Former SpaceX Engineers Are Reviving a Forgotten Nuclear Project. AI Data Centers Could Be the Customers.

A $400 million nuclear experiment that sat dormant for nearly a decade is getting an unexpected second life as former SpaceX engineers race to find enough electricity to power the data centers behind the artificial intelligence technology.

According to a report from The Washington Post, nuclear startup Applied Atomics is working with BWXT to revive mPower, a small modular reactor project that was abandoned in 2017 after years of development. The centerpiece is a roughly 120-foot-tall test facility outside Lynchburg, Virginia, designed to mimic the workings of a nuclear power plant without using nuclear fuel.

Applied Atomics co-founder Ben Kellie, a former SpaceX engineer who led engineering for the company’s inaugural West Coast launch from Vandenberg, sees the dormant facility as a shortcut in an industry where developing new reactor technology can take years and consume billions of dollars.

“We’re standing on the shoulders of giants here, right?” Kellie told The Washington Post. “We’re starting with something that had $400 million in investment put into it, and significant time and effort. That’s a leg up.”

The renewed interest comes as the AI boom dramatically increases electricity demand. Data centers require enormous amounts of continuous power, putting pressure on electric grids and pushing technology companies to search for reliable sources that can operate around the clock while producing relatively low carbon emissions.

That search has helped revive interest in nuclear energy, including small modular reactors, or SMRs, which developers hope can be manufactured and deployed more quickly than conventional nuclear plants.

Applied Atomics’ approach differs from some of the more experimental reactor technologies attracting Silicon Valley investment. Rather than relying on novel fuels and cooling systems, mPower is essentially a smaller version of the light-water reactors already operating across the United States.

The proposed reactor would generate about 195 megawatts of electricity, enough to power roughly 155,000 homes. Applied Atomics says the first commercial reactor could begin operating in about five years if development proceeds as planned.

The company envisions customers installing multiple reactors together, potentially five or 10 at a single site, creating enough capacity for major data center or industrial projects. There is also a more unconventional possibility: putting the reactors at sea.

Applied Atomics and BWXT are working with Core Power, which is exploring floating nuclear power plants that could be anchored offshore and supply electricity to nearby grids, data centers or industrial facilities.

“We have great interest from utilities in the U.S.,” Core Power CEO Mikal Boe told the Post. “Several are engaging with us on where they could put this.” The concept is not entirely new. Babcock & Wilcox, which later became BWXT, worked on nuclear-powered merchant ships beginning in the 1960s.

Government-backed demonstration vessels operated from the United States and West Germany, but the technology could not compete economically with conventional marine fuel, and economics also helped kill the original mPower project.

Development began in 2009, but the U.S. shale boom sent natural gas prices tumbling and undermined the business case for new nuclear plants. The 2011 Fukushima disaster in Japan further damaged enthusiasm for nuclear construction. By 2017, BWXT had ended the project.

Today, the calculation has changed as electricity demand is climbing, while technology companies are increasingly willing to pay premiums for dependable, low-emissions power. Yet the obstacles that have plagued nuclear development remain.

Small reactors still produce radioactive waste, and the United States has no permanent national repository for spent nuclear fuel. Projects must also navigate lengthy regulatory processes, community concerns and the industry’s history of construction delays and cost overruns.

David Schlissel, a longtime nuclear consultant to consumer and environmental organizations, cautioned that the current excitement resembles previous attempts to revive the industry.

“We are now living in about the fifth supposed ‘nuclear renaissance’ in the U.S.,” Schlissel told the Post, arguing that repeated promises of cheaper and faster reactors have yet to materialize commercially.

Kellie believes lessons from SpaceX could change that equation. He argues nuclear’s biggest problem is no longer figuring out how reactors work but learning how to build them efficiently, repeatedly and at predictable costs.” We know how to split the atom,” Kellie said. “We’ve been doing it for a very long time.”