Chip titan Nvidia (NVDA) has agreed to provide a financial guarantee of up to $105 billion to back OpenAI’s 20-year lease for a multi-gigawatt data center in Pike County, Ohio, developed by SoftBank’s SB Energy.
📊 Key Capital Figures & Infrastructure Metrics
- Nvidia Financial Guarantee: Up to $105 billion covering lease, power payments, and site residual value for OpenAI.
- Direct Equity Investment: Nvidia will invest $1.5 billion directly into SB Energy (following a $1 billion joint investment by OpenAI and SoftBank).
- Power Capacity Scale: Total planned site capacity of up to 8 gigawatts (GW), with the initial 800 megawatts (MW) online by 2028.
- Regional Power Expansion: SoftBank and SB Energy plan to build ≥10 GW of new power generation and invest $4.2 billion in local grid infrastructure with AEP Ohio.
- Massive Revenue Revenue Potential:
- Initial 4.25 GW capacity could generate ~$200 billion in revenue for Nvidia.
- Long-term potential of $600 billion in revenue by 2030 through supplying 16 GW of computing power to OpenAI across site expansions.
- Economic Impact: Projected to create ~35,000 construction jobs through 2032 and 2,500 permanent operational jobs.
💡 Strategic Vendor Dynamics & Grid Bottlenecks
- Exclusive Hardware Deployment: Nvidia secures status as the exclusive chip provider, locking in long-term compute sites that can be upgraded across future GPU generations.
- Solving Power Constraints: Direct intervention addresses critical land, power, and grid capacity bottlenecks currently slowing U.S. hyperscale data center development.
- Vendor Financing Debate: CEO Jensen Huang rejected claims of “circular financing,” stating Nvidia is deploying its scale to secure long-lived physical infrastructure for its next-generation AI compute factories.
💡 The Strategic Takeaway
Nvidia is using its massive balance sheet to act as both chip supplier and financial backstop for next-generation AI infrastructure. By guaranteeing power and lease liabilities, Nvidia ensures that power-constrained AI labs can deploy compute at scale, locking in hundreds of billions in future chip demand.