Virginia is the data center capital of the world, home to the globe’s largest concentration of the energy– and water-devouring gluttons. Nearly a quarter of the electricity used in the state in 2023 was gobbled up by the increasing number of data centers serving Google, Amazon, Microsoft, and other tech giants.

That growing power demand can reach as high as 1 gigawatt per data center for large campuses, enough to power up to 1 million homes. A report released by the U.S. Department of Energy late last year estimated that electricity needed for data centers in the U.S. tripled over the past decade and is projected to double or triple again by 2028. By then, the increased computing power needed for artificial intelligence will equal the electricity demand of 22% of U.S. households, according to an analysis published in May by MIT Technology Review.

Part of the plan to feed that hunger for more energy is a campaign by state and local officials led by Gov. Glenn Youngkin to make the commonwealth a pioneer in emerging smaller-scale nuclear technologies, including small modular reactors and fusion.

“The governor laid out an all-of-the-above energy plan for us to rise and meet that need to ensure that we continue the economic development growth that we’ve been experiencing,” says Glenn Davis, director of the Virginia Department of Energy. “It truly is an all-of-the-above [plan], but with a significant focus on … clean baseload power in the form of small modular reactors,” also known as SMRs.

The state campaign dovetails with a national movement to revitalize and grow nuclear energy as a leading 21st-century power source, especially via microscale reactors. Proponents envision a not-too-distant future in which nuclear power is ubiquitous, perhaps even available on demand via portable reactors small enough to fit inside a tractor-trailer truck.

In May, President Donald Trump signed an executive order to speed the investment and deployment of advanced nuclear reactors. And practically every Big Tech company is forging nuclear power deals with big utilities: Amazon.com has signed multiple agreements to explore SMR development, including with Dominion Energy. This year, Meta signed a 20-year deal with Constellation Energy to help revive an Illinois nuclear power plant. That followed Microsoft’s September 2024 deal with Constellation to revive a portion of the Three Mile Island power plant, the Pennsylvania facility that in 1979 experienced a partial nuclear meltdown, the worst nuclear disaster in U.S. history.

The argument for nuclear power is that it can produce the huge amount of electricity demanded by data centers, while renewable energy sources like wind and solar cannot. A December 2024 study by Virginia’s Joint Legislative Audit and Review Commission found that to meet the demand needs of the data center industry, solar farms would have to be added at twice the annual rate they were added in 2024, and the amount of new wind generation needed would exceed the potential capabilities of all Dominion’s Energy offshore wind sites, both planned and under construction. Large natural gas plants would also need to be added at an equal or faster rate than their busiest build period.

Any mention of nuclear, though, evokes thoughts of the disasters at Three Mile Island and in Russia and Japan, but advocates argue that was your grandparents’ technology. Today’s tech is smaller, cheaper and safer, they say, noting that nuclear power already accounts for 30% of Virginia’s electricity. There’s an opportunity for Virginia to be at the forefront of filling supply chain needs by home growing the companies producing the energy.

“Virginia is leading the nation with … small modular reactor development,” Davis says, “and it’s going to provide huge economic and job opportunities for the next generation of Virginians.”

Risky business?

But being first with emerging technologies like small modular reactors or nuclear fusion is risky. Except for two nuclear reactors that came online in Georgia recently after years of delays at a cost of $30 billion, twice the original estimate, no new nuclear facilities have been built in the United States for three decades.

The Chinese, the Koreans, and the Russians have been building reactors continuously for five decades. The United States has not, making the supply chain and labor expertise lacking here. Fixing that gap will require investment spanning years.

Only two small modular reactors of the kind championed by Youngkin are operational globally, one a floating nuclear power plant in Russia, and the other in China. But more than 80 SMR designs in 19 countries are under development. Plans to build an SMR for a Utah energy provider were abandoned after years of delays, as the cost ballooned from an initial estimate of $3 billion to more than $9 billion. The utility that had agreed to purchase the power backed out after the cost per megawatt-hour increased by more than 53%, from $58 to $89.

John Parsons, deputy director for research at the Center for Energy and Environmental Policy Research at the Massachusetts Institute of Technology, says proponents are not acknowledging the challenges. Only two SMR designs have been approved by the Nuclear Regulatory Commission. Because the reactors are smaller, they are less costly to build, but supply power at a higher price because they don’t benefit from economies of scale. And Parsons says the SMRs will not be built in time to meet near-term data center demands.

“I personally welcome the renewed interest in nuclear, and I think it can serve the people of Virginia, as well as my region here in New England and elsewhere,” he says. “I’m a little concerned about a mismatch between the time horizon on which some of these are being advertised, although I think that will take care of itself. My bigger concern is more at the national level that there are serious concerns about cost. I believe those concerns are more of a challenge than they are of an immovable obstacle. But they are a challenge. They need to be addressed, and leaving it up to each individual utility or state to handle it is probably not as effective a solution as if the national government treated it … [as] the challenge it is.”

Parsons calls the push for small modular reactors fine as a test run, but not an efficient solution to energy needs. “With small modular reactors, the problem is, in the public conversation, this is just a branding exercise,” he says. “People who advertise small modular reactors are just doing it as a way to push aside what they think of as the old, costly, bad nuclear.

“If we took this challenge seriously, we should be building several of the large reactors, doing the work necessary to rebuild the national capability to execute on these [larger] construction projects,” he adds, noting that the cost of electricity from small reactors will be higher.

Another MIT expert, Jacopo Buongiorno, director of the university’s Center for Advanced Nuclear Energy Systems, agrees that large-scale reactors are more efficient. But there’s less risk to companies and investors from putting in $3 billion for an SMR instead of $10 billion for a large reactor. “It may be a good solution in the near term, as people will be more comfortable with making investments in nuclear,” he adds. “But I think long term, as far as providing large chunks of electricity on the grid, large reactors are probably the way to go.”

Optimistically, Buongiorno says, new U.S. nuclear projects will come online in the mid- to late 2030s: “You don’t snap your fingers and make it happen.”

Bridging the gap

The cost of energy produced by small modular reactors won’t be a problem in Virginia, though, Davis says, because purchasers like data center operators will agree to the higher prices. “Once you know what you’re getting up front [the price], there’s no risk,” he says. “The risk is there could be cost overruns, and what Gov. Youngkin has required is that the technology companies and the off-takers bear that cost overrun risk.”

But M.V. Ramana, the chair in Disarmament, Global and Human Security at the School of Public Policy and Global Affairs at the University of British Columbia, isn’t so sure.

Ramana has been a critic of SMRs because of their cost and construction delays. He questions whether agreements with data center operators would be ironclad. “It’s not going to be an unconditional thing, saying you’re going to pay how much you ask for, or you’re willing to wait for however long,” he says. “So, there are going to be conditions, and the question is whether the nuclear industry is going to be able to meet those conditions.”

Davis acknowledges that new nuclear projects will not be operating in time to meet near-term demand from data centers. “We’re working to bring about additional natural gas deployment to help bridge the gap to the early 2030s, and natural gas will be a player far beyond 2030 as well,” he says. “We’re counting on additional natural gas deployment or additional natural gas generation to help with some of the significant demand increases as Virginia continues to have significant economic development opportunities and job growth.”

While there are plans to bring more nuclear power to the state, none of them have firm timelines. Last year, Dominion Energy issued a request for proposals to evaluate developing an SMR at the North Anna nuclear power plant in Louisa County. And Amazon and Dominion signed an October 2024 agreement to jointly explore ways to advance SMR development and financing. Dominion’s most recent integrated resource plan calls for small modular reactors beginning in the mid-2030s.

“We’re still evaluating the proposals,” Dominion spokesman Tim Eberly says. “If we do move forward with developing an SMR, it will likely be at North Anna, but we have not yet committed to moving forward.”

Dominion is also leasing land at the James River Industrial Park in Chesterfield County to Commonwealth Fusion Systems, an MIT spinoff that plans to build the nation’s first fusion reactor.

Nuclear fusion, which powers the sun, creates heat that can be converted to electricity by heating two isotopes of hydrogen to 100 million degrees Celsius and fusing them. It’s the opposite of fission and produces much less nuclear waste. The fusion reactor, if built, will be about the size of a big-box store. Commonwealth Fusion, which has raised $2 billion from Bill Gates, Google and Italian oil and gas giant Eni, is working on a fusion demonstration machine at its Massachusetts headquarters, something that will pave the way for its first commercial fusion plant planned for Chesterfield.

While Dominion is offering technical expertise, it has no investment in the project and has not agreed to purchase any power from it. Google has agreed to purchase 200 megawatts, half the plant’s output, when it is built. There is no timeline other than that it will be completed in the early 2030s. Virginia’s Department of Energy and Chesterfield County will contribute $1 million apiece and the county will provide $10 million more in “longer-term support,” according to Youngkin. The plant equipment will be exempt from sales and use taxes.

Fusion has been the butt of jokes as being a breakthrough technology that is always a decade away. Earlier this year, a French fusion reactor set a new record for sustained reaction — lasting just more than 22 minutes.

Noting that Commonwealth Fusion spun out of MIT research, Parsons says, “I’m happy to be a cheerleader for that, but we’re running a marathon, and it hasn’t gotten through the first quarter of the marathon yet. It’s nowhere near the finish line, and people who are near the finish line and are cheering are just deluding themselves.”

Southwest power-up?

Nevertheless, in Wise County, local officials are cheering for a study they say could lead to Southwest Virginia’s return as the king of energy in the commonwealth, a crown it held during the reign of coal. In 2022, Youngkin stood on a former coal mine site in Wise County and said the area’s 100,00 acres of former coal land would make a good location for an SMR.

Those plans, however, have evolved.

Now, Wise County is the beneficiary of a $197,500 grant, partially funded by the state, to determine what kind of advanced nuclear reactor would be best, how it could be funded, and if it could be used for research and power at the University of Virginia Wise. “We are just now starting the process,” says Duane Miller, executive director of the LENOWISCO Planning District Commission, which serves the counties of Lee, Wise and Scott, and the city of Norton.

A timeline has yet to be established for the project. “The goal would be to have something within the next 10 years, but it could be sooner with how technology is evolving on this, or it could take longer,” Miller says, adding that regional officials would reassess if any red flags were raised. “I always tell people to complete a journey, you’ve got to start it.”

Aiding on the trek will be the Virginia Innovative Nuclear Hub, established this year by the Youngkin administration. Initially funded with $1.2 million from the Virginia Department of Energy, the hub is a collaborative research initiative involving the University of Virginia, Virginia Tech, Virginia Commonwealth University and Liberty University. About two-thirds of the funding will go toward examining “the feasibility, design, and implementation strategy” for establishing a research SMR in Virginia, while the remainder will be focused on researching “materials degradation in nonaqueous environments,” a challenge for next-generation nuclear reactors.

To begin developing SMRs here, Davis says, a supply chain needs to be built in the state. The first one or two SMRs may need offshore manufacturing help, perhaps from South Korea. “Then you’ll see manufacturing build up in Virginia, so that third, fourth, fifth reactor is mostly manufactured domestically,” he says.

The second part of ramping up will be regulatory, he adds, noting that the Trump administration is already streamlining processes with the Nuclear Regulatory Commission.

When he looks ahead to the 2030s, Davis sees nuclear in Virginia as part of that all-of-the-above energy strategy envisioned by Youngkin that will attract businesses and projects like data centers, creating jobs and rejuvenating the economy in places like Wise County.

“There was a time Southwest Virginia was the energy capital of the East Coast,” he says. “I believe they will be again, and it’s going to be led by small modular reactors down there.”

The Associated Press contributed to this story.