Nuclear Plant Restart Race: Which of 3 Revivals Is Closest to Power

A nuclear plant that was shut down for being too expensive to run is now being loaded with fuel again. At Palisades in Michigan, that process stalled in recent weeks because a single fuel assembly tilted out of vertical. The pause raises a simple question about any nuclear plant restart: can a reactor that has started decommissioning really come back?

Three shuttered plants, Palisades, Crane and Duane Arnold, are being revived with up to $4.4 billion in Energy Department financing and long-term power contracts from big technology companies. Together they would add roughly 2.25 GW of capacity.

That is enough to serve about 2 million homes and businesses. It also creates a real question for your portfolio: is this a durable investment signal, or an expensive experiment that could slip at every stage?

Here you get a clear read on which plant is furthest along, where the real risks sit, and what the push means for your exposure to uranium and power demand.

Three plants, three very different restart paths

Read the announcements together and the three projects look like one coordinated revival. Put them side by side, ordered by how close each is to producing power, and they come apart.

Palisades: closest, but already delayed

Holtec began loading the reactor’s 204 fuel assemblies on 30 August 2026, one of the last major steps before restart. Loading then paused when one assembly tilted after the handling machine released it. On 13 October 2026, the Nuclear Regulatory Commission (NRC) approved a one-time licence amendment that allows special equipment to move the tilted assembly to the spent-fuel pool.

NRC finding The regulator said the fuel rods remained intact and safe.

The earlier hope of a 2026 return is now obsolete. Power purchase agreements require service by March 2027, which leaves little room for further surprises. Total restoration costs are reported at roughly $2 billion once state support is included.

For readers wanting the technical detail, our full explainer on the Palisades tilted fuel assembly covers the NRC’s preliminary no-significant-hazards finding and what the unpublished root-cause evaluation could reveal.

Crane: the 2027 target and the grid-rights shortcut

Formerly Three Mile Island Unit 1, Crane carries a $1.6 billion restart budget and a $1 billion DOE loan that closed in November 2025. Earlier descriptions pointed to 2028, but CEO Joseph Dominguez has repeated a 2027 target this year, and that is the company’s most recent position. Regulators issued a finding of no significant environmental impact in September 2026, and Constellation wants to operate through 19 April 2034.

Constellation has also asked the Federal Energy Regulatory Commission (FERC) to transfer capacity rights from its Eddystone station to Crane. Keeping an existing grid position avoids a multi-year queue to reconnect.

Duane Arnold: the long runway

Iowa’s only nuclear plant shut in August 2020 after a derecho, a severe line of windstorms, damaged it weeks before its planned closure. NextEra will commit at least $1.6 billion, backed by DOE financing of up to $1.9 billion that closed on 8 September 2026. NRC decisions are expected around January 2028.

The spread from March 2027 to early 2029 tells you that only Palisades is close enough for near-term revenue to matter. The other two remain regulatory and execution stories.

Why a plant closed for being too costly is now worth reviving

To see why the economics flipped, start with why these plants closed. Entergy shut Palisades in 2022 when it left the wholesale power market. Exelon closed Crane in 2019, citing losses and Pennsylvania lawmakers’ refusal to subsidise it, while replacing Duane Arnold with cheaper wind was projected to save Iowa customers $300 million.

A restart differs from a new build because most of the expensive parts already exist. The site, turbines, grid connection and operating history are in place, so the work centres on inspection, repair, upgrades and relicensing. That is why sponsors and the DOE describe restarts as faster and cheaper than building new large reactors.

The idle-car comparison Todd Allen of the University of Michigan compares a restart to reviving a car that has sat unused, but on a far larger scale: rehiring and retraining staff, checking equipment, securing fuel and having regulators reinstate the licence. He considers it a lot to track but fairly straightforward.

The bigger change is who buys the power. A power purchase agreement (PPA) is a long-term contract in which a buyer agrees to purchase electricity from a specific plant. Microsoft signed a 20-year deal for Crane in September 2024, Google has a 25-year agreement at Duane Arnold, and Palisades has PPAs with Wolverine Power Cooperative and Hoosier Energy.

Most terms are undisclosed. One analyst estimate put Crane’s price at about $110-$115 per megawatt-hour, but that is an estimate, not a published term. NextEra CEO John Ketchum has said the Google deal means Iowa residential customers bear none of Duane Arnold’s costs.

Six conditions now make restarts viable:

1. Revenue certainty: Committed 20-25 year buyers replace reliance on wholesale prices.
2. 24/7 demand: Data centres need continuous power, which favours nuclear over intermittent sources.
3. Avoided new-build costs: Existing infrastructure removes the largest capital items.
4. Preserved grid value: Existing interconnection rights skip lengthy queues.
5. Federal financing: DOE loans lower capital costs and execution risk.
6. Policy value: Corporate buyers increasingly want firm carbon-free power.

Demand is the backdrop. In its high-demand case, the Energy Information Administration (EIA) projects data-centre server electricity use at more than 16 times 2020 levels by 2050.

The revenue case now rests on a handful of tech counterparties, so you should treat those contracts as the foundation of the whole thesis. When the next restart is announced, look for three things: a committed buyer, preserved grid rights and a financing backstop.

Who carries the risk: taxpayers, ratepayers or investors?

Sponsors present a confident case: dependable, emissions-free power, delivered faster than any new plant, on sites that already work. The experts are far less united.

The expert split

Lyman’s concern is specific. Most U.S. plants were built mainly between 1970 and 1990 and designed for the natural hazards understood then, while climate change is making extreme weather worse. On the technical side, Utility Dive found no obvious major barrier at Duane Arnold, which operated safely and closed for economic reasons.

The financing deserves a closer look. The DOE amounts, up to $1.5 billion, $1 billion and up to $1.9 billion, total about $4.4 billion, but they are financing ceilings, not cost guarantees. Interest rates, maturities, fees and collateral have not been published for any of the three loans.

The risk checklist

• Equipment: Decades-old systems need inspection, testing and possible replacement.
• Safety basis: Original designs may not reflect current extreme-weather risk.
• Execution: The Palisades tilted assembly shows surprises can stall progress.
• Cost: Inspections may uncover corrosion or required modifications.
• Ratepayers: Full rate-allocation terms are not public for any project.
• Regulatory timing: Crane depends on NRC and FERC approvals; Duane Arnold’s licensing runs into 2028.

Loans are meant to be repaid, but the government absorbs credit and completion risk. You should read federal backing as de-risking for sponsors, not a guarantee for anyone else. Schedule and regulatory risk can be tracked; loan terms and rate allocation remain opaque.

The restart loans are not the only federal balance sheet exposure in the sector, since DOE nuclear supply chain loans for large reactor components show the same pattern of government absorbing credit and completion risk.

What the restart push signals for uranium, nuclear stocks and data-centre power

The commodity numbers are the clearest part of the picture.

Restarts add fuel demand just as the global uranium supply gap widens, with mining analysts projecting that reactor requirements could outpace production well before 2040.

Supply shortfall Reuters reported that AI-driven data-centre demand and new reactor construction are deepening a uranium supply shortfall, especially in the U.S., where mine output has fallen to historic lows.

• Long-term contract price: about $97/lb (September 2026)
• Spot price: a 19-year high in October, alongside a raised long-term forecast of $120/lb
• Cameco, in one 2026 forecast: revenue guidance of $3.13-$3.37 billion, deliveries of 29-32 million pounds and a realised price of $85-$89/lb

Equities are messier. Recent coverage described a broader nuclear sell-off, with money managers turning more defensive on generators such as Constellation and Vistra. No reliable event-specific moves exist for Constellation, NextEra, Vistra or Oklo, and Oklo sits firmly in the speculative AI-and-nuclear theme rather than the restart story.

Hyperscalers need firm power in the late 2020s, before new large reactors or transmission lines can realistically arrive. Restarts fill part of that gap, but 2.25 GW is small against projected demand. You should read them as a signal of scarcity in firm clean power, not a supply answer.

Globally, Palisades is expected to become the first plant to resume power after starting decommissioning. Elsewhere, restarts and refurbishments offer general lessons:

• Canada: Refurbishments demand extensive component replacement and major capital.
• Armenia: Life extension requires regulator approval and safety upgrades.
• Belgium: Safety reviews and politics are intertwined.
• Japan: Post-Fukushima restarts took years of review and local consent.
• India: Construction, supply-chain and regulatory constraints delay capacity.
• South Korea: Policy reversals still depend on licensing and execution.

The common thread is that these projects are site-specific, expensive and sensitive to schedule. “Restarts mean buy nuclear” skips most of the chain.

What the first restarts will prove, and what they will not

Palisades is the nearest test of whether a plant that has entered decommissioning can truly return, and its outcome will set expectations for Crane and Duane Arnold. Three variables matter most:

1. NRC clearance to resume fuel loading at Palisades, and whether the March 2027 contract deadline holds.
2. Crane’s FERC and NRC approvals for a 2027 return.
3. Any disclosure of loan and rate-allocation terms.

A successful restart would strengthen the firm-clean-power thesis. Timing slippage and undisclosed terms, though, mean any position sizing should reflect real uncertainty rather than headline enthusiasm.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results. Financial projections and forward-looking statements are speculative and subject to market conditions, regulatory decisions and company performance.