Trump DOE gives Microsoft partner $1B loan to restart Three Mile Island reactor

The Trump administration will provide Constellation Energy a $1 billion Department of Energy loan to restart Three Mile Island Unit 1, a project Constellation pegs at $1.6 billion with refurbishment due in 2028 after Microsoft agreed to buy all output from the 835 MW plant for 20 years; analysts at Jefferies estimate Microsoft may pay roughly $110–$115/MWh. That price is cheaper than building new nuclear but represents a substantial premium to wind, solar and even renewables paired with batteries, yet major tech firms (Microsoft and Meta) are increasingly using long-term nuclear offtakes to secure 24/7 power for data centers and AI growth. The financing comes via the DOE Loan Programs Office—whose loan track record and an IRA-created reinvestment program have enabled similar deals—underscoring how federal credit plus corporate contracts can make nuclear restarts investable despite cost headwinds, with implications for utility capital allocation and grid reliability planning.

Analysis

The Trump administration approved a $1.0 billion Department of Energy loan via the Loan Programs Office to Constellation Energy to restart Three Mile Island Unit 1, a project Constellation pegs at $1.6 billion with refurbishment expected to complete in 2028 and Microsoft contracted to buy all output from the 835 MW plant for 20 years. Jefferies estimates Microsoft may be paying roughly $110–$115 per MWh under that offtake, a level described as cheaper than building new nuclear but materially higher than wind, solar, or renewables paired with batteries per Lazard cost comparisons.

The economics reflect a corporate desire—particularly among hyperscalers—for reliable 24/7 clean power to support data centers and AI growth, evidenced by a contemporaneous Meta deal for 1.1 GW of nuclear clean attributes. That demand can make nuclear restarts investable despite a premium to intermittent renewables, but the margin between PPA price and alternative LCOEs highlights sensitivity to power-market pricing and corporate willingness to pay a reliability premium.

The LPO financing materially reduces capital availability risk for Constellation but carries program reputational history (Solyndra) and a reported 3.3% default rate after recoveries; recent comparable LPO activity includes a $1.6 billion loan to AEP. Investors should therefore weigh improved credit-backed project viability against construction, regulatory and political execution risks and monitor milestone delivery to 2028.

Key Decisions for Investors

• Consider selective, modestly sized exposure to Constellation Energy (CEG) or utility restart financings where a long‑term corporate PPA and LPO credit materially de‑risk cash flows, while monitoring capex execution against the $1.6bn budget and 2028 timeline
• Evaluate Microsoft (MSFT) exposure qualitatively—an implied $110–$115/MWh offtake signals a corporate willingness to pay for 24/7 clean power that could support future offtakes but is a meaningful premium to wind/solar LCOEs
• Use hedges or staging (options, tranches) to protect positions from construction, cost‑overrun or regulatory delays given LPO program political sensitivity and the legacy of Solyndra despite a reported 3.3% default rate after recoveries
• Monitor follow‑on transactions (Meta, AEP) and Lazard LCOE trends as leading indicators for broader capital reallocation toward nuclear restarts; increase conviction only if additional corporates sign similar long‑dated PPAs