Nvidia moves deeper into AI infrastructure
Nvidia said Monday it will invest $1.5 billion in SB Energy, a data center and power developer tied to SoftBank and OpenAI. The investment gives Nvidia a central role in OpenAI’s Ports-Pike data center project near Cincinnati, Ohio: according to SEC filings cited in the report, Nvidia will be the sole supplier of compute infrastructure for the facility.
In this context, “compute infrastructure” means the hardware layer that runs AI workloads, including accelerated servers, chips, and related systems. For a company like OpenAI, that infrastructure is not just equipment; it is the physical capacity that determines how much AI training and inference can be supported over time.
A deal built around chips, credit, and scale
The agreement goes well beyond a conventional investment. Nvidia will also provide up to $105 billion in credit to help build the facility. The project could begin at 4.25 gigawatts and eventually expand to 8 gigawatts, according to Nvidia’s SEC documents. Gigawatts are a measure of power capacity and are increasingly used to describe the scale of the largest AI data center developments.
Key figures from the project include:
- Nvidia investment in SB Energy: $1.5 billion
- Nvidia credit support: up to $105 billion
- Initial data center scale: 4.25 gigawatts
- Potential expanded scale: 8 gigawatts
- Planned natural gas power plant: 9.2 gigawatts
- Estimated power plant cost: $33 billion
SB Energy’s existing investors include SoftBank and OpenAI. SoftBank previously held $5.8 billion worth of Nvidia stock, which it sold in November to help fund other AI investments. That detail highlights how tightly connected the AI infrastructure ecosystem has become: capital, chips, cloud capacity, and model development are increasingly financed through overlapping relationships.
Power supply becomes a strategic constraint
The Ports-Pike project is notable because the data center is paired with a major power development. SB Energy plans to build a 9.2-gigawatt natural gas power plant on land owned by the U.S. Department of Energy. The site previously enriched uranium for the U.S. nuclear arsenal and U.S. Navy submarines.
The power plant is expected to cost $33 billion. BloombergNEF says the cost of building natural gas power plants has risen 66% over the past two years, helping explain the scale of the budget. For AI data centers, energy availability is becoming as important as chip supply. Large model training requires sustained high-intensity computation, while inference demand grows as AI services gain users.
Energy markets may feel the pressure
The report also notes that by the time SB Energy’s plant and others are completed, they may be competing for natural gas with export markets. That convergence could triple natural gas prices in some parts of the country. The point is not that one data center will determine national energy prices, but that AI infrastructure expansion can spill into power generation, fuel supply, capital spending, and public land use.
For Nvidia, the deal secures future demand for its hardware and extends its influence beyond chip sales into financing and infrastructure. For OpenAI and its partners, guaranteed access to Nvidia systems and large-scale credit support may reduce uncertainty around one of the most difficult parts of AI expansion: turning capital plans into usable compute capacity.
The next AI race is physical
This investment shows that the AI race is no longer only about model design or chip performance. It is becoming a contest over chips, credit, electricity, and land. Large AI developers may increasingly align with semiconductor suppliers and energy developers to secure long-term capacity. At the same time, gas prices, construction costs, permitting, and regional power constraints will shape how fast new AI data centers can come online. Nvidia’s move is both a supply-chain lock-in and a bet that demand for AI compute will keep rising.