Small Modular Reactors (SMRs), Explained
Small modular reactors, or SMRs, are nuclear reactors designed to be much smaller than the gigawatt-scale plants that make up today’s nuclear fleet, and to be built largely in factories rather than assembled piece by piece on site. The idea has been discussed for decades, but it has drawn intense new interest as electricity demand rises and buyers — including data center operators — look for large amounts of firm, carbon-free power. This explainer covers what SMRs are, how they differ from conventional reactors, and where the technology actually stands.
What makes a reactor “small” and “modular”
The “small” refers to output. SMRs are generally defined as reactors producing up to about 300 megawatts of electricity, versus roughly 1,000 megawatts or more for a typical large reactor. Some designs, sometimes called microreactors, are smaller still, in the single- or low-double-digit megawatt range. The “modular” refers to how they are meant to be built: as standardized units manufactured in a factory and shipped to the site, where several modules can be installed together and added over time. In theory, factory production and repetition drive down cost and construction risk — the two problems that have plagued large nuclear projects.
The main technology families
SMR designs fall into a few broad categories. Light-water SMRs use the same basic water-cooled approach as the existing fleet, shrunk and simplified; they are the closest to established regulatory precedent. A second group uses alternative coolants — molten salt, high-temperature gas such as helium, or liquid metal like sodium — which can run hotter and, proponents argue, offer efficiency and safety advantages. High-temperature designs are also of interest for industrial heat, not just electricity. Each family carries different fuel, materials, and licensing questions.
The safety argument
A central selling point is passive safety. Smaller reactor cores produce less heat, and many designs are engineered so that if something goes wrong, natural forces like convection and gravity carry heat away without operator action or external power. Advocates say this can reduce the size of the emergency planning zone around a plant and make siting more flexible, including near industrial users. These claims are part of what regulators must evaluate rather than settled fact, and they vary by design.
Where the technology actually stands
Interest and money have surged, but deployment is early. In the United States the Nuclear Regulatory Commission has certified or reviewed a handful of designs, and several companies are pursuing licenses for specific sites. The first commercial SMRs in the U.S. are still in development, and the sector has already seen setbacks: one prominent light-water SMR project was canceled after projected costs rose and subscribing utilities pulled back. That episode underscored that the core challenge — delivering nuclear power at a competitive, predictable cost — has not yet been solved at commercial scale.
At the same time, demand-side pull is real. Technology companies seeking round-the-clock clean power for data centers have signed agreements to purchase output from planned reactors, and some have backed specific developers. Those deals provide a potential customer base that earlier nuclear projects lacked, though most are contingent on reactors that do not yet exist.
How this era differs from the last
Nuclear power’s history is defined by large plants that ran badly over budget and behind schedule, and that record shapes every conversation about SMRs. The core argument for the new approach is that the industry’s cost problem is partly a construction problem: building enormous, first-of-a-kind structures on site, each one slightly different, invites delays and overruns. Standardized, factory-built modules are meant to convert nuclear construction into something more like manufacturing, where each unit is cheaper and more predictable than the last. Whether that logic holds in practice is the central unproven claim of the entire sector, and it will not be settled until multiple identical units are actually built in sequence.
Realistic timelines
Enthusiasm has run ahead of steel in the ground. Licensing a novel reactor design, securing fuel, arranging financing, and completing construction is a multi-year undertaking even in the best case, and the first commercial units in the United States are still working through those stages. Announced power-purchase deals and site plans are meaningful signals of demand, but they are commitments to buy from reactors that mostly do not yet exist. A useful habit when reading SMR news is to distinguish among four very different milestones: a design announced, a design licensed, a project financed, and a reactor operating. Most current headlines concern the first two.
The open questions
Several issues will determine whether SMRs live up to the promise. Cost is first: the factory-production thesis only works if enough identical units are ordered to achieve economies of scale, and early units are expensive. Fuel is second: some advanced designs require a more enriched fuel that is not yet widely produced. Licensing is third: regulators are adapting frameworks built for large light-water reactors to novel designs. Waste and public acceptance remain, as with all nuclear power. None of these is necessarily disqualifying, but each is unresolved.
The bottom line
SMRs represent a bet that nuclear power can be made smaller, standardized, and cheaper to build, and that firm clean electricity will command a premium as demand climbs. The engineering concepts are credible and the customer interest is genuine, but the technology is still moving from design and licensing toward first commercial operation. The honest status is promising and unproven — a sector to watch closely rather than one that has arrived.
This is an evergreen explainer and is updated as SMR designs are licensed and built.