What Is a Small Modular Reactor?
A small modular reactor (SMR) is a nuclear fission reactor with an electrical output typically below 300 megawatts (MWe)—compared to 1,000 to 1,600 MWe for conventional large-scale nuclear plants—designed to be factory-manufactured in standardized modules and assembled on-site rather than custom-built. The modular design is intended to reduce construction costs, shorten deployment timelines, enable incremental capacity additions, and make nuclear power economically viable for locations, grids, and applications that cannot accommodate a large conventional plant.
SMR technology encompasses several distinct reactor designs: light-water reactors (a scaled-down version of conventional nuclear technology), advanced gas-cooled reactors, molten salt reactors, high-temperature gas reactors, and microreactors below 10 MWe. The majority of SMR projects in advanced development in the United States, Canada, and the United Kingdom are based on light-water reactor designs, which benefit from the deepest regulatory familiarity and the most extensive operating history.
The AI infrastructure connection
The convergence of AI infrastructure buildout and SMR commercialization is the defining energy story of 2024-2026. Hyperscale data centers require between 100 MW and 1 GW of continuous, reliable power with near-100 percent uptime—a demand profile that solar and wind generation cannot meet without massive battery storage and grid backup. SMRs offer 24/7 carbon-free baseload power, independent of weather and grid interconnection queues, making them a theoretically ideal match for behind-the-meter data center power.
Microsoft signed a 20-year power purchase agreement with Constellation Energy in September 2023 to restart Unit 1 at Three Mile Island—a large conventional plant, not an SMR, but reflecting the same underlying logic. Amazon, Google, and Microsoft have all signed preliminary SMR agreements or development partnerships: Amazon with X-energy, Google with Kairos Power, and Microsoft with Terrapower and Helion (fusion). The first commercial SMR deployments in North America are targeted for the late 2020s to early 2030s.
Bankability and the commercialization challenge
Despite the strategic logic, SMR commercialization faces significant bankability challenges. The cost of the first several deployments is highly uncertain: unlike mature technologies with established reference plants, SMR capital costs remain estimated rather than observed. NuScale Power, which received the first SMR design certification from the U.S. Nuclear Regulatory Commission in 2022, cancelled its flagship Carbon Free Power Project in Utah in November 2023 after cost estimates for the first 462 MW plant rose from $6.1 billion to over $9 billion, and the subscriber utility pool collapsed.
The bankability problem is structural: SMRs need several deployments to drive down cost through learning and standardization, but each early deployment must be financed at pre-learning cost levels that make the economics marginal without policy support. This is the classic first-mover problem in new energy technology, addressed in previous clean energy cycles through production tax credits, loan guarantees, and contract-for-difference mechanisms. The U.S. Inflation Reduction Act’s nuclear production tax credit and DOE loan guarantees are the primary policy instruments supporting SMR commercialization in the U.S.
Regulatory pathway
SMRs must receive design certification from national nuclear regulators—the Nuclear Regulatory Commission in the U.S., the Canadian Nuclear Safety Commission, the Office for Nuclear Regulation in the UK—before deployment. Design certification is a multi-year process that constitutes a major time and cost barrier for new entrants. The NRC completed its first SMR design certification (NuScale VOYGR) in 2022 after a decade-long review process. Several additional SMR designs are in the pre-application or active review phase.
Why it matters for emerging markets
SMRs are relevant to emerging markets as both an energy sovereignty instrument and a potential AI infrastructure enabler. Countries with limited grid capacity, high diesel dependency, or constrained interconnection infrastructure are candidates for SMR deployment once commercialization costs fall. Several Middle Eastern and Southeast Asian governments have expressed interest in SMR procurement as part of energy-transition and AI-infrastructure strategies. The critical caveat is timeline: commercial SMR deployments at scale are unlikely before the early 2030s, and the regulatory, financing, and supply-chain infrastructure required for emerging-market deployment is not yet in place.
Related Juncture terms
- Grid Interconnection
- Behind-the-Meter Generation
- Federal Compute Zone
- AI Special Economic Zone
- Bankability
- Safeguards
Related analysis
- [When Digitization Becomes Dependency: How AI Infrastructure Is Redrawing the Sovereign Risk Map](https://juncturepolicy.org/moldova-india-ai-sovereignty-gap-emerging-markets/)
- [Fortress AI Infrastructure](https://juncturepolicy.org/fortress-ai-infrastructure)
- [MDB Nuclear Bankability](https://juncturepolicy.org/mdb-nuclear-bankability)
Sources
- U.S. Nuclear Regulatory Commission, NuScale Power Module design certification, 2022 (nrc.gov)
- U.S. Department of Energy, Advanced Reactor Demonstration Program documentation (energy.gov)
- IAEA, Advances in Small Modular Reactor Technology Developments, 2022 edition (iaea.org)
- NuScale Power, Carbon Free Power Project termination notice, November 2023 (nuscalepower.com)