President Joe Biden has set ambitious goals for fighting climate change: To cut U.S. carbon emissions in half by 2030 and to have a net-zero carbon economy by 2050. The plan requires electricity generation—the easiest economic sector to green, analysts say—to be carbon-free by 2035.
Where is all that clean electricity going to come from?
A few figures from the U.S. Energy Information Administration (EIA) illustrate the challenge. In 2020 the United States generated about four trillion kilowatt-hours of electricity. Some 60 percent of that came from burning fossil fuels, mostly natural gas, in some 10,000 generators, large and small, around the country. All of that electricity will need to be replaced—and more, because demand for electricity is expected to rise, especially if we power more cars with it.
Renewable energy sources like solar and wind have grown faster than expected; together with hydroelectric, they surpassed coal for the first time ever in 2019 and now produce 20 percent of U.S. electricity. In February the EIA projected that renewables were on track to produce more than 40 percent by 2050—remarkable growth, perhaps, but still well short of what’s needed to decarbonize the grid by 2035 and forestall the climate crisis.
This daunting challenge has recently led some environmentalists to reconsider an alternative they had long been wary of: nuclear power.
Nuclear power has a lot going for it. Its carbon footprint is equivalent to wind, less than solar, and orders of magnitude less than coal. Nuclear power plants take up far less space on the landscape than solar or wind farms, and they produce power even at night or on calm days. In 2020 they generated as much electricity in the U.S. as renewables did, a fifth of the total.
But debates rage over whether nuclear should be a big part of the climate solution in the U.S. The majority of American nuclear plants today are approaching the end of their design life, and only one has been built in the last 20 years. Nuclear proponents are now banking on next-generation designs, like small, modular versions of conventional light-water reactors, or advanced reactors designed to be safer, cheaper, and more flexible.
“We’ve innovated so little in the past half-century, there’s a lot of ground to gain,” says Ashley Finan, the director of the National Reactor Innovation Center at the Idaho National Laboratory.
Yet an expansion of nuclear power faces some serious hurdles, and the perennial concerns about safety and long-lived radioactive waste may not be the biggest: Critics also say nuclear reactors are simply too expensive and take too long to build to be of much help with the climate crisis.
Bombs into plowshares
A test reactor at the Idaho National Laboratory, where Finan now works, produced the first electrical power from nuclear energy in 1951. Its success was soon trumpeted in President Dwight Eisenhower’s famous “atoms for peace” speech to the United Nations in 1953. Arjun Makhijani, a nuclear physicist who runs the non-profit Institute for Energy and Environmental Research, points out that the speech was given shortly after a thermonuclear test blast in the Soviet Union, when atomic fears were at a peak.
“Basically, he said this is too doom and gloom—give me something good to say,” Makhijani explains. Eisenhower’s speech ushered in a new nuclear era: Global interest in nuclear power spiked, and countries around the world began building large reactors, often with technology and expertise from the United States.
By 1996, nuclear power provided 17.6 percent of the world’s electricity. Today, that’s down to about 10 percent. The Fukushima accident in 2011 is a major reason for the decline. Japan’s 48 nuclear reactors have largely stayed offline since then; Germany has closed 11 of its 17 reactors and intends to close the remaining six by 2022. Belgium, Spain, and Switzerland are also phasing out their nuclear programs.
The United States, still the world’s largest producer by far of nuclear electricity, currently has 94 reactors in 28 states. But after the Three Mile Island accident in 1979, when a reactor partially melted down near Middletown, Pennsylvania, enthusiasm for nuclear energy dimmed.
The average age of American power plants, which are licensed to run for 40 years, is 39; in the last decade, at least five have been retired early, largely because maintenance costs and cheaper sources of power made them too expensive to operate.
The most recent closure came just last week, on April 30, when the second of two reactors was shut down at the Indian Point power plant, on the Hudson River north of New York City. Until a few years ago, those reactors had supplied a quarter of the city’s power. Nationwide, the EIA predicts that nuclear power generation will decline 17 percent between 2018 and 2025.
Late and over budget
While environmental opposition may have been the primary force hindering nuclear development in the 1980s and 90s, now the biggest challenge may be costs. Few nuclear plants have been built in the U.S. recently because they are very expensive to build here, which makes the price of their energy high.
Jacopo Buongiorno, a professor of nuclear science and engineering at MIT, led a group of scientists who recently completed a two-year study examining the future of nuclear energy in the U.S. and western Europe. They found that “without cost reductions, nuclear energy will not play a significant role” in decarbonizing the power sector.
“In the West, the nuclear industry has substantially lost its ability to build large plants,” Buongiorno says, pointing to Southern Company’s effort to add two new reactors to Plant Vogtle in Waynesboro, Georgia. They have been under construction since 2013, are now billions of dollars over budget—the cost has more than doubled—and years behind schedule. In France, ranked second after the U.S. in nuclear generation, a new reactor in Flamanville is a decade late and more than three times over budget.
“We have clearly lost the know-how to build traditional gigawatt-scale nuclear power plants,” Buongiorno says. Because no new plants were built in the U.S. for decades, he and his colleagues found, the teams working on a project like Vogtle haven’t had the learning experiences needed to do the job efficiently. That leads to construction delays that drive up costs.
Elsewhere, reactors are still being built at lower cost, “largely in places where they build projects on budget, and on schedule,” Finan explains. China and South Korea are the leaders. (To be fair, several of China’s recent large-scale reactors have also had cost overruns and delays.)
“The cost of nuclear power in Asia has been a quarter, or less, of new builds in the West,” Finan says. Much lower labor costs are one reason, according to both Finan and the MIT report, but better project management is another.
The MIT study suggests that standardizing reactor designs and building the same reactor many times is a key to reducing costs. One way to do that may be with small modular reactors (SMRs), which are generally considered to be less than 300 megawatts, compared to the 1,000 megawatts of a traditional nuclear power plant. Their smaller size, Buongiorno says, may allow these reactors’ components to be built in factories, allowing for economies of production, and reducing construction times and uncertainties. The small reactors could either be used individually or combined to make a single large power plant.
In the U.S., a company called NuScale has recently received design certification approval from the Nuclear Regulatory Commission for its SMR, the first and only company to do so. Its reactor is a miniaturized version of a traditional reactor, in which pressurized water cools the core where nuclear fission is taking place. But in the NuScale design, the whole reactor is itself immersed in a pool of water designed to protect it from accidental meltdown.
NuScale hopes to build 12 of these reactors to produce 720 megawatts at the Idaho National Laboratory as a pilot project. It’s been supported by the U.S. Department of Energy, which has approved up to $1.4 billion to help demonstrate the technology. NuScale plans to sell the plant to an energy consortium called Utah Associated Municipal Power Systems.
Last year, eight of the 36 utilities in the consortium backed out of the project, citing the cost. The company recently announced the project would be delayed to 2030, and the cost would rise from $4.2 billion to $6.1 billion.
Nuclear opponents point to this latest disappointment as yet another example of why nuclear isn’t up to the task.
“If your first SMR isn’t built until the late 2020s, and then you have to turn it on, not to mention set up a whole new global supply chain, are you going to reach zero emissions by 2035?” asks IEER’s Makhijani. “You can’t make a significant contribution in time.” He adds that the industry’s long history of overruns and delays are especially problematic when considering climate commitments. “There’s no room for significant mistakes.”
A variable and uncertain grid
On an electric grid, supply has to precisely match the constantly fluctuating demand; at the moment, there are no large storage reservoirs for electrons, like the ones we have for water. Renewables make this balancing act harder, because they produce a fluctuating supply of electricity—when it’s cloudy, or the wind isn’t blowing, the grid needs more energy from other sources.
The future of nuclear power will depend in part on how well it can balance a grid that increasingly relies on renewables. Nuclear has traditionally been considered a baseload source of energy—the reactors run as often as possible to spread their enormous fixed costs over the largest number of kilowatt-hours. Unlike gas turbines, which can be turned on and off in seconds to “follow the load,” reactors take an hour or more to cut their production in half.
It’s not that reactors can’t follow the load; they’re just slower. “They can and do, because they have to,” Buongiorno says. “It’s just never an attractive economic proposition.”
Last fall, the DOE awarded $80 million each to two companies working on advanced reactor designs intended in part to address this problem. The first, TerraPower, a startup founded by Bill Gates, is working on a sodium-cooled reactor that, instead of using its heat directly to drive a turbine and generate electricity, stores the heat in a tank of molten salt, where it can be tapped to generate electricity when needed.
The second grant went to a company called X-energy for a gas-cooled reactor that operates at very high temperatures, producing steam that would be suitable for industrial processes as well as generating electricity. That kind of “load-switching,” Finan and Buongiorno both say, can help nuclear reactors manage variable demand for electricity—while at the same time helping to decarbonize industry. Small reactors might even be sited right next to a factory that requires both heat and electricity. The high-level radioactive waste they produce, however, would need to be transported to a centralized location for management.
But while promising, none of these new designs are moving quickly enough to meet Biden’s targets. DOE officials called their decision to support these two pilot projects, which aim to be fully operational by 2028, “their boldest move yet.”
Meanwhile, there’s a more direct way to balance the variability of renewables: store electricity in batteries. The market for utility-scale battery storage is exploding; it increased by 214 percent in 2020, and the EIA predicts that battery capacity will surge from its current 1,600 megawatts to 10,700 by 2023.
Makhijani thinks nuclear power isn’t going to be needed to balance the grid. A study he conducted in 2016 for the state of Maryland found that increased battery storage, combined with incentives to consumers to reduce their electricity use at peak times, would almost allow utilities to balance the variability of renewables.
They’d just need to store a little energy as hydrogen, which can be produced by running renewable electricity through water and then converted back to electricity in a fuel cell. That process is currently very expensive, Makhijani says, but “as long as it’s not giant amounts, it’s affordable.”
A window of opportunity
Worldwide, nuclear power could be a significant player in the coming decades. China, the largest greenhouse gas emitter, increased its nuclear output 6 percent in 2020 and currently has 17 new reactors under construction, according to the World Nuclear Association, a trade group. India is building six. The U.S. is unlikely to match that anytime soon.
Experts differ sharply on the need to build new nuclear power plants in the U.S. Some models suggest that it would be possible with the right policy incentives to meet Biden’s 2035 target for decarbonizing the grid by building out only renewables.
Existing nuclear plants are another story. The benefit of keeping them online for now is more widely accepted—although Makhijani, for one, argues that their carbon-free energy could be replaced more cheaply by investing in new wind and solar.
Because they’re already built, these reactors are essentially sunk costs, and since most have been online for decades, they’ve already depreciated. Still, in many places their energy has to compete on the market, which some may fail to do. That was one factor behind the decision to shut down Indian Point, the plant’s owner, Entergy Corporation, has acknowledged.
The status of existing plants has big implications: Including Indian Point, seven gigawatts of nuclear power are in danger of going offline before 2026 due to depressed electricity prices.
“Taking out nuclear power plants completely destroys gains with renewables,” Buongiorno says. When the San Onofre Nuclear Generating Station, which produced about 8 percent of California’s electricity, closed in 2013, the local cost of electricity increased, and carbon dioxide emissions in California increased by 9.2 million tons the following year.
The MIT report found that in the next decade, the most cost-efficient, reliable grid comes from an energy mix. “Our analysis shows a big share of nuclear, a big share of renewables, and some storage is the best mix that is low-carbon, reliable, and at the lowest cost,” Buongiorno says.
Co-author Michael Corradini, the former director of the Wisconsin Energy Institute, says federal policies that reward the most cost-effective, low-carbon energy—regardless of the technology—make the most sense. Taxing carbon is one example of a technology-neutral energy policy; a renewable energy standard, of the kind Biden proposed in his infrastructure package, might be another. “If you tax carbon, people are going to switch fuels to things that are more economical,” Corradini says.
At the end of the day, “we need an all-of-the-above policy.”
Lois Parshley is a freelance journalist. Follow her on Twitter @loisparshley.