The Outlook for U.S. Nuclear: Restarts, Uprates, New Build, and When Ratepayers Feel It

Noreva, an energy market data and analytics provider (formerly Karbone Research), tracks capacity forecasts across six ISOs and power forecasts across all seven U.S. ISOs, and nuclear has moved from a retirement story to a build story faster than most models assumed. A Michigan reactor is loading fuel to become the first U.S. plant to return from decommissioning. A Pennsylvania unit shuttered in 2019 is restarting under a 20-year Microsoft power purchase agreement. Wyoming has its first reactor under construction in the state’s history. This piece answers the four questions Noreva’s clients keep asking: what is actually new, how likely new nuclear is to get built, why the market is shifting back toward it, and if it does not get built, when ratepayers feel the impact through capacity and energy prices.

What’s Actually New With Nuclear in the U.S. Right Now?

Three restarts are furthest along. Constellation announced a 20-year power purchase agreement with Microsoft on September 20, 2024 to restart Unit 1 of Three Mile Island, renamed the Crane Clean Energy Center, covering roughly 835 MW of output, with the plant then expected online in 2028 (Constellation Energy, September 20, 2024). Constellation has since accelerated that timeline: after PJM granted an accelerated interconnection approval, the company said the restart was on track for 2027, a year ahead of the original date (Constellation Energy, “One Year Later: Crane Clean Energy Center Still in the Spotlight and Ahead of Schedule,” September 23, 2025). DOE closed a loan supporting the restart, one of two nuclear-restart loans the agency counts in its first-year tally under the current administration’s nuclear push (DOE, “One Year After Executive Orders: U.S. Nuclear Energy Renaissance Is in Full Swing,” May 23, 2026). In Michigan, Holtec’s Palisades plant, more than 800 MW, is aiming to be the first U.S. reactor to resume operating after entering decommissioning: the NRC approved transfer of the operating license to Palisades Energy LLC on July 24, 2025 and rescinded the certifications that had prohibited operation on August 25, 2025 (NRC, Palisades Nuclear Plant reactor status, accessed September 21, 2026; Utility Dive, August 28, 2025). Workers began loading the plant’s 204 fuel assemblies the week of August 25, 2026, after cracked steam generator tubes pushed the original end-2025 target back; the company has not set a new restart date (Michigan Public, September 1, 2026). In Iowa, NextEra and Google announced on October 27, 2025 a deal to restart the 615 MW Duane Arnold plant, offline since 2020, under a 25-year Google power purchase agreement with first power targeted for the first quarter of 2029 (NextEra Energy, October 27, 2025).

License extensions are compounding the restarts. Thirteen reactors, including Turkey Point, Peach Bottom, Surry, North Anna, Dresden and Browns Ferry, now hold NRC subsequent license renewals extending operating life to 80 years; three more (Nine Mile Point Unit 1, Cooper and Ginna) are under active NRC review, and more than twenty additional plants have filed letters of intent to apply between 2025 and 2034 (NRC, Status of Subsequent License Renewal Applications, accessed September 21, 2026). None of that adds new MW, but it removes retirement dates that ISO and utility planning models used to assume at a 60-year ceiling.

Uprates are moving alongside the restarts. Southern Company’s Hatch, Vogtle and Farley plants are adding a combined 345 MWe across six reactors as part of a $26.5 billion DOE loan package that closed in February 2026, and DOE’s UPRISE initiative targets 2.5 GW of fleet-wide uprate capacity by 2027 and 5 GW by 2029 (DOE, May 23, 2026). Constellation and Walmart signed a long-term power purchase agreement covering the Dresden Clean Energy Center in Illinois for roughly 176 MW, including 30 MW tied to a planned uprate, split across two 15-year terms starting in 2029 and 2030, Walmart’s first nuclear PPA (Constellation Energy, June 23, 2026).

Federal action set the pace for all of it. President Trump signed an executive order on May 23, 2025 directing the NRC to decide new-reactor licenses within 18 months and existing-reactor continued-operation cases within one year, with a stated goal of moving U.S. nuclear capacity from roughly 100 GW today to 400 GW by 2050 (The White House, “Ordering the Reform of the Nuclear Regulatory Commission,” May 23, 2025). In June 2026, DOE announced up to $17.5 billion in conditional loan commitments to accelerate 10 large Westinghouse AP1000 reactors at five sites, an 11 GW combined target the department says could pull deployment timelines forward by as much as three years (DOE, “Department of Energy Announces American Nuclear Supply Chain Loans,” June 23, 2026).

On new build specifically, TerraPower’s Natrium plant at Kemmerer, Wyoming became the first commercial reactor to receive an NRC construction permit in nearly a decade, and the first non-light-water design approved in more than 40 years, when the NRC voted to award the permit on March 4, 2026 (TerraPower, “NRC Approves the Natrium Reactor Construction Permit,” March 4, 2026); construction started April 23, 2026 on the 345 MW sodium-cooled fast reactor (TerraPower, “TerraPower Commences Construction on America’s First Utility-Scale Advanced Nuclear Power Plant,” April 23, 2026). Kairos Power broke ground on its Hermes 2 demonstration reactor in Oak Ridge, Tennessee on April 17, 2026, the first deployment under its October 14, 2024 master plant development agreement with Google to build up to 500 MW of small modular reactor (SMR) capacity by 2035, with first deployment targeted for 2030 (Kairos Power, “Google and Kairos Power Partner to Deploy 500 MW of Clean Electricity Generation,” October 14, 2024; Kairos Power, “Kairos Power Breaks Ground on Hermes 2 Demonstration Plant,” April 17, 2026); Hermes 2 is expected to supply up to 50 MW to the Tennessee Valley Authority grid, and the NRC has extended its construction completion deadline to April 30, 2029. Dow and X-energy’s Long Mott Generating Station at Seadrift, Texas, four Xe-100 units totaling roughly 320 MWe, had its NRC construction permit application submitted on March 31, 2025 and cleared NRC environmental review in 2026, with startup targeted for the early 2030s (Dow, March 31, 2025; ANS Nuclear Newswire, May 20, 2026).

How Likely Is New Nuclear to Actually Get Built?

The three tracks above do not carry the same odds. Restarts are close to certain: each has a defined asset, an existing environmental footprint and a signed offtake agreement, and the main variable is how much equipment work remains, which is exactly what pushed Palisades past its original target. Uprates carry similarly low technical risk because they modify plants already operating, though the aggregate volume DOE is targeting depends on loan disbursement and utility capital budgets across many separate projects. New large reactors and SMRs face a different profile: first-of-a-kind design, supply chain and construction-schedule risk with a poor track record in the United States. The clearest example is still Southern Company’s own Vogtle Units 3 and 4, which entered commercial operation in July 2023 and April 2024 at a final cost of $36 billion and a 15-year schedule, both more than double the project’s original estimates, with Georgia Power’s 45.7% ownership share alone driving a 25% increase in residential rates (PowerMag, “What Was Learned From Building New Nuclear Reactors?,” April 1, 2025). DOE’s current AP1000 loan program finances supply-chain components ahead of construction specifically to reduce that risk for the next wave.

U.S. Nuclear Build Tracks, September 2026 Status

Why Is the Power Market Shifting Back Toward Nuclear?

Four forces are pulling capacity and power markets back toward nuclear, and all four show up in data Noreva already tracks.

First, accreditation. PJM’s marginal ELCC class ratings for the 2027/2028 Base Residual Auction give nuclear a 95% class rating, the highest of any resource type PJM accredits (PJM, ELCC Class Ratings for the 2027/2028 Base Residual Auction, 2025). Gas combined cycle sits at 74% and solar at 7-8% in the same rating set, so a nuclear MW clears an auction at nearly its full nameplate value while a solar MW clears at a fraction of it. For the full ISO-by-ISO accreditation comparison, see Noreva’s UCAP, ELCC and Firm Capacity analysis.

PJM Marginal ELCC Class Ratings, 2027/2028 Delivery Year: Nuclear vs. Two Reference Classes

Second, price. PJM’s capacity market has cleared at the FERC-approved price cap in three consecutive Base Residual Auctions: $329.17/MW-day for the 2026/2027 delivery year (PJM Base Residual Auction results, July 22, 2025), $333.44/MW-day for 2027/2028 (PJM Base Residual Auction results, December 17, 2025), and $325/MW-day for 2028/2029 (PJM Base Residual Auction results, July 14, 2026). Three straight auctions at the cap are a structural signal, and that is the backdrop against which every restart, uprate and new-build decision above is being financed.

Third, demand. ERCOT’s large-load interconnection queue grew nearly 300% from year-end 2024 to year-end 2025, reaching more than 233 GW of requests with data centers accounting for over 70% of that total (Utility Dive, citing ERCOT, January 6, 2026). PJM’s own 2026 Load Forecast Report, posted January 14, 2026, trimmed its near-term summer 2026 peak-load forecast for large loads by 0.7% against the 2025 Long-Term Load Forecast Report after tightening its data-center vetting process, even as its 10-year net energy growth forecast still averages 5.3% a year (PJM, 2026 PJM Load Forecast Report, January 14, 2026). Tighter vetting today is not the same as less demand tomorrow, and Noreva has separately flagged potential unmet demand by 2030 even if every currently queued project completes on schedule (see Noreva’s coverage of how data center builders are solving the power problem), which is why hyperscalers increasingly want firm, carbon-free supply under long-term contracts instead of merchant exposure. Nuclear’s twenty-to-twenty-five-year fixed-term PPA structure, backed by a single large firm resource, fits that demand better than a portfolio of shorter renewable contracts.

Fourth, gas price risk. Noreva’s own natural gas forecasting coverage, launched August 5, 2026, models a high-price scenario in which constrained supply growth, LNG export capacity expanding roughly 160% by 2030, and AI-driven power demand push select U.S. gas hubs above $10/MMBtu. For any counterparty locking in twenty-year power costs, that scenario raises the relative value of a fuel-free, firm resource like nuclear before even accounting for its capacity value.

If It Doesn’t Get Built, When Will Ratepayers Feel the Impact?

If restarts slip further and new build stalls, the counterfactual is gas and delayed retirements. The same $26.5 billion DOE loan package that funds the Hatch, Vogtle and Farley uprates also funds 5.3 GW of new gas generation in Southern Company’s territory (DOE, May 23, 2026), and Noreva has separately documented how “reliability must run” contracts that keep aging coal plants online past their planned retirement dates carry their own cost, with capacity payments in some regions projected to rise sharply through 2050 if coal retirements are not offset by new firm capacity (see Noreva’s analysis of why keeping coal online will be so expensive).

What determines when ratepayers see this is timing more than magnitude. PJM’s recent Base Residual Auctions have cleared one to two years ahead of their delivery years while the auction calendar catches up, with PJM targeting a return to a three-year forward cycle (PJM, July 22, 2025). Any further tightening therefore reaches capacity bills within one to two delivery years, and energy bills as soon as gas prices move. Energy prices carry the faster mechanism of the two, since they track the marginal fuel cost in each hour, increasingly gas: if new nuclear supply is delayed and gas fills the gap, ratepayers are exposed to whatever Noreva’s high-price gas scenario prices in, without even the one-to-two-year lag the capacity auction provides. This article does not estimate retail bills; the auction and forward-curve mechanisms above are the two channels through which a delayed nuclear buildout reaches a ratepayer’s bill, and the size of that effect at the meter is set by state regulation and utility rate design rather than by the wholesale market, so it varies too widely across jurisdictions for a single number to carry meaning.

How Should Developers and Investors Model Nuclear Into a Merchant Curve?

Noreva builds its capacity and power forecasts by ISO and by resource class instead of off one blended curve, and nuclear shows why that matters. A restart or uprate adds firm, high-ELCC MW into an auction that has cleared at the cap three times running, and because nuclear’s 95% PJM ELCC rating means nearly all of its nameplate capacity counts, that MW pulls the marginal clearing price down less than a lower-accreditation resource of the same nameplate size would. A new-build reactor adds no MW for years and carries schedule and cost risk that belongs in a merchant curve as a range, not a point estimate. Noreva’s capacity forecasts and asset valuation work build these scenarios ISO by ISO and resource class by resource class, so a nuclear-adjacent asset, a competing gas peaker, or a storage project sitting behind the same interconnection queue can each be underwritten against the accredited-MW curve nuclear is actually moving.

Frequently Asked Questions: U.S. Nuclear Outlook

What’s new with nuclear in the USA?

Three restarts are underway: the Crane Clean Energy Center in Pennsylvania, Palisades in Michigan, and Duane Arnold in Iowa. Southern Company and Constellation are adding uprates at existing plants, thirteen reactors now hold 80-year NRC license renewals, and construction has started on TerraPower’s Natrium reactor in Wyoming, the first new commercial reactor design approved by the NRC in decades.

What’s the likelihood of new nuclear getting built in the US?

Restarts and uprates are highly likely because they modify existing, previously licensed assets. New large reactors and SMRs carry first-of-a-kind cost and schedule risk, the kind that left Vogtle’s Units 3 and 4 at a final cost of $36 billion, more than double original estimates, and DOE’s current loan structure is designed specifically to reduce that risk for the next wave of AP1000s.

To what extent is the market shifting back toward nuclear development, and why?

Nuclear carries the highest capacity accreditation of any resource type in PJM’s 2027/2028 ELCC class ratings at 95%, PJM’s capacity market has cleared at the FERC-approved cap in three straight auctions, data centers want firm long-term power under contract, and Noreva’s own natural gas forecast flags a scenario where gas prices rise structurally over the next decade. Together those four forces make a firm, twenty-year nuclear PPA look more attractive than it did five years ago.

If new nuclear doesn’t get built, when will ratepayers feel the impact?

Through two channels on two different timelines: PJM’s capacity auction, which has recently cleared one to two years ahead of the delivery year while PJM works back toward its usual three-year-forward cycle, so a shortfall reaches capacity bills within one to two delivery years; and energy prices, which move faster still because they track the marginal fuel, increasingly gas, in each hour.

How does a nuclear restart change a PJM capacity forecast?

A restart adds firm, high-ELCC megawatts into an auction that has already cleared at the FERC-approved cap in three straight Base Residual Auctions, and because nuclear’s PJM ELCC rating is 95%, nearly all of that nameplate capacity counts toward accredited MW. A new-build reactor adds no MW for years and carries first-of-a-kind cost and schedule risk that belongs in a forecast as a range, not a fixed date.

Model Nuclear Restarts and New Build With Noreva

Nuclear’s accreditation value, its capacity timeline, and the gas price risk it hedges against are all moving pieces in the same merchant curve. Noreva’s capacity forecasts and asset valuation platform tracks restarts, uprates and new build ISO by ISO, alongside the PJM, MISO, NYISO, ISO-NE, SPP and CAISO capacity auctions those projects clear into.

Book a demo to see how Noreva prices nuclear into your capacity and power curves.