clock menu more-arrow no yes mobile
A hilly landscape with scattered trees and the sun rising in the background. Oversized nuclear power sources are emanating light in different colors amongst the landscape: a small modular reactor, a MARVEL microreactor, a portable reactor, and power plant cooling towers.
Advanced nuclear reactors, especially smaller designs, could help solve many of the persistent problems with nuclear energy.
Paige Vickers for Vox

Smaller, cheaper, safer: The next generation of nuclear power, explained

The nuclear industry’s big bet on going small.

IDAHO FALLS, Idaho — Inside the Transient Reactor Test Facility, a towering, windowless gray block surrounded by barbed wire, researchers are about to embark on a mission to solve one of humanity’s greatest problems with a tiny device.

Next year, they will begin construction on the MARVEL reactor. MARVEL stands for Microreactor Applications Research Validation and EvaLuation. It’s a first-of-a-kind nuclear power generator, cooled with liquid metal and producing 100 kilowatts of energy. By 2024, researchers expect MARVEL will be the zero-emissions engine of the world’s first nuclear microgrid here at Idaho National Laboratory (INL).

“Micro” and “tiny,” of course, are relative. MARVEL stands 15 feet tall, weighs 2,000 pounds, and can fit in the trailer of a semi-truck. But compared to conventional nuclear power plants, which span acres, produce gigawatts of electricity to power whole states, and can take more than a decade to build, it’s minuscule.

For INL, where scientists have tested dozens of reactors over the decades across an area three-quarters the size of Rhode Island, it’s a radical reimagining of the technology. This reactor design could help overcome the biggest obstacles to nuclear energy: safety, efficiency, scale, cost, and competition. MARVEL is an experiment to see how all these pieces could fit together in the real world.

“It’s an applications test reactor where we’re going to try to figure out how we extract heat and energy from a nuclear reactor and apply it — and combine it with wind and solar and other energy sources,” said Yasir Arafat, head of the MARVEL program.

The project, however, comes at a time when nuclear power is getting pulled in wildly different directions.

Germany just shut down its last nuclear reactors. The US just started up its first new reactor in 30 years. France, the country with the largest share of nuclear energy on its grid, saw its nuclear power output decline to the lowest levels since 1988 last year. Around the world, there are currently 60 nuclear reactors under construction, with 22 in China alone.

But the world is hungrier than ever for energy. Overall electricity demand is growing: Global electricity needs will increase nearly 70 percent by 2050 compared to today’s consumption, according to the Energy Information Administration. At the same time, the constraints are getting tighter. Most countries in the world, including the US, have now committed to zeroing out their net impact on the climate by the middle of the century.

To meet this energy demand without worsening climate change, the US Energy Department’s report on advanced nuclear energy released in March said “the U.S. will need ~550–770 [gigawatts] of additional clean, firm capacity to reach net-zero; nuclear power is one of the few proven options that could deliver this at scale.”

Photo of MARVEL project lead Yasir Arafat in front of a prototype of the reactor
MARVEL program director Yasir Arafat, in front of a prototype microreactor, said the project will study how small nuclear reactors could function on a power grid.
Idaho National Laboratory/US Department of Energy

The US government is now renewing its bets on nuclear power to produce a steady stream of electricity without emitting greenhouse gases. The Bipartisan Infrastructure Law included $6 billion to keep existing nuclear power plants running. The Inflation Reduction Act, the US government’s largest investment in countering climate change to date, includes a number of provisions to benefit nuclear power, including tax credits for zero-emissions energy.

“It’s a game changer,” said John Wagner, director of INL.

The tech sector is jumping in, too. In 2021, venture capital firms poured $3.4 billion into nuclear energy startups. They’re also pouring money into even more far-out ideas, like nuclear fusion power. Public opinion has also started moving. An April Gallup poll found that 55 percent of Americans favor and 44 percent oppose using nuclear energy, the highest levels of support in 10 years.

But nuclear energy is still facing some long-running headwinds. It’s the one power source whose operating costs have actually increased over time. Recent construction efforts have run years behind schedule and billions of dollars over budget. Most reactors still rely on enriched uranium, a pricey fuel to mine and process. Finding a place to store nuclear waste remains a problem. The workforce needed to build and operate plants has withered, due to the decades between reactor builds. And now, with rising interest rates, it’s more expensive to finance ambitious energy projects.

Can the nuclear energy industry invent its way out of its toughest problems?

Advocates certainly hope so, and the potential for nuclear energy to meet the challenge of climate change is immense. Many new nuclear power technologies are now in design and testing phases. But one of the most promising strategies for nuclear is to go big by going small.

The new generation of nuclear power, explained

Splitting atoms is the largest source of greenhouse gas-free electricity in the US and the second-largest in the world behind hydropower. Nuclear fission produces 10 percent of the world’s electricity. The US has the largest nuclear reactor fleet in the world, with 92 reactors across 53 power plants in 28 states.

The current crop of nuclear reactors use a variety of different design approaches, tailored to their specific needs. That helped these power plants better fit into the power grids when they were initially built, but it made it harder for them to adapt to changing demands and for newer plants in other places to learn from them.

To understand what sets the new reactor designs apart, it helps to know how earlier designs worked. Generally, civilian nuclear reactors are divided into “generations” that refined the technology, economics, and safety with each iteration.

The first generation of reactors were proofs of concept, according to Jess Gehin, associate director for nuclear science and technology at INL. From there, they scaled up in size and added safety features to make them more usable in the real world, forming the second generation. The bulk of the world’s operating nuclear reactors right now are second-generation designs. They are also the foundation of most business models and the basis for nuclear energy regulations.

More recent third-generation reactors advance this with improved safety features. “Several of those have been built that actually start moving away from the active safety systems to more passive systems,” Gehin said. The recently opened reactor at the Vogtle Electric Generating Plant in Georgia is a design called AP1000. It’s considered a generation three-plus reactor that uses fewer moving parts than conventional designs and can cool off on its own should something go wrong. “You can go 72 hours without any operator interaction,” Gehin said.

Photo of the Transient Reactor Test Facility
The Transient Reactor Test Facility at Idaho National Laboratory will host the MARVEL microreactor.
Umair Irfan/Vox

Fourth-generation reactors are now in the works. Unlike current reactors that mainly use water to control the reaction and to stay cool, these designs use other materials like liquid metal, pressurized gas, and molten salt. The advantage is that they can reach higher operating temperatures, which can lead to greater efficiency. Industrial processes like steel production could also draw on that extra heat.

Many fourth-generation designs can also use cheaper, lower-grade nuclear fuels. That’s one of the approaches being developed by TerraPower, a nuclear company founded by investor Bill Gates. Some fourth-generation designs can even use waste from other reactors. They can also integrate equipment that allows them to ramp up and down more readily to scale with energy demands.

These combined effects improve the economics of nuclear power, streamlining the overall process from reducing fuel costs to generating power more effectively to reducing waste and to improving safety.

Nuclear can do more than generate electricity

Some of the most significant advances in nuclear energy, however, may not be in the reactors themselves. Their biggest benefits could come from rethinking how they fit into the existing power infrastructure.

The Energy Department has suggested that hundreds of sites for coal power plants, which are rapidly shutting down across the country, could be repurposed for nuclear energy. The advantage is that they already have many of the necessary permits and the equipment to plug into the power grid, saving some of the startup costs of a new plant.

Most conventional reactors are optimized to run flat out, with a steady output of energy. But demand on the power grid varies widely as lights switch on in the evening or heaters turn on during the day. While some nuclear power plants can ramp up and down, it’s not always easy. Windy and sunny days can also mean that there’s a surfeit of cheap electrons from renewables and undercut nuclear electricity on price. And since nuclear plants have high fixed costs even when they’re turned down, they prefer to stay up and sell as much of their electricity as possible.

Now, engineers are planning nuclear reactors with this capricious demand in mind. “New reactors are designed to be dispatchable and flexible,” said Christine King, director of the Gateway for Accelerated Innovation in Nuclear at INL.

Photo of scientist operating remote armature
A researcher at Idaho National Laboratory investigates nuclear materials inside a containment chamber.
Umair Irfan/Vox

One idea is to integrate energy storage. Molten salt, for instance, can be used to store heat from a nuclear reactor for hours at a time and dispatch it as needed. Another approach is to use the heat from a reactor not just to boil water but to provide industrial heat to factories. Researchers are also designing reactors that can produce hydrogen when they have excess power, which in turn can run fuel cells in cars or put electrons back on the grid.

Electricity from nuclear power plants doesn’t necessarily have to feed into the power grid either, according to King. It can instead power dedicated processes like capturing carbon dioxide directly from the air. Capturing this carbon dioxide is a highly energy-intensive process, though, and nuclear could provide the requisite power without making the problem worse. That captured carbon could then serve as a building block for synthetic fuels, particularly for sectors that are hard to electrify, like aviation and shipping.

It’s hard to build anything these days

The virtues of advanced nuclear reactors are all great in theory. In practice, building anything big is really, really hard.

Bent Flyvbjerg, a professor at the IT University of Copenhagen and a professor at the University of Oxford, recently co-authored a book called How Big Things Get Done. It examines why so many major infrastructure projects like high-speed trains, IT systems, and even home renovations run behind schedule and over budget. Often, these problems arise from a failure of planning, inadequate expertise, political pressure, and limited experience.

Nuclear energy brings even more unique challenges. One is that the technology itself is evolving, so it’s difficult to learn from past efforts to build reactors. Nuclear regulators also built their rules around second-generation designs. So as engineers come up with new ways to split atoms, nuclear observers also have to come up with new standards to make sure they’re safe. The back-and-forth between developers and regulators adds another layer of complexity to the design process.

And anytime there’s a problem with a nuclear power plant anywhere, regulators step up their scrutiny. “Once they had adapted to a certain set of standards, they would be raised because there was a nuclear incident or accident,” Flyvbjerg said.

Most existing commercial reactors also don’t scale up and down easily, so they have to start with bigger, more expensive designs at the outset. That means they have to recover that cost over decades, but if utilities get their electricity demand forecasts wrong, then nuclear power plants end up having to raise their prices or lose money. With new reactors being built for the first time, there’s little experience to draw on. Builders often encounter unanticipated problems that require more money and resources to fix.

The Vogtle Plant was nearly six years behind schedule, and its cost was almost double its initial budget of $14 billion, for example. Utilities in South Carolina abandoned a $9 billion effort to build two AP1000 reactors in 2017. If you’re an investor or a public utility, it’s enough to grind your molars into dust. Developing fourth-generation reactors stands to be an even more expensive, time-consuming process.

But there are some potential ways to chip away at these monumental challenges. One way is for governments to step in and provide research support to these new designs and test them out.

Photo of the Materials and Fuels Complex and Idaho National Laboratory
Idaho National Laboratory builds and tests nuclear reactors to help the industry figure out how they’ll work in the real world.
Umair Irfan/Vox

For the nuclear industry, the hot new strategy is to scale down with small modular reactors, or SMRs. Rather than building huge, customized reactors on site, companies like NuScale are developing smaller reactors, on the order of 10 to 50 megawatts, that can be built in factories and trucked or shipped around the world. The standardized designs could save costs. And by starting small and scaling up, they could meet a variety of use cases.

This approach has already caught eyes around the world. The US Navy already operates more than 200 small nuclear reactors to power submarines and aircraft carriers. The test is to see whether the business case makes sense on land. China and Russia are already running SMRs, and 19 countries are developing them. Canadian Prime Minister Justin Trudeau said in April that Canada is making “a return to nuclear, which we’re very very serious about, and investing in some of the small modular reactors.” One of NuScale’s first commercial SMR plants in the world is now planned in Romania in 2028.

“This is the right experiment to be doing,” Flyvbjerg said.

And with designs like MARVEL, researchers are investigating even smaller reactors that can power remote communities, back up renewables, or provide emergency power after a disaster. As reactors get smaller, though, the question is how many it will take in order to achieve economies of scale.

“A lot of learning has to do with how many you build,” said Gregory Nemet, a professor at the University of Wisconsin Madison and author of How Solar Energy Became Cheap.

Technologies like wind turbines, photovoltaic panels, and lithium-ion batteries saw huge price drops in part because it was easy to build a lot of them, so small improvements in performance had big ripple effects. If smaller nuclear reactors could achieve even a fraction of these cost declines, they could finally push the cost curve of nuclear power in the other direction.

It’s not clear how much advanced nuclear will cost

Balancing the books may prove to be a bigger obstacle for nuclear power than splitting the atom.

A new report from the National Academy of Engineering says the economics of nuclear power “is perhaps the largest challenge to the commercial success of advanced reactors.” Advanced nuclear reactors are especially tricky to game out.

“Let me just say that anyone making estimates of what it will cost to produce electricity from these power plants has got to have a whole series of embedded assumptions, there’s a lot of uncertainty,” said Richard Meserve, a former chair of the nuclear regulatory commission and a co-author of the report, during a briefing about the report.

Another big issue is that most countries still don’t have a long-term solution for dealing with nuclear waste, which can remain hazardous for hundreds of years. It’s a huge technical and political problem.

And while there is more demand for clean energy, interest rates are rising, making it more expensive to borrow money to build anything, let alone financially risky novel reactors. INL’s Wagner noted that US reactor construction halted in the ’80s due in part to high interest rates at the time. “When interest rates go to 10, 12, 15 percent, what happens? You’ve got cost overruns,” he said.

At the same time, the world is about to overrun its carbon budget and overshoot the goal of limiting warming to less than 2.7 degrees Fahrenheit (1.5 degrees Celsius) this century.

The US has now committed to cutting its greenhouse gas emissions in half by 2030 compared to 2005 levels. It’s unlikely that new nuclear power plants will make much progress toward that goal, now less than seven years away. But the US and more than 130 countries in the world want to eliminate their contributions to climate change entirely by 2050. That goal demands far cleaner, more abundant, and reliable energy than we have now.

Nuclear could help the world achieve this. It’s a risky and expensive investment, but the foundations for this future have to be laid now.


A big El Niño is looming. Here’s what it means for our weather.

Down to Earth

In defense of flies. Yes, really.

Future Perfect

UN numbers say meat is bad for the climate. The reality is worse.

View all stories in Science