SoftBank and SB Energy Commit to 10 Gigawatts of New Power Generation for US Data Centers

A Former Nuclear Site Becomes an AI Power and Compute Campus

US Energy Secretary Chris Wright and Commerce Secretary Howard W. Lutnick joined a groundbreaking on March 20, 2026, for a public-private partnership between the Department of Energy, American Electric Power (AEP), and a SoftBank Group-led consortium anchored by SB Energy. Under the agreement, SB Energy will build 10 GW of new power generation — including at least 9.2 GW of natural gas generation — connected to the local grid, specifically to power a new 10 GW data center development at the Portsmouth Site, a former Department of Energy uranium enrichment facility in Pike County, Ohio.

AEP will build 765-kilovolt transmission infrastructure to move that power, backed by a $4.2 billion commitment from SB Energy. SB Energy is leasing federal land at the site and has also committed to helping fund ongoing environmental cleanup and revitalization of the former uranium enrichment facility, which remains under the DOE’s Office of Environmental Management. Initial development is expected on 189 acres of the site, with the project projected to generate roughly 10,000 construction jobs and more than 2,000 permanent positions once operational.

A Power-First Build for a Named AI Data Center

The Portsmouth campus is not being built speculatively — it is being developed specifically as the site of what DOE materials describe as one of the world’s largest AI data centers, with initial development on 189 acres of the site. That sequencing — a hyperscale power buildout committed alongside a named, purpose-built AI data center rather than built speculatively and leased out later — reflects how AI capacity planning in 2026 increasingly starts with securing power rather than securing chips, since power availability has become the tighter bottleneck in many US regions.

Why Natural Gas, and Why Now

The DOE-Commerce partnership frames the buildout around energy affordability and grid reliability, positioning the new generation as protecting nearby communities from data-center-driven electricity price increases. The heavy reliance on natural gas — at least 9.2 GW of the total 10 GW commitment — reflects a broader pattern across large US AI infrastructure deals in 2026: renewable generation alone cannot be built and interconnected fast enough to match the pace of AI data center demand, pushing developers toward gas generation that can be sited and brought online on a construction timeline measured in a few years rather than a decade.

1. Expect power procurement to increasingly precede compute procurement

The SB Energy structure at Portsmouth shows a sequencing shift: securing multi-gigawatt power generation and grid interconnection now often happens before — or alongside — building out the AI compute campus itself, rather than after a data center site is already built.

2. Watch former industrial and nuclear sites become AI infrastructure targets

The choice of a former uranium enrichment facility — with existing federal land access, grid connections, and DOE relationships — suggests other decommissioned industrial or energy sites in the US could become attractive real estate for similarly scaled AI campuses.

3. Track the natural gas dependency as a policy and emissions flashpoint

With 9.2 of the committed 10 GW coming from natural gas, this deal is likely to draw continued scrutiny over the climate implications of AI infrastructure growth, even as the DOE frames the project around affordability and reliability rather than emissions.

What the Deal Signals for AI Infrastructure Buildout

The Portsmouth project is one of the clearest examples yet of AI infrastructure investment functioning as energy infrastructure investment: a $4.2 billion transmission commitment, a 10 GW generation buildout, and a 10,000-job construction program are being justified primarily by the compute needs of a single long-term tenant. For governments and grid operators elsewhere, the model illustrates both the opportunity (private capital funding new baseload generation and grid upgrades) and the risk (regional electricity markets becoming structurally tied to the capital expenditure cycles of a handful of AI companies and their infrastructure partners).

Frequently Asked Questions

Sources & Further Reading

• DOE Fact Sheet — Ensuring Affordable Energy Access in Ohio While Powering the Future
• DOE — Portsmouth Site SB Energy Partnership
• AEP, SoftBank Ink Deal with US DOE for 10 GW Data Center Power — Rigzone
• BloombergNEF — AI Data Center Build Advances at Full Speed: Five Things to Know