President Trump signed an executive order setting a goal of at least three new privately built advanced reactors reaching criticality—sustaining a controlled nuclear chain reaction—by July 4, 2026. The goal was met and exceeded. (Natalie Behring/Getty Images)

Four years ago, the operators of Michigan’s Palisades nuclear plant took its reactor offline for what they assumed would be the last time. The plant had generated electricity for more than half a century, but its long-term contract to sell the power was expiring, and its operator, Entergy, was leaving the merchant nuclear business. In May 2022, the company announced that it had permanently shut the plant down. Workers removed the fuel, and a new owner, Holtec International, took over to decommission the facility.

Today, workers are reversing the process. On August 30, Holtec began loading fuel back into the reactor, preparing to restore more than 800 megawatts of generating capacity—enough to power hundreds of thousands of homes. The plant is expected to return to service later this year or early 2027. If successful, it would be the first time a U.S. nuclear plant has been restarted after being retired and decommissioned.

For much of the past decade, the decline of atomic energy seemed inexorable. Plants were closing around the country. Indian Point, which supplied about a quarter of the electricity to New York City and Westchester County, was shuttered in 2021. Three Mile Island in Pennsylvania, the site of America’s worst commercial nuclear accident, closed its last reactor in 2019. California’s largest utility announced plans in 2016 to close Diablo Canyon, the state’s last nuclear plant.

Broad forces drove the decline. Cheap natural gas made nuclear less competitive, while renewable-energy mandates and subsidies drove investment toward wind and solar power. Electricity consumption in America had been essentially flat for nearly two decades. As the nation’s fleet of large reactors aged, it became harder to justify further investment. The decline was apparent enough for James B. Meigs to ask in these pages in 2019: “Have we reached the beginning of the end of the atomic era?”

The answer, it’s now clear, is no. After years of uncertainty, nuclear power is undergoing a genuine revival in the United States. Shuttered plants like Palisades are preparing to restart, while California has reversed course to keep Diablo Canyon running. Developers are working on new fuels, safety systems, and reactor designs, including small modular reactors (SMRs) and microreactors intended for settings beyond the traditional utility power plant.

Investment is powered by renewed electricity demand, driven largely by artificial intelligence and data centers. Microsoft is helping finance the restart of a reactor at Three Mile Island. Amazon has invested in developing small reactors. Google has contracted for nuclear power extending decades into the future. These companies need enormous quantities of electricity to train and run AI models. Nuclear’s ability to produce round-the-clock power on a small footprint and without emissions makes it well suited to meet those needs.

Federal policy has shifted as well, with support spanning both parties and several administrations. Congress passed nuclear-licensing reforms signed by Donald Trump in 2019 and Joe Biden in 2024. Trump’s first administration launched a program to help private companies demonstrate advanced reactors; the Biden administration subsequently provided financing for projects. Trump has since made expanding nuclear power a major energy-policy priority, moving to accelerate reactor testing and overhaul licensing rules at the Nuclear Regulatory Commission.

“The NRC is in the most comprehensive period of reform since the 1970s,” says Chairman Ho K. Nieh. The changes reflect how dramatically nuclear power’s fortunes have shifted—from an energy source that, not long ago, seemed politically, economically, and even technologically unviable to one now poised for a renaissance. That shift is due primarily to nuclear power’s ability to provide what America’s economy increasingly requires: a vast supply of steady, clean electricity delivered reliably to a grid experiencing rapidly growing demand.

Construction on a small modular reactor (SMR) site near Idaho Falls, Idaho.
Construction on a small modular reactor (SMR) site near Idaho Falls, Idaho (Josh Edelson/AFP/Getty Images)

Earlier hopes of a nuclear renaissance did not come to fruition. In the mid-2000s, developers proposed dozens of new reactors, but almost none was built. Only two, both at Plant Vogtle in Georgia, emerged from that wave.

Nieh attributes the failed revival to what he calls “the three Fs”: the financial crisis, fracking, and Fukushima. After the financial crisis, new plants became harder to finance. Developers had to commit enormous sums up front long before earning any return, and persistent delays and cost overruns bedeviled the industry. Meantime, the advent of hydraulic fracturing made natural gas cheap and abundant, which in turn made nuclear a less attractive long-term investment. Then, in 2011, a tsunami disabled backup power and cooling at Japan’s Fukushima plant, causing core meltdowns in three reactors. “Once Fukushima happened, that was a straw that broke the camel’s back,” he says. “All the interest in new builds just evaporated.”

Advances in reactor design and operations have made plants much safer, but by the 2010s, the industry had largely lost its ability to build new ones. The projects that did proceed were costly and slow. Plant Vogtle’s two new reactors were supposed to enter service in 2016 and 2017, for a combined price tag of $14 billion. When they finally came online in 2023 and 2024, it was at more than twice the expected cost. Today, America’s commercial fleet has fewer than 100 reactors, with an average age of about 44 years.

Even during the peak climate-hawk years of the mid-2010s, nuclear remained sidelined. Shutting down nuclear plants became a rallying cry for Green New Deal–era climate activists. States adopted renewable-energy mandates and subsidies that tilted the market toward wind and solar while generally excluding nuclear. Meigs described a “widespread belief that nuclear power simply isn’t necessary.” If wind and solar were about to take over, why keep a reactor open, let alone build more?

In New York, Governor Andrew Cuomo spent years trying to close Indian Point, portraying it as a danger to the New York metropolitan area. When its last reactor was about to shut down in 2021, his office issued a celebratory announcement. The closure, Cuomo said, moved New York “a big step closer to reaching our aggressive clean energy goals.” Yet, within a few years, natural gas had largely replaced the plant’s carbon-free output.

The reality is that wind and solar could not, on their own, replace Indian Point’s 24/7 generation. The grid still required a firm power supply when the wind stalled or the sun went down. By the time Indian Point closed, three new gas-fired plants had entered service in the region. Fossil-fuel generation has since replaced roughly 92 percent of the electricity lost from Indian Point.

Cuomo’s successor is now trying to reverse course. In a July interview, Governor Kathy Hochul called closing Indian Point “one of the biggest mistakes in the state’s history.” She is pursuing an “all-of-the-above” strategy that includes adding 5 gigawatts of new nuclear capacity—more than twice what Indian Point supplied.

The renewed appetite for nuclear power, however, does not by itself make a new plant financially viable. Investors must still account for licensing delays and changing requirements. “Regulatory predictability attracts investment,” says Nieh. “Capital will leave if the risks are too high.”

For advanced-reactor developers, that uncertainty includes how to satisfy requirements written for an earlier generation of nuclear technology. The NRC’s licensing framework was designed for the massive, water-cooled plants of the past. Today’s reactors, however, may use different coolants, fuels, or safety systems and don’t fit that mold. These companies have had to petition the NRC for exemptions from rules written for those older models, a process that can add years and additional layers of uncertainty.

“We’ve had our regulations in place for decades, and they’re not wine. They don’t get better with age,” Nieh says. “We’ve been trying to force-fit new technologies into these old frameworks that aren’t fit for [the] purpose.”

At 12:20 am on July 4, as the nation prepared to mark its 250th anniversary, engineers at Idaho National Laboratory had yet another reason to celebrate. A test reactor built by Aalo Atomics achieved criticality, meaning that it sustained a controlled nuclear chain reaction. The company reached the milestone less than eight months after construction of the reactor began. The experiment used a full-size core operating at essentially zero power to test the nuclear components and controls for Aalo’s planned 10-megawatt reactors.

Aalo’s test was authorized under a new Department of Energy pilot program that President Trump created by executive order in May 2025. The order set what seemed a lofty goal: at least three new privately built advanced reactors reaching criticality by Independence Day 2026. At the time, Kathryn Huff, a nuclear-engineering professor who had led the DOE’s Office of Nuclear Energy under President Biden, told NPR that calling the schedule aggressive was “a pretty big understatement.” Research reactors typically took at least two years to build after construction began, she said.

The goal was not just met but exceeded. Antares Nuclear achieved criticality on June 4, followed by Valar Atomics and Deployable Energy later that month, while Aalo’s July 4 test brought the total to four.

The deadline was part of a series of executive orders Trump signed to accelerate the expansion of nuclear power. One directed an overhaul of NRC regulations and operations to speed up new reactor deployments. Another accelerated testing under the DOE’s authority and directed the department to create the pilot program that set the July 4 deadline. A third sought to “reinvigorate the nuclear industrial base” by strengthening the nation’s domestic fuel cycle. The fourth promoted reactors for national-security uses, including military installations and certain AI infrastructure. The orders also set a goal of quadrupling American nuclear capacity from roughly 100 gigawatts to 400 gigawatts by 2050.

The DOE has since cut its reactor-authorization process from 17 steps to 11 and eliminated more than 900 pages of guidance. It used its authority over research reactors to enable companies such as Aalo to build and test prototypes under DOE supervision without first securing an NRC license for commercial operation. The NRC has finalized licensing rules for advanced reactors. Developers can use this framework to meet safety standards without seeking exemptions from requirements for older, light-water plants. Trump’s order also directs the NRC to set firm deadlines of 18 months for new reactor applications and one year for license renewals.

For Nieh, the reforms involve how the NRC views its mission. Trump’s orders direct the commission to revise its mission statement expressly to include enabling the safe use and deployment of civilian nuclear technologies. Nieh views the changes, along with the bipartisan ADVANCE Act signed by Biden in 2024, as critical for an agency whose decisions affect whether developers can attract investment and move forward with construction. “We’re not here just to move paperwork from one inbox to another,” he says. “We want to make it safe, but we also want to make it possible.”

NASA Administrator Isaacman Tours Idaho National Laboratory, Idaho Falls, Idaho.
NASA executives tour Idaho National Laboratory. Nuclear power is returning because America has rediscovered the value of abundant, dependable electricity, just as building reactors has begun to seem possible again. (Natalie Behring/Getty Images)

With these regulatory changes underway, private capital is flowing into nuclear technology at record levels. According to PitchBook, global venture funding for nuclear startups reached a record $6.2 billion in 2025; by July 2026, investors had committed another $4.5 billion. Most of it is going to American firms.

Much of that money is aimed at small modular reactors, or SMRs: factory-built reactors, generally 300 megawatts or less, that can be shipped in pieces and assembled onsite. These smaller reactors reduce the need for the custom engineering and extensive field labor that make one-off projects like Plant Vogtle so costly and slow. Dozens of small modular and other advanced reactor designs are now under development nationally, though none yet holds a commercial operating license.

Engineers have also been working to address the safety issues that events like Fukushima exposed—in particular, the possibility that a loss of cooling could cause reactor fuel to overheat and melt. Many SMR designs rely on gravity, convection, and natural circulation to cool the reactor without the need for powered pumps or an operator at the switch. Some new large reactors work this way, too. Westinghouse’s AP1000, the design used at Vogtle, can maintain core cooling for 72 hours without AC power or operator action.

Microreactors take the smaller-scale approach even further. These designs typically produce less than 10 megawatts and are small enough to move by truck to power military bases, mines, and communities far from a large electrical grid. Several units can also be grouped together to serve larger customers. Aalo, for example, plans to combine its reactors into pods to power data centers.

Some newer designs combine nuclear generation with energy storage that can supply extra electricity when customers need it. In Kemmerer, Wyoming, TerraPower is developing a molten-salt heat-storage system that will enable the plant to maintain a steady output of 345 megawatts, then tap stored heat to generate up to 500 megawatts during periods of peak demand. The NRC issued a construction permit in March, and the project is expected to be completed by 2031.

Federal support is also helping offset some of the initial risk of new projects. The Department of Energy backed the Palisades restart with a $1.52 billion loan and recently selected the Tennessee Valley Authority and Holtec for cost-shared funding for small-reactor projects in Tennessee and Michigan. In June, the DOE’s Office of Energy Dominance Financing announced a conditional commitment of $17.5 billion in loans for components for ten new AP1000 reactors. The financing will help developers purchase reactor vessels, steam generators, and other long-lead-time components while they complete licensing and site preparation, which could slice up to three years off the project timeline.

The most important new customers for nuclear power may be the companies building AI data centers. Lawrence Berkeley National Laboratory estimates that data centers could account for as much as 12 percent of U.S. electricity consumption by 2030, up from roughly 4 percent in 2023. The companies building them have strong incentives to finance vast new supplies of reliable power.

Restarting retired nuclear plants is one way to do so. Microsoft is backing a 20-year agreement to restart a reactor at Three Mile Island. Google has signed a 25-year agreement supporting the restart of Iowa’s Duane Arnold plant. Tech firms are bankrolling new reactor designs as well. In 2024, Amazon led an investment of roughly $500 million in X-energy, a developer of SMRs. The funding supports reactor licensing, fuel production, and early development of a 320-megawatt project in Washington State. The firms aim to bring more than 5 gigawatts of nuclear power online by 2039 to fuel Amazon’s operations. “The US cannot win the AI race against China without nuclear power,” X-energy chief executive Clay Sell told the Financial Times earlier this year.

None of this guarantees a nuclear renaissance. The United States still operates fewer reactors than it did a decade ago, and no commercial SMR is yet producing electricity. But regulatory changes, along with advances in nuclear technology, have made building reactors faster and cheaper a more credible prospect. Safety is now more of a political obstacle than a technical one. (Nuclear’s safety record, measured by deaths per unit of electricity, is comparable with that of wind and solar.) Meanwhile, new financing models are drawing more private capital into nuclear projects, shifting the risk of construction delays and cost overruns from utility ratepayers to investors. A proposed deal from Brookfield illustrates the approach: to complete the abandoned V.C. Summer reactors in South Carolina, private investors would bear most of the construction risk rather than recover the project’s capital costs through customers’ utility bills.

The case for a nuclear renaissance is stronger than ever. Licensing is getting faster and more predictable, while greater access to capital is making new projects easier to finance. Electricity demand is surging in the age of AI, and companies are willing to sign decades-long contracts to secure reliable power. Nuclear is returning, in short, because America has rediscovered the value of abundant, dependable electricity, just as building reactors has begun to seem possible again.

Donate

City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading