Our Take
China's nuclear strategy works because it chose one design, built it at scale, and moved fast; the US is betting smaller reactors will solve a problem that is actually about governance and capital, not physics.
Why it matters
Nuclear capacity is the real constraint on US data center buildout and grid decarbonization. If China locks in cost leadership on large reactors while the US waits for microreactors to prove themselves, the West cedes a critical infrastructure advantage.
Do this week
Energy procurement teams: model your 2028-2032 load forecasts assuming US large-reactor supply stays flat and price pressure comes from imported reactor designs or long-term power purchase agreements tied to Chinese nuclear output.
China built eight new reactors while the US completed none
China started construction on six new large reactors in 2025 and two more in the first five months of 2026. The country has roughly doubled its nuclear fleet since 2016, reaching nearly 60 gigawatts of capacity (per MIT Technology Review, citing official Chinese data). The United States, by contrast, completed Unit 3 at Georgia's Plant Vogtle in that same period. Unit 4 is still under construction.
The pace difference is structural, not accidental. China's average build time for a new reactor sits between five and seven years. The global average is nine years. The two most recent US reactors took about 15 years each (company-reported). China achieves this speed through standardization: a single uniform project management system handles design, licensing, and construction across all plants. New reactors are built in batches of six or more to capture economies of scale.
France, the other nuclear incumbent, connected one reactor to the grid in December 2024, its first in over 20 years. Like the US, it has spent recent years financing smaller modular reactors as a hedge against the cost and complexity of large-scale projects.
The West is betting on smaller; China is executing at scale
The US and Europe have pinned nuclear growth on microreactors and small modular reactors (SMRs). The logic is sound on paper: smaller up-front capital requirements, factory-built components, faster licensing. Last week, California-based Antares achieved criticality on its Mark-0 microreactor as part of a Department of Energy pilot targeting three test units by July 2026. The company plans to produce electricity in late 2027 and deploy in the field by 2028.
Big Tech is throwing capital at these ventures, hoping microreactors can power data centers without the regulatory and timeline risk of large plants.
But here is the cost math China understands: larger reactors produce more electricity per unit cost. Smaller reactors require less capital upfront but generate electricity more expensively per megawatt-hour. China is on course to overtake both the US and the European Union in installed nuclear capacity by 2030. It is also deploying its first operational SMR, the Linglong-1, this year, showing it is not abandoning small reactors. It is simply not waiting for them to solve a problem that standardization and state coordination already solved for large ones.
Data center operators and utilities must plan for a two-speed nuclear future
US electricity demand from AI inference and data centers is growing faster than new reactor supply can match. China's nuclear output will not directly serve US demand, but it signals a durable cost and speed advantage that will reshape global supply chains for reactor components, fuel, and operating expertise. If US microreactor projects slip or cost overruns hit large-reactor construction, utilities and tech companies will face a multi-year gap between peak demand and available capacity.
The real question is not whether microreactors are a better design. It is whether the US can standardize project delivery, compress licensing timelines, and sustain capital investment at the scale China has already achieved. Neither small nor large reactor wins on technology alone. China is winning on execution.