OpenAI and SoftBank Expand AI Infrastructure Through $1B SB Energy Partnership

What was announced

On January 9, 2026, OpenAI and SoftBank Group announced a strategic partnership with SB Energy to expand AI and energy infrastructure in the United States. According to OpenAI, the transaction is part of the Stargate framework and is intended to speed deployment of large AI data center campuses tied to rising compute demand. The announcement frames this as infrastructure execution rather than a model launch, but the practical effect is significant: it secures power-linked capacity for future AI workloads.

The companies said OpenAI and SoftBank Group are each investing $500 million in SB Energy, totaling $1 billion. OpenAI also said it signed a 1.2 GW data center lease for initial buildout in Milam County, Texas, and selected SB Energy to build and operate that site. In parallel, SB Energy disclosed an additional $800 million of Redeemable Preferred Equity from Ares to support growth. Taken together, the financing package combines equity capital, leasing commitments, and project-level infrastructure execution.

Why this matters for AI operations

For AI operators, the key takeaway is tighter coupling between model roadmaps and power-aware data center development. OpenAI’s statement emphasizes speed, cost discipline, and integrated energy delivery, signaling that access to electricity, land, and construction timelines is becoming a strategic bottleneck equal to GPUs. In that context, a 1.2 GW initial lease is not a routine colocation update; it is a utility-scale planning decision aligned with long-horizon inference and training demand.

OpenAI also linked the deal to the previously announced $500 billion Stargate commitment introduced in January at the White House. While the SB Energy release does not claim that all Stargate capital is deployed, it shows concrete execution steps: named partners, a defined site, and stated financing. The companies additionally formed a non-exclusive preferred partnership model for future builds, which implies repeatable project structures rather than one-off procurement.

Execution signals to monitor

• Whether initial facilities enter service starting in 2026 as described by SB Energy.
• How fast Milam County construction translates into usable AI compute capacity.
• Whether the non-exclusive preferred model expands to additional U.S. campuses.
• How community commitments on jobs, workforce development, and grid modernization are implemented locally.

OpenAI and SB Energy also stated that the Milam County design aims to minimize water usage and add new generation capacity to support energy needs while protecting Texas ratepayers. That operating detail matters because future AI infrastructure plans are increasingly evaluated on resilience, permitting feasibility, and local grid impact, not only on hardware performance.