Summary: Around January 9, 2026, OpenAI and SoftBank declared investments of $500 million each into SB Energy — a combined $1 billion — with the same set of announcements bundling a 1.2 GW Stargate data-center lease. SoftBank’s role carries its own weight: OpenAI is not swinging alone but beside a large investor, dividing both the risk and the commitment. For observers, it adds another concrete exhibit that AI capital expenditure is reshaping the energy market: once model companies hold equity in energy producers, the boundary between the two industries has effectively dissolved.
Around January 9, 2026, OpenAI and SoftBank declared investments of $500 million each into SB Energy — a combined $1 billion — with the same set of announcements bundling a 1.2 GW Stargate data-center lease. The package is positioned as the newest addition to the Stargate infrastructure buildout.
The figures deserve a pause. A billion dollars of equity into an energy developer is not a compute purchase; it is a stake in the supply chain that makes compute feasible. The lease converts that stake into guaranteed capacity. Read side by side, the two describe a program planning infrastructure on decade-long horizons.
Three Components of the Deal
Taken apart, the agreement holds:
- Equity: $500 million from OpenAI and $500 million from SoftBank, $1 billion total into SB Energy
- Capacity: a 1.2 GW lease of Stargate data-center capacity
- Strategic position: an extension of the Stargate compute construction program
Why an Energy Company
The AI data-center bottleneck keeps migrating upstream. GPUs can be had for capital; gigawatt-scale power demands grid capacity, land, and time. This deal promotes “power supply” from a procurement relationship to an equity-bound partnership: a model company and a major investor hold shares in the energy producer, then lock future capacity in place with a lease. For a program like Stargate that must scale dependably across years, it is the most direct instrument for shrinking uncertainty on the electricity side.
A sequencing logic runs beneath it as well. Chips can be ordered and delivered inside a procurement cycle; power interconnects and generation cannot. Locking electricity years in advance effectively locks the option to train and serve future models — and the failure to lock it is precisely how expansion plans go quiet and die.
The Logic of Stargate’s Expansion
A 1.2 GW lease announces that Stargate now counts capacity in gigawatts rather than rooms or clusters. SoftBank’s role carries its own weight: OpenAI is not swinging alone but beside a large investor, dividing both the risk and the commitment. When compute demand gets inked at this scale, the entire supply chain behind data centers and power generation must retune its tempo.
For competitors, the message reads: Stargate is not a single campus but a program that compounds — each lease and investment makes the next phase easier to finance and harder to outrun.
The Compute Race Moves Upstream
The 2026 compute race stopped being only about buying GPUs: contracting power, leasing capacity, and investing in energy infrastructure all belong to the battlefield now. For the power and data-center industries, that is unambiguous good news — AI giants are turning into long-horizon, large-scale, predictable demand. For observers, it adds another concrete exhibit that AI capital expenditure is reshaping the energy market: once model companies hold equity in energy producers, the boundary between the two industries has effectively dissolved.
Expect more deals wearing this shape. When the binding constraint is power rather than silicon, the natural move at frontier scale is to stop buying capacity and start owning the inputs to it. The companies that treat energy as a core competency — not a facilities line item — will set the tempo of the next buildout cycle.