Small modular reactors have been described as the next big thing in energy for nearly two decades. What has changed is the money behind that description. Governments are funding supply chains, tech companies are signing power deals years before a single reactor exists, and vendors are reporting real revenue for the first time. But the number of SMRs actually generating electricity anywhere in the world can still be counted on one hand. The gap between financial momentum and physical deployment is the real story behind the renaissance narrative.
The U.S. Department of Energy has moved from research funding to direct deployment support. In May 2026, DOE selected eight companies to share more than $94 million in cost-shared funding aimed at clearing licensing, supply chain, and site-preparation bottlenecks for Generation III+ SMRs, following an $800 million Tier 1 award to the Tennessee Valley Authority and Holtec Government Services in December 2025 for projects in Tennessee and Michigan. That funding builds on a March 2025 solicitation that made $900 million available specifically to de-risk near-term SMR deployment. A separate DOE analysis found that replicating the $51 billion in incentives the federal government spent on renewables over the past decade could deliver a return three times cheaper per kilowatt-hour if applied to SMRs instead the kind of cost argument now shaping U.S. energy policy.
International tracking shows a wide gap between announced projects and actual output. The International Atomic Energy Agency tracks more than 80 SMR designs in various stages of development worldwide, and the World Nuclear Association's SMR Global Project Tracker catalogs more than 100 design concepts across over 20 countries. Yet the IAEA also confirms only four SMRs are currently in advanced stages of construction globally, located in Argentina, China, and Russia, and separate reporting citing World Nuclear Association data puts total commercially operating SMR units worldwide at just four as of early 2026 Russia's floating Akademik Lomonosov plant and China's Linglong One demonstration reactor among them. The International Energy Agency's own 2026 review notes only one 125 MW commercial SMR under construction in China and one 300 MW unit in Russia, with the U.S., Canada, and the U.K. still expected to begin construction in the near term rather than having already started.
NuScale Power, the only SMR vendor with U.S. Nuclear Regulatory Commission design approval, reported full-year 2025 revenue of $31.5 million, down from $37.0 million in 2024 as licensing revenue from its Romania-based RoPower technology agreement wound down through the year. The company ended the first quarter of 2025 with $521.4 million in cash and short-term investments after raising $102.4 million through an at-the-market share program, giving it a funding runway even as commercial revenue remains small relative to its market capitalization. That gap between balance-sheet strength and operating revenue is common across the sector: investors are pricing in a 2030s deployment wave that has not yet translated into meaningful sales for any SMR vendor.
The clearest new demand signal is coming from technology companies, not utilities. Deloitte's power and utilities research projects that U.S. data center electricity demand could rise roughly fivefold to 176 gigawatts by 2035, and estimates that new nuclear capacity including SMRs could realistically meet about 10% of that increase, requiring 35 to 62 gigawatts of new nuclear generation this decade. Boston Consulting Group's own modeling, cited in trade press coverage, projects 10 to 30 gigawatts of installed SMR capacity by 2040, equivalent to roughly 50 operating units a fraction of the more than 80 designs currently in development. Recent corporate deals illustrate the shift: Google signed an agreement with Kairos Power targeting 500 megawatts of SMR capacity by 2035, and Amazon has funded small modular reactor projects in Washington state, both moves aimed at securing carbon-free power for AI data centers rather than replacing conventional grid generation.
Economics remain the sector's biggest open question. Deloitte's analysis puts 2024 nuclear construction costs at $6,417 to $12,681 per kilowatt, compared with just $1,290 per kilowatt for new natural gas capacity — a five-to-ten-times premium that SMR vendors argue factory fabrication will eventually close, but has not yet closed at commercial scale. Vendor-submitted cost data tracked by independent nuclear-economics analysts puts first-of-a-kind SMR costs at $4,000 to $7,000 per kilowatt for eventual nth-of-a-kind builds, still above large gas plants but positioned as competitive with new large-scale nuclear. Until a first-of-a-kind U.S. project is actually built and its real costs disclosed, these figures remain vendor projections rather than proven outcomes.
The honest answer sits between the two labels. Government funding, corporate power-purchase agreements, and a genuine data-center demand driver are real and growing, distinguishing this cycle from earlier false starts in the 2000s and 2010s. But with only four SMRs operating commercially worldwide, IEA-projected capacity of just 10 to 25 gigawatts by 2035 representing 1% to 3% of global nuclear capacity, and no U.S. construction license yet issued as of early 2026, the physical evidence still lags far behind the financial and political enthusiasm. Small modular reactors look less like a bubble about to pop and more like an industry in its expensive, slow-moving infancy — one that will be judged not by funding announcements, but by how many of these 80-plus designs actually reach the grid this decade.