Nvidia announced a $1.5 billion investment into data center developer SB Energy and provided $105 billion of credit support for an OpenAI project in Ohio. The deal is described as the first implementation of Nvidia’s $500 billion "Land, Power, Shell" (LPS) strategy.
Under the arrangement, SB Energy is responsible for the land, power and building work. Nvidia supplies the chips and also provides a financial backstop: if the tenant (OpenAI) vacates the facility and the assets cannot be resold, Nvidia will cover the resulting shortfall. OpenAI signs the lease for the campus, and Nvidia is the exclusive hardware supplier for 20 years.
The campus is planned to deliver 4.25 gigawatts of capacity. Rather than simply writing a check to a developer, Nvidia effectively purchased the shelf‑space for OpenAI’s GPU purchases over the next two decades, structuring the facilities so that only Nvidia products fit. According to the presented figures, OpenAI will need about $600 billion of Nvidia compute by 2030 and cannot fully finance its own buildout; Nvidia therefore provided financing on the condition that every chip installed belongs to Nvidia.
This transaction is characterized as the first "shelf" under the LPS strategy; the $500 billion plan envisions replicating similar deals for other labs and projects, using the same supplier lock‑in.
Why this matters
The deal changes the balance between suppliers and customers in the AI infrastructure market: Nvidia is acting simultaneously as landlord (providing the real‑estate and capacity arrangement), bank (financing and guaranteeing the buildout) and exclusive hardware vendor. The structure secures long‑term demand for Nvidia’s chips while reducing OpenAI’s immediate capital requirements.
At the same time, the model creates significant dependency: as long as the lease and guarantee structure are in place, the installed hardware and future purchases within these facilities are effectively tied to Nvidia.
Key figures and timeline
- Nvidia investment in SB Energy: $1.5 billion
- Credit support for the OpenAI project: $105 billion
- Campus capacity: 4.25 gigawatts
- Exclusive supplier term: 20 years
- Nvidia LPS strategy target: $500 billion
- Stated OpenAI need for Nvidia compute by 2030: $600 billion
The transaction illustrates how hardware manufacturers can use financial arrangements to lock in market share and assume roles beyond component supply, extending into financing and asset risk management within AI infrastructure.
