Any return-to-service path requires extensive inspections, component replacements, and multi-year coordination with the U.S
Decision Focus
Constellation Energy has advanced regulatory filings and public statements supporting a potential restart of Three Mile Island Unit 1, the Pennsylvania pressurized water reactor closed in September 2019. The company frames the project explicitly around growing electricity demand from AI data centers and large industrial buyers — operators who need firm, carbon-free capacity that variable renewables cannot guarantee continuously. For Global Heads of Data Center Energy, the signal is specific: nuclear restart economics are now being tested against data center PPA structures, and the outcome will define what long-duration baseload supply looks like in PJM for the next decade.
90-Second Brief
In recent days, constellation Energy is pursuing a restart study for Three Mile Island Unit 1, targeting AI data center operators and industrial buyers as anchor offtakers. Any return-to-service path requires extensive inspections, component replacements, and multi-year coordination with the U.S. Nuclear Regulatory Commission, meaning supply would not materialize quickly even if the economics clear. IRA production tax credits for nuclear facilities improve the financial case, and nuclear capacity factors, described in industry sources as often above 90 percent, are positioned as the differentiator over wind and solar in long-term contract structures.
What Is Really Happening?
Low-cost natural gas and flat demand made nuclear economics untenable when Unit 1 closed in 2019. Neither condition holds today. AI compute has pushed load growth projections sharply upward across PJM, and the source article notes that AI-related electricity consumption in the United States could more than double by the early 2030s — a projection that, if directionally accurate, would stress every major market for carbon-free firm power.
Constellation, described in the source article as the operator of the largest fleet of nuclear plants in the United States, is positioning that operational expertise as the differentiator for this demand environment. From a grid standpoint, PJM has a direct interest in any firm, dispatchable zero-carbon resource that reduces reliance on gas peakers as intermittent penetration rises. The alignment between Constellation’s commercial interest, the regional grid’s reliability requirement, and hyperscalers’ procurement pressure gives this proposal structural weight beyond a standard development-stage announcement.
The IRA’s production tax credit framework — applicable to existing and restarted nuclear facilities — reduces the minimum viable contract price Constellation would need to make a restart pencil. Pennsylvania and neighboring states are also evaluating zero-emission credit programs that could stack a second revenue layer. That dual policy support changes the negotiating context: the seller’s cost floor is moving, and contract terms that looked economically impossible two years ago may now be within range.
Why It Matters for Global Heads of Data Center Energy
The first-order impact is on PPA strategy for PJM assets. If a nuclear restart at this scale demonstrates that facility-specific, long-duration clean power contracts can clear economically, it opens a procurement channel without the capacity factor variability or curtailment risk embedded in most renewable structures. For teams running 24/7 CFE programs, anchoring a portion of load to a baseload asset with a 90-plus percent capacity factor is a materially different operational position than relying on hourly matching and REC arbitrage.
The second-order impact is on interconnection exposure. A revived Three Mile Island would inject firm zero-carbon capacity into PJM without the queue delays that new generation assets carry. Data center operators with existing or planned load in PJM gain a potential supplier that sidesteps the three-to-seven-year interconnection timeline constraining nearly every other capacity option. In a market where grid access is the primary growth constraint, that matters.
The third-order impact is contractual. If Constellation secures an AI-linked offtake agreement pricing capacity against facility output rather than a market index, it creates a contract template other nuclear operators and data center energy teams will reference. Operators not in early-stage discussions as anchor positions fill will negotiate from a progressively weaker structural position.
Forward View
If restart economics clear, the immediate watch item is which hyperscaler or colo operator signs the first anchor PPA. That agreement will establish the pricing architecture — capacity price, production credit flow-through, term length — that shapes subsequent deals across the sector. Teams not yet in early discussions face a narrowing window to influence contract structure before anchor positions are committed.
A second front is NRC process pacing. The multi-year coordination requirement is a fixed constraint, but NRC milestone timing determines when firm capacity can realistically enter a forward procurement roadmap. Energy planning teams building three-to-five-year power purchase schedules need to track specific NRC filing progress, not just headline restart news.
A third signal is Pennsylvania’s clean capacity credit framework. State-level support directly affects Constellation’s capital commitment decision — the trigger for everything downstream, including NRC coordination, customer contract execution, and workforce mobilization.
What Is Still Uncertain
No customer commitment is confirmed. The source article states clearly that the restart decision depends on a complex mix of market prices, capital costs, and regulatory conditions. IRA production tax credit eligibility is described as potential, not guaranteed, and precise credit values depend on rulemaking not fully resolved. NRC coordination scope and duration are characterized in general terms with no published filing schedule. Community and political acceptance — covering nuclear risk, waste management, and emergency planning — remains an active debate that economic modeling alone cannot resolve and that can affect both timeline and final investment decision.
One Question for Your Team
If nuclear facility-specific PPAs become a viable procurement channel in PJM within three to five years, does your current long-term power roadmap include a defined position — or a deliberate decision not to take one — and do you have the commercial framework to evaluate what terms would make that position rational at your scale?
Sources
- Ad-hoc-news — AI data centers put nuclear in the spotlight, Constellation Energy’s Three Mile Island plan gains traction (Link)
