Billions in funding to go toward redeveloping former Ohio uranium site into new hub

More than $33 billion will be invested to redevelop the former Portsmouth Gaseous Diffusion Plant into a 10-gigawatt AI data-center hub under a public–private partnership with the U.S. government, SoftBank, and SB Energy. The plan includes a new natural gas power plant to supply the site and a $14 million SoftBank investment in local schools, with sponsors citing national-security-grade AI capacity and significant regional job creation potential.

Analysis

This project will re-price the marginal economics of building AI infrastructure in non-coastal US markets by internalizing generation, transmission, and land risk in a single site. That vertical integration favors large balance-sheet owners and EPCs who can carry long construction timelines and take on remediation/permits — expect differential bid premiums of 15–30% on major turnkey contracts versus piecemeal builds, and multi-year lock-ins on long-term power contracts.

A material second-order effect is on the regional power and gas markets: a sustained, concentrated load will push local capacity prices and basis differentials wider, creating outsized near-term earnings for merchant generators and pipeline owners that can secure firm transport and capacity contracts. Conversely, it will compress margins for renewables that cannot guarantee dispatchable baseload without complementary firming, increasing demand for dual-fuel or gas-backed generation and long-duration storage.

Regulatory, permitting, and national-security reviews are the single largest reversal risk and create a chunky timeline: expect high-probability schedule slips measured in quarters-to-years rather than weeks. Construction wage inflation, localized housing tightness, and supply-chain premium on heavy electrical gear will reallocate value to EPCs and equipment manufacturers while making smaller local contractors cost-constrained — a classic winner-take-most outcome for the first wave of contractors and partners.

Key Decisions for Investors

• Long Digital Realty (DLR) — 12–24 month horizon. Rationale: scale player with balance sheet to capture long-term leases and wholesale AI demand. Entry: on any pullback to the 50-day moving average or on a 8–12% drawdown; target +30–40% upside if leasing accelerates, downside -20% if hyperscaler capex slows.
• Long NRG Energy (NRG) or Vistra (VST) — 6–18 month horizon. Rationale: merchant generators able to supply firm, dispatchable capacity and capture higher local wholesale spreads and capacity revenues. Position: buy incrementally as forward spark spreads widen; expected upside +25–50% in a scenario of multi-year firm power demand, tail risk is gas-price surge compressing margins (manage with 5–7% position sizing).
• Long Jacobs Engineering (J) or Fluor (FLR) — 12–36 month horizon. Rationale: large EPCs that win turnkey remediation, grid interconnection and data-center construction work. Trade: buy on evidence of awarded contracts or fiscal-quarter guide-ups; asymmetric payoff as fixed-fee contracts reprice higher, downside limited by backlog diversity.
• Options play on Nvidia (NVDA) — 9–18 month horizon. Rationale: core AI hardware exposure benefits if large-scale US AI campuses drive sustained GPU demand. Trade: buy a calendar or vertical call spread (long-dated call, sell shorter-dated higher strike) to capture sustained demand while limiting premium paid; reward skewed materially higher if enterprise AI deployments accelerate, risk limited to premium paid.