AINEWS
Two nuclear bets, neither proved at scale
China's standardized gigawatt reactors and America's factory built small reactors both depend on a learning curve. Neither has reached the volume needed to show which one bends.
China's standardized gigawatt reactors and America's factory built small reactors both depend on a learning curve. Neither has reached the volume needed to show which one bends.
China is betting that the path to cheap nuclear power runs through giant, standardized reactors built one at a time on site. The United States is betting the answer is small reactors built in factories and shipped out. The evidence on whether either model actually works is, for now, mostly absent.
The reason this matters now is that electricity demand is no longer a flat line to manage. Data centers, building electrification, and the slow reshoring of energy-heavy industry are all pulling grids in new directions at once, and clean firm power, the zero-carbon kind available around the clock, is the part no one has figured out how to build cheaply. MIT Technology Review's June 2026 analysis of the two strategies frames the split as a comparison of two learning curves, not two countries' construction speed.
China has nearly doubled its nuclear fleet since 2016, reaching roughly 60 gigawatts of capacity. Almost every new unit going online is a gigawatt-scale pressurized-water reactor, the same basic design that has powered most of the world's commercial nuclear plants for decades, just standardized at around 1,000 megawatts apiece. The model assumes that doing the same thing many times in a row eventually makes each one cheaper, the way solar panel factories did, and that standardization is the only way to get there.
The United States has built exactly two reactors since 2016: Vogtle Unit 3 and Unit 4 in Georgia, both of which are widely reported to have come in years late and well over their original budgets. France, the world's second-largest nuclear fleet, has also added almost no new reactors in that period, even though roughly two-thirds of its electricity already comes from nuclear. The most recent US milestone of a different kind came when a microreactor developer reported achieving criticality, the point at which a chain reaction becomes self-sustaining, inside a new Department of Energy pilot program—a milestone reported by the Associated Press—a step toward showing that smaller, factory-style designs can go critical on American soil at all.
The small modular reactor thesis is that building reactors in a factory, the way aircraft or cars are built, removes the local permitting and supply-chain penalties that broke Vogtle's budget. The Chinese thesis is that doing the same big reactor over and over, with as little variation as possible, drives costs down through repetition. Both bets depend on a learning curve. Neither has reached the volume needed to prove its curve is real.
The cleanest test of the US factory thesis is unit cost on the first commercial SMRs, not the demonstration units, and whether customers sign long-term offtake contracts at prices that beat combined-cycle gas. The cleanest test of the Chinese thesis is whether gigawatt-PWR costs in China are still falling as the fleet grows past 60 GW, or whether they have flattened the way reactor costs did in the West. Neither test is finished.
What the reader is actually being asked to pay for, through rates and taxes, is a bet on which curve bends. If the US SMR bet works, future clean firm capacity is cheaper to build and faster to deploy, and the worst economic case for nuclear disappears. If the Chinese bet is the one that pays off, the lesson is that the United States needs to standardize on a single design and stop treating every reactor as a one-off construction project. If neither works, the gap between rising electricity demand and zero-carbon supply has to be filled by something that is not yet visible on a grid interconnection queue.