the Ohio project comes as the company expands its involvement in AI infrastructure financing.
The chipmaker recently partnered with major financial institutions including Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to develop financing platforms aimed at mobilizing more than $500 billion in capital for AI infrastructure.
The initiative reflects NVIDIA’s belief that financing constraints could become one of the biggest obstacles to the next stage of AI expansion.
NVIDIA’s Expanding Role
GPU supplier
↓
Infrastructure partner
↓
Investor
↓
Financing facilitator
↓
AI ecosystem platform
This strategy could help NVIDIA create more demand for its processors while helping customers secure the capital required to deploy them.
Investor Concerns Are Increasing
NVIDIA’s decision to reduce the Ohio guarantee also illustrates the growing scrutiny surrounding AI infrastructure spending.
The AI industry has attracted enormous amounts of capital, but investors are increasingly asking whether projected AI demand will justify the cost of building massive data centers.
Companies must consider:
- GPU utilisation
- Electricity costs
- Financing costs
- AI model demand
- Customer commitments
- Hardware depreciation
- Construction costs
- Long-term revenue visibility
If demand grows as expected, large data centers could become highly valuable infrastructure assets.
If demand falls short, however, companies could face substantial unused capacity and financial obligations.
OpenAI Needs More Computing Capacity
For OpenAI, the Ohio project is part of its broader effort to secure enough computing infrastructure to support future AI models and services.
The company’s AI systems require enormous amounts of computing power both during training and when responding to users.
As ChatGPT usage grows and OpenAI develops increasingly capable models, access to large-scale computing infrastructure becomes strategically important.
OpenAI’s Compute Requirements
More users
↓
More inference
↓
Larger models
↓
More training
↓
More GPUs
↓
More data centers
↓
More power
The Ohio campus could therefore provide OpenAI with a significant long-term source of computing capacity.
SoftBank’s SB Energy Is Developing the Facility
The data-center project is being developed by SB Energy, a SoftBank subsidiary.
The facility is planned for southern Ohio and is expected to become one of the largest AI infrastructure projects in the country.
The project involves substantial power and construction requirements, meaning its development will depend on coordination among technology companies, energy providers, lenders and government authorities.
Development Ecosystem
SoftBank
↓
SB Energy
↓
Data-center development
+
OpenAI
↓
Long-term capacity
+
NVIDIA
↓
AI hardware and financing
+
Financial institutions
↓
Debt and capital
The structure demonstrates how modern AI infrastructure requires cooperation across multiple industries.
NVIDIA Still Benefits From OpenAI’s Expansion
Even with the reduced guarantee, NVIDIA has a strong strategic incentive to support OpenAI’s infrastructure growth.
OpenAI is one of the world’s largest buyers of advanced AI computing hardware.
More data-center capacity means more opportunities for NVIDIA to sell GPUs, networking systems and complete AI computing platforms.
Strategic Relationship
NVIDIA financing
↓
OpenAI infrastructure
↓
More computing capacity
↓
More NVIDIA systems
↓
Higher chip demand
The relationship therefore extends beyond the financing agreement itself.
The Deal Highlights Circularity Concerns
Large AI infrastructure transactions have increasingly attracted attention because technology companies can simultaneously act as investors, suppliers and customers.
NVIDIA can provide financing support while also supplying the GPUs that populate the resulting data centers.
OpenAI can commit to purchasing computing capacity while using NVIDIA hardware to operate its models.
AI Infrastructure Cycle
NVIDIA
↓
Financing
↓
Data center
↓
NVIDIA GPUs
↓
OpenAI
↓
AI services
↓
Revenue
↓
More computing demand
↓
More NVIDIA hardware
This structure can accelerate growth, but investors may question how much of the ecosystem’s expansion is being supported by external demand versus financing arrangements among industry participants.
Financing Is Becoming as Important as Hardware
The shift demonstrates how the AI industry’s next bottleneck may not simply be access to chips.
Companies also need capital to build the physical infrastructure required to use those chips.
The industry therefore increasingly depends on:
- Banks
- Private-equity firms
- Infrastructure investors
- Energy companies
- Cloud providers
- Chipmakers
- AI companies
This could turn AI data centers into a major asset class for institutional investors.
AI Data Centers Are Becoming Infrastructure Assets
Traditional data centers were generally built around cloud computing, enterprise applications and internet services.
AI data centers require significantly greater power density and specialised computing infrastructure.
This is creating a new category of infrastructure investment.
Traditional Data Center
Cloud workloads
↓
Moderate power requirements
↓
Standard computing
VS
AI Data Center
AI training
+
AI inference
↓
High-density GPUs
↓
Extreme power demand
↓
Advanced cooling
↓
Large infrastructure investment
The Ohio project is an example of this transition.
The Reduced Guarantee Does Not Mean the Project Is Cancelled
NVIDIA’s decision to reduce its guarantee should not be interpreted as a cancellation of the Ohio data center.
The revised arrangement is instead designed to phase NVIDIA’s financial support.
OpenAI remains expected to pursue the full 10-gigawatt lease, while NVIDIA would initially back only the first stage.
This could allow construction to proceed while giving both companies more time to evaluate the economics of later phases.
What It Means for NVIDIA
For NVIDIA, the move represents a more cautious approach to AI infrastructure investment.
The company can continue supporting OpenAI’s expansion while limiting its direct financial exposure.
That could reassure investors concerned about NVIDIA becoming too deeply involved in financing its own customers.
At the same time, NVIDIA remains strategically tied to the growth of AI data centers because its chips and networking systems are central to many of these facilities.
What It Means for OpenAI
OpenAI still gains access to a potentially enormous source of future computing capacity.
However, the company will need to manage the financial obligations associated with leasing and operating such a large facility.
The economics of the project will ultimately depend on whether demand for AI services continues to grow fast enough to justify the infrastructure.
What It Means for the AI Industry
The revised Ohio agreement is another sign that the AI infrastructure boom is entering a more financially disciplined phase.
The industry’s earlier focus was largely on securing GPUs and announcing massive capacity targets.
The next phase will increasingly involve questions about:
- Who pays for infrastructure?
- Who guarantees the debt?
- How quickly will capacity be used?
- Who bears the risk if AI demand slows?
- How long will the hardware remain economically useful?
These questions will become more important as AI infrastructure commitments reach hundreds of billions of dollars.
Key Numbers at a Glance
| Metric | Figure |
|---|---|
| Planned Ohio capacity | 10 gigawatts |
| Earlier NVIDIA guarantee | About $250 billion |
| Revised NVIDIA backing | Less than $120 billion |
| Initial phase | About 5 gigawatts |
| Potential full project scale | Hundreds of billions of dollars |
| Developer | SB Energy |
| Parent company | SoftBank |
| Main customer | OpenAI |
| Main AI hardware supplier | NVIDIA |
| Expected lease | Full 10-gigawatt capacity |
Infographic: NVIDIA and OpenAI Ohio Data Center
OPENAI
↓
10-GIGAWATT OHIO DATA CENTER
↓
SB ENERGY
↓
DATA CENTER DEVELOPMENT
↓
NVIDIA
↓
REVISED GUARANTEE
↓
LESS THAN $120 BILLION
↓
FIRST PHASE
↓
ABOUT 5 GIGAWATTS
↓
OPENAI LEASES CAPACITY
↓
NVIDIA SYSTEMS
↓
AI COMPUTING
What Investors Should Watch
Investors should watch whether NVIDIA and OpenAI finalize the revised agreement and whether construction proceeds according to schedule.
Other important indicators include:
- Final financing terms
- OpenAI’s lease commitments
- Construction progress
- GPU deployment
- Data-center utilisation
- AI demand growth
- NVIDIA’s infrastructure investments
- Power availability
- Financing costs
- Expansion plans for later phases
The performance of these projects could influence how investors value the broader AI infrastructure boom.
The Bigger Picture
NVIDIA’s decision to reduce its proposed financial guarantee for OpenAI’s Ohio data center reflects a growing shift in the AI industry from aggressive infrastructure expansion toward more carefully structured financing. The original proposal would have exposed NVIDIA to roughly $250 billion of financial backing, while the revised arrangement would initially provide less than $120 billion for the project’s first phase. The move allows NVIDIA to remain closely involved with OpenAI’s infrastructure expansion without immediately taking on the full financial risk of a 10-gigawatt campus.
The development also highlights the enormous capital requirements behind the AI boom. Building advanced AI infrastructure requires not only GPUs but also power, cooling, buildings, networking and long-term financing. As companies announce increasingly large data-center projects, investors are likely to focus more closely on utilisation, returns and who ultimately bears the risk. NVIDIA’s partnership with major financial institutions to mobilize more than $500 billion for AI infrastructure shows that financing is becoming a central part of the industry’s growth strategy.
Looking Ahead
The Ohio project remains strategically important for OpenAI because access to massive computing capacity will be essential as the company expands its AI models and services. A phased financing structure could allow the first 5 gigawatts of capacity to move forward while giving OpenAI, NVIDIA and their financial partners more time to assess demand and the economics of later stages. If AI demand continues growing rapidly, the remaining capacity could eventually be developed with additional financing arrangements.
For NVIDIA, the reduced guarantee could represent a more disciplined approach to its increasingly broad role in the AI ecosystem. The company has powerful incentives to support new data centers because those facilities create demand for its GPUs and networking systems, but excessive financial exposure could create risks for shareholders. The Ohio deal therefore offers a test of whether NVIDIA can continue accelerating AI infrastructure growth while keeping its own balance-sheet and investment risks under control.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
