A small modular reactor (SMR) is an emergent class of nuclear fission reactors with a rated electrical power of less than 300 megawatts (MWe), which use modular design principles to achieve streamlined construction and enhanced scalability compared to large light-water reactors. Many SMR designs are intended to be built in factories and transported to installation sites as prefabricated modules, while others are designed for flexible multi-unit configurations.
The term SMR refers to the physical size, electrical capacity, and modular construction approach. Reactor technology varies significantly among SMR designs. As of March 2026, most SMR designs are light-water reactors (LWRs), however SMR concepts encompass various reactor types, including generation IV, thermal-neutron reactors, fast-neutron reactors, molten salt reactor, and gas-cooled reactor models. Many SMRs also incorporate passive safety features.
Commercial SMRs are designed to deliver an electrical power output as low as 10 MWe and up to 300 MWe per module. Reactors below 10 MWe in capacity are considered nuclear microreactors. SMRs may also be designed purely for desalination or process heat rather than electricity. These SMRs are measured in megawatts thermal (MWt). Many SMR designs plan for customers to simply add modules to achieve a desired electrical output rather than scaling the size of the reactor. These reactors are also expected to enhance safety through passive safety systems that operate without external power or human intervention during emergency scenarios, although this is not specific to SMRs but rather a characteristic of most modern reactor designs. SMRs are also claimed to have lower power plant staffing costs, as their operation is fairly simple, and are claimed to have the ability to bypass financial and safety barriers that inhibit the construction of conventional reactors.
SMRs have attracted strong interest from technology companies, such as Google and Microsoft, for use in powering data centers to meet demand driven by the AI boom. Modular reactors are expected to reduce construction costs and time compared to large light-water reactors, as well as allow data center operators to draw power exclusively from a behind-the-meter SMR without purchasing electricity from the power grid.
Definition According to the American Nuclear Society (ANS), while small or modular reactor concepts date back to the 1950s, the term "small modular reactor" first entered common use in the late 1970s. Since then, the definition of an SMR has remained somewhat controversial.
John Fabian, writing in Nuclear Newswire, stated in 2026 that until around 2011, the abbreviation "SMR" referred either to "small modular reactor" or "small and medium sized reactor". According to Fabian, "the way 'SMR' is being used today does not seem consistent and is design dependent." For example, as of 2026 the United States Nuclear Regulatory Commission (NRC) defines SMRs as only light-water reactors under 300 MWe, while non-LWR designs are considered "advanced reactors". At the same time, the World Nuclear Association defines SMRs as any reactor under 300 MWe so long as it is designed with modular technology. According to Fabian, the classification of reactors as SMRs based on their power output has remained consistent, with reactors under 300 MWe being classified as SMRs.
As of 2023, there was broad consensus in the nuclear industry that SMRs are defined as nuclear reactors designed intentionally for low power (below 300 MWe) and with modular design applied either to individual components or the entire assembly.
Operational SMRs , only China and Russia have successfully built operational SMRs. Russia has been operating a floating nuclear power plant Akademik Lomonosov, in Russia's Far East (Pevek), commercially since 2020. China's pebble-bed modular high-temperature gas-cooled reactor HTR-PM was connected to the grid in 2021.
As of 2025, there were 127 modular reactor designs, with seven designs operating or under construction, 51 in the pre-licensing or licensing process, and 85 designers in discussions with potential site owners.
Background While small modular reactors have experienced a resurgence in the 2000s, they are not a novel concept, and have been under development for many decades.
Early small reactors Small reactors have been used for military use since the 1950s for nuclear marine propulsion. The thermal output of the largest naval reactor as of 2025 is estimated at 700 MWt (the A1B reactor), similar in capacity to some large SMR plant designs. Naval nuclear reactors have had an excellent record of safety. According to public information, the US Navy's Naval Reactors program has never succumbed to a meltdown or radioactive release over its 60 years of service. In 2003, Admiral Frank Bowman backed up the Navy's claim by testifying no such accident has ever occurred. Several nuclear-powered commercial vessels have also been constructed, including four nuclear freighters and nine nuclear icebreakers.
Lawrence R. Hafstad, the Director of Reactor Development at the United States Atomic Energy Commission, proposed building small, cheap, transportable nuclear power packages for use in remote areas. Alvin M. Weinberg, Director of Research at Oak Ridge National Laboratory (ORNL), wrote in support of Hafstad's proposal in 1952, adding that the naval reactors could provide a basis for developing nuclear power packages. He suggested that building many small reactors for remote sites would reduce the financial risk of establishing a nuclear industry:
Between 1954 and 1977, the US Army experimented with powering military installations with small land-based reactors during the Army Nuclear Power Program. The preliminary design for one such reactor was produced by ORNL under Weinberg's supervision. One of the Army reactors, PM-2A, was noted as extensively utilizing prefabricated modules to achieve lower construction cost.
Military small reactors are quite different from commercial SMRs. Historically, the military relied on highly enriched uranium (HEU) to power their reactors and not the low-enriched uranium (LEU) fuel used by SMRs. This is because submarine reactors are severely space-constrained and require higher power density than civilian reactors. Many naval reactors also operate for over a decade or more without refueling.
Early land-based reactors constructed in the United States were relatively small, such as the 60 MW Shippingport Atomic Power Station and the 250 MW Indian Point Unit 1. However, these reactors were intended to be scaled up into large central-station power plants, to take advantage of economies of scale. Soon after the first successful nuclear demonstrations, utilities raced to scale up plants before operational experience could be gained from smaller units. While traditional engineering wisdom dictates that a system should be scaled up twofold, by the time many nuclear plants with capacities exceeding 1000 MW were being ordered and constructed, no reactor larger than 250 MW was operational. While nuclear power plants were formerly designed as effectively standardized designs, the growth of operational issues and new safety requirements led to plants being designed and built as "one-of-a-kind" designs, further increasing cost, construction, and licensing issues. Reactor orders soon fell off sharply, and none of the plants ordered after 1974, except Watts Bar Unit 2, were ever constructed. The nuclear industry, throughout the following decades, would focus on optimizing the existing large reactors, whose average capacity factors rose from 50% to 90% by the mid-2000s.
Renewed interest In 1982, the Electric Power Research Institute (EPRI) conducted a study that recommended smaller, less capital-intensive, and inherently safe reactors be designed instead of large light-water reactors. Weinberg, then the director of the Institute for Energy Analysis, conducted a related study into inherently safe reactors. It concluded that such a system was possible, and selected the 400 MW Swedish Process Inherent Ultimate Safety (PIUS) and 100 MW American Modular High Temperature Gas-cooled Reactor (MHTGR) designs as the most intrinsically safe. At this time, multiple high-temperature gas-cooled reactors were being designed with the intention of building a large plant out of several modular low-power reactors. The PIUS reactor, with its reactor vessel situated within a large pool of borated water, and the robust TRISO fuel of the General Atomics MHTGR influenced subsequent SMR design features. Several small prefabricated reactor designs were developed around the same time in the UK, France, and Germany, however none were constructed. The EPRI study also led to the Department of Energy's (DOE) Advanced Light Water Reactor program, which resulted in the development of two 600 MW LWRs, the General Electric ABWR and later SBWR, and the Westinghouse AP-600 designs. However, these were later scaled up significantly into the ESBWR and AP-1000 reactors. Two AP-1000 reactors were later constructed as Vogtle Units 3 and 4, together costing $30 billion due to delays and cost overruns.
The DOE also ran an Advanced Liquid-Metal Reactor program, which led to the development of the Power Reactor Inherently Safe Module (PRISM) SMR by General Electric in the 1980s and 1990s. Each 1400 MW PRISM plant would be composed of nine 160 MW reactor modules, which would each be factory-fabricated and shipped by rail to the construction site. PRISM's use of multiple factory-fabricated modules to comprise a large power plant and its extensive use of passive safety features were both carried into later SMR designs. In 1983, the International Atomic Energy Agency started the Small and Medium Power Reactor Project Initiation Study to survey designs for small and medium sized reactors. The Nuclear Energy Agency launched a follow-on study in 1991 that evaluated 17 small or medium-sized reactor concepts, including several from the DOE's Advanced Light Water and Liquid-Metal programs.
Government support for reactor design steadily declined, leading to the DOE being issued no budget for nuclear development at all in 1998. However, in 1999 the DOE began the Nuclear Energy Research Initiative (NERI), followed by the Generation IV International Forum in 2000, and later the Global Nuclear Energy Partnership (GNEP) in 2006. The NERI program resulted in the development of multiple SMR designs, including liquid metal-cooled reactors and SMRs designed for process heat, as well as others based on existing LWR technology. In 2001, a report to Congress examining 50 MWe small modular reactors for powering remote areas, "found no substantive technical issues to hinder development and deployment of SMRs, and initial estimates of the electricity generation costs are comparable to, if not better than, those for current electricity supplies in typical remote areas." One concept, the IRIS SMR, would later receive significant commercial interest and funding under the GNEP, while another, the MASLWR, later formed the basis for NuScale's SMR. The GNEP and its associated programs also catalyzed significantly greater industry and utility interest in developing SMRs.
The term "small modular reactors" as opposed to "small-and-medium-sized reactors" was brought to wider use when US Secretary of Energy Steven Chu identified "small modular reactors" as "America's new nuclear option" in a 2010 Wall Street Journal op-ed, where he stated "SMRs would be ready to 'plug and play' upon arrival [on site]" and be more affordable. He announced that President Barack Obama had requested $39million for a new SMR design and licensing program.