Nuclear startups are back in the game with small reactors and big challenges

The nuclear industry is living in the fog of a renaissance. Old factories are being renovated, and investors are flooding startups with money. In just the last few weeks of 2025, nuclear startups are taking off Raised $1.1 billionpretty much Investor optimism Smaller nuclear reactors will succeed where the broader industry has faltered recently.

Conventional nuclear reactors are huge pieces of infrastructure. The newest reactors built in the United States — Vogtle 3 and 4 in Georgia — contain tens of thousands of tons of concrete, are powered by 14-foot-tall fuel accumulators, and each generate more than a gigawatt of electricity. But it was also eight years late and more than $20 billion over budget.

A new crop of nuclear startups hope that by downsizing the reactor, they can avoid both problems. Need more power? Just add more reactors. They argue that small reactors can be built using mass production techniques, and as companies produce more parts, they should get better at making them, which would lead to lower costs.

The magnitude of this benefit is an expert matter Still searchingBut today’s nuclear startups depend on being bigger than zero.

But manufacturing is not easy. Just look at Tesla’s experience: the company struggled mightily to produce the Model 3 in large numbers profitably — and had the advantage of being in the auto industry, where the US still had a lot of experience. American nuclear startups do not have this advantage.

“I have a number of friends who work in the nuclear supply chain, and they can talk about five to 10 items that we don’t make in the United States.” Milo WernerThe general partner at DCVC told TechCrunch. “We have to buy them from abroad. We have forgotten how to make them.”

Werner knows a thing or two about manufacturing. Before becoming an investor, she worked at new product introduction Tesla, and before that, she did the same at FitBit, launching four factories in China for the wearables company. Today, in addition to investing in DCVC, Werner co-founded NextGen Industry Group, which works to promote the adoption of new technologies in the manufacturing sector.

When companies of any size want to manufacture something, they face two main challenges, Werner said. The first is capital, which is often the biggest constraint since factories are not cheap. Fortunately for the nuclear industry, this shouldn’t be a big problem. “They’re overwhelmed in D.C. right now,” she said.

But the nuclear industry is not immune to another challenge that all manufacturers face: a lack of human capital. “We haven’t built any industrial facilities in 40 years in the United States,” Werner said. As a result, we lose muscle memory. “It’s like we sit on the couch and watch TV for 10 years and then we get up and try to run a marathon the next day. That’s not good.”

After decades of offshoring, the United States lacks people with experience in plant construction and operations. “There are certainly some people in the United States who do this, but we don’t have the number of people we need for everyone to have a full staff of experienced manufacturing workers.” She’s not just talking about machine operators, but everyone from plant supervisors to financial managers to board members.

The good news is that Werner sees a lot of startups, nuclear and otherwise, building early versions of their products close to their technical team. “This draws manufacturing closer to the U.S. because it allows them to have that improvement cycle.”

To reap the benefits of mass manufacturing, it pays for startups from all fields to start small and then scale. “Really leaning into modularity is very important for investors,” she said. A modular approach helps companies start producing small quantities early so they can collect data about the manufacturing process. Ideally, this data will show improvement over time, which can comfort investors.

The benefits of mass manufacturing do not happen overnight. Companies often anticipate the cost reductions that can result from learning through manufacturing, but it may take longer than they expect. “A lot of times it takes years, like a decade, to get there,” Werner said.