Yes, the US is on track to build new nuclear power plants, though the timeline and scale remain uncertain. The last large reactor to come online, Vogtle Unit 4 in Georgia, began commercial operation in March 2024. No traditional large-scale reactors are currently under construction anywhere in the country. But a combination of federal policy, surging electricity demand from AI data centers, and new reactor designs is setting the stage for a nuclear expansion not seen in decades.
The Department of Energy has set a target of 200 gigawatts of new nuclear capacity by 2050, which would triple the country’s current nuclear output. Getting there will require overcoming the cost overruns, regulatory delays, and fuel supply gaps that have stalled nuclear construction for years.
Where Things Stand Today
The US currently has no nuclear reactors under construction. Vogtle Units 3 and 4 in Georgia, which came online in 2023 and 2024 respectively, were the first new reactors built in the country in over 30 years. Each generates about 1,114 megawatts, enough to power roughly 500,000 homes. But the Vogtle project became a cautionary tale: it ran billions of dollars over budget and years behind schedule, bankrupting one of its original builders.
That history has made utilities wary of committing to traditional large reactor projects. The next wave of construction is instead focused on smaller, more modular designs that their developers say can be built faster, cheaper, and with more flexibility.
The First New Reactors Taking Shape
TerraPower, the nuclear company backed by Bill Gates, received a construction permit from the Nuclear Regulatory Commission for its Natrium reactor in Kemmerer, Wyoming. The company plans to begin construction in the coming weeks, with a projected completion date of 2030. The Natrium design uses liquid sodium as a coolant instead of water, which allows it to operate at higher temperatures and store energy for use during peak demand.
Not every new design has made it this far. NuScale Power, the only company to have its small modular reactor design fully certified by the NRC, canceled its flagship project in Idaho after costs spiraled. The plant was originally supposed to deliver electricity at $55 per megawatt-hour, but estimates climbed to $89 per MWh. Utilities that had signed on to buy the power began withdrawing, and without enough customers, the project collapsed. The Department of Energy had committed $1.4 billion to support it.
That cancellation highlighted a core challenge: even with government backing, first-of-a-kind nuclear projects carry enormous financial risk. Costs tend to drop with subsequent builds, but someone has to pay for the first one.
Tech Companies Are Driving Demand
The single biggest shift in the nuclear landscape is the entrance of major technology companies. AI data centers consume enormous amounts of electricity around the clock, and companies like Amazon, Microsoft, and Google need carbon-free power sources that can run 24/7, something solar and wind alone can’t provide.
Amazon has signed three separate nuclear energy agreements. In Washington state, the company is partnering with Energy Northwest to develop four advanced small modular reactors using X-energy’s technology. The first phase would generate about 320 megawatts, with the option to scale up to 960 megawatts. Amazon’s investment also includes manufacturing capacity to support more than five gigawatts of new nuclear projects using the same technology. In Virginia, Amazon is working with Dominion Energy to explore building a small modular reactor near an existing nuclear station, targeting at least 300 megawatts of new capacity in a region where power demand is projected to grow 85% over the next 15 years.
Microsoft signed a deal to purchase power from a restarted Three Mile Island reactor in Pennsylvania. Google has announced its own nuclear purchase agreements. These commitments matter because they give reactor developers something that has been difficult to secure: guaranteed customers willing to pay for the power before a plant is even built.
Restarting Shuttered Plants
Building new reactors takes years. Restarting ones that already exist is faster. Constellation Energy is working to bring the Three Mile Island Unit 1 reactor back online, now renamed the Christopher M. Crane Clean Energy Center. The plant operated for more than 40 years before shutting down in 2019 for economic reasons, not safety concerns. (It’s a separate unit from the one involved in the famous 1979 accident.)
Constellation has submitted licensing requests to the NRC to resume power operations, and the agency began restart inspection activities in early 2026. The Palisades plant in Michigan is also being considered for a restart, which would make it the first US nuclear plant to return to service after entering the decommissioning process.
These restarts are appealing because the expensive infrastructure already exists. The reactor vessels, turbines, cooling systems, and grid connections are in place. The main costs involve refurbishing equipment, replacing fuel, and satisfying current safety requirements.
Federal Policy Is More Supportive Than Ever
The policy environment for nuclear energy has shifted significantly. The Inflation Reduction Act created a production tax credit for existing nuclear plants worth up to 1.5 cents per kilowatt-hour for facilities that meet prevailing wage requirements. This credit, available through 2032, helps keep aging plants economically competitive and prevents more premature shutdowns.
Congress also passed the ADVANCE Act in 2024, which directs the NRC to streamline its licensing process for new reactor designs. The law establishes expedited review procedures for qualifying applications and specifically addresses licensing reactors at former coal plant sites and other industrial locations, where grid connections and community infrastructure already exist.
These two laws work in tandem. The tax credits reduce the financial risk for investors, while the licensing reforms are meant to shorten the decade-plus timeline that has historically made nuclear projects so expensive. Whether the NRC can actually speed up its reviews without cutting corners on safety remains an open question.
The Fuel Supply Problem
Many advanced reactor designs require a specialized fuel called high-assay low-enriched uranium, or HALEU. This fuel is enriched to a higher concentration than what conventional reactors use, but well below weapons-grade levels. The problem: HALEU is not currently available from domestic suppliers. Russia has been the primary global source, and geopolitical tensions have made that supply chain unreliable.
The Department of Energy is running a HALEU Availability Program to build domestic enrichment and processing capacity, with multiple contracts underway for enrichment services, fuel transportation, and manufacturing. These programs have target dates in 2025 and 2026, but scaling up to commercial volumes will take longer. If the fuel supply isn’t ready when reactors are, deployment timelines will slip regardless of how fast the plants themselves are built.
Public Support Is Growing
Public opinion has moved in nuclear energy’s favor. About 59% of US adults now support building more nuclear power plants, according to a Pew Research Center survey from 2025. That’s a notable increase from just a few years ago. Support is higher among Republicans (69%) than Democrats (52%), but both groups have become more favorable since 2020. Americans still prefer expanding solar (77%) and wind (68%) over nuclear, but the gap is narrowing.
This shift matters politically. Nuclear projects require local community acceptance, state-level permitting, and often legislative support for rate increases that utilities pass on to customers. Broader public backing makes all of those steps easier.
How Many Plants Could Actually Get Built
The Department of Energy’s target of 200 gigawatts of new nuclear capacity by 2050 is ambitious by any measure. The entire US nuclear fleet today generates about 100 gigawatts. Tripling that in 25 years would require building reactors at a pace the country hasn’t achieved since the 1970s and 1980s.
The realistic near-term picture is more modest. TerraPower’s Natrium plant, if completed on schedule in 2030, would be a single demonstration project. Amazon’s SMR partnerships could add roughly a gigawatt over the next decade. A handful of plant restarts might contribute another two to three gigawatts. Together, these projects could prove that new nuclear construction is financially viable in the US, which would be the necessary first step toward anything approaching the DOE’s 2050 vision.
The biggest variable is cost. If the first wave of small modular reactors can be built close to their projected budgets, it will unlock financing for dozens more. If they follow the pattern of Vogtle and NuScale, with ballooning costs and retreating investors, the expansion will stall. The next five years will largely determine which path the country takes.