Nuclear Small Modular Reactors Market

Data centre operators signing power deals for the 2030s have done more to finance small reactors than a decade of energy policy did. Douglas Insights values the nuclear small modular reactors market at USD 4.80 billion in 2025 and expects USD 21.0 billion by 2035, compounding at 15.88%. The estimate is built from the bottom up: roughly 1,600 megawatts electric of small modular reactor capacity in active construction and pre-construction spending in 2025, at an average annual spend of USD 3.0 million per megawatt covering reactor module and nuclear island equipment, engineering, design and licensing, construction and balance of plant, and fuel supply and fabrication, triangulated against project milestones, licensing records, government funding and developer disclosures. Capacity in active construction grows 18.0% a year as the first projects move from licensing to build, while spend per megawatt falls 1.8% a year as designs mature, though first of a kind cost risk remains high. This study sits within our nuclear power coverage and follows the published Douglas Insights methodology.

Why did NuScale’s Utah small modular reactor project collapse before construction?

NuScale and Utah Associated Municipal Power Systems ended the Carbon Free Power Project in November 2023 because it was unlikely to win enough utility subscription to proceed, as their joint release filed with the SEC states. Rising cost under inflation and higher interest rates drove members away, and that first of a kind cost risk now shapes every SMR project.

The project’s projected cost rose to a level its customers would not pay, which remains the central risk every small modular reactor project faces. For years, small modular reactors were promoted as the answer to nuclear power’s problems: instead of building enormous, bespoke plants that routinely ran years late and billions over budget, factories would build smaller standardised reactor modules that could be assembled quickly and cheaply. The most advanced project in the United States, a plant using NuScale’s design for a group of Utah utilities, was cancelled in 2023 after its estimated power price rose sharply amid inflation and higher interest rates, and not enough utilities would commit. The episode showed that the promised cost advantages of small reactors depend on building many identical units, while the first few units carry all the costs of new designs, supply chains and licensing. Since then, the sector has regained momentum, supported by electricity demand from data centres, government backing and a first wave of projects actually entering construction, but the cost question remains open until the first units are built and their real costs are known. The exclusive chapter of this report tracks first of a kind cost estimates against actual spending by project, since that evidence will decide the sector’s future.

Which small modular reactor spending counts toward the USD 4.80 billion?

The USD 4.80 billion counts 2025 spending on small modular reactor projects in active construction and pre-construction, about 1,600 megawatts electric at USD 3.0 million per megawatt, across nuclear island equipment, construction, engineering and licensing, and fuel. Large reactors, microreactors, research reactors, uranium mining and the operation of finished plants are excluded.

The small modular reactor study covers spending on small modular reactor projects, generally reactors of up to about 300 megawatts electric per module, including light water and advanced designs. Reactor module and nuclear island equipment covers the reactor pressure vessel, steam generation, safety systems and other nuclear equipment. Engineering, design and licensing covers design completion, safety analysis and regulatory licensing. Construction and balance of plant covers civil construction, turbines, generators and conventional plant. Fuel supply and fabrication covers fuel for initial loading, including high assay low enriched uranium required by many advanced designs. Large conventional nuclear reactors, microreactors below a few megawatts used for remote sites, research reactors not producing power, uranium mining and conversion, and operations of completed plants sit outside the boundary. Value is measured at annual project spending. The model also splits spending by reactor type, between light water small modular reactors and advanced designs cooled by sodium, gas or molten salt, and by customer, between utilities, technology and data centre offtakers, and industrial users.

Why are Google and Amazon paying to bring small reactors onto the grid?

Google signed with Kairos Power on 14 October 2024 for 500 megawatts of advanced reactors by 2035 (Kairos Power release), and Amazon backed X-energy two days later with a target of more than 5 gigawatts of US projects by 2039 (X-energy release). Both want firm, carbon free power for data centres that never switch off.

The reason is that artificial intelligence data centres need enormous amounts of electricity around the clock, and companies with carbon targets want it to be clean and reliable, which nuclear power uniquely provides. Solar and wind are cheap but intermittent, and pairing them with enough storage to power a data centre continuously is expensive. Nuclear plants run at high output all day and night with very low emissions. As data centre electricity demand surged, several of the largest technology companies signed agreements to support new nuclear capacity, including small modular reactor developers, to secure future clean power, and some agreed to restart or extend existing reactors. These agreements provide something small reactor developers had lacked: creditworthy customers willing to commit to buying power from first of a kind plants. Together with government support in the United States, Canada, the United Kingdom and elsewhere, and national goals to expand nuclear capacity substantially, this has turned small modular reactors from a perpetual promise into a sector with projects under construction. It has not removed cost and execution risk, but it has created demand willing to absorb some of it.

Which forces lift SMR capacity in construction from 1,600 to about 8,400 megawatts?

Four forces lift small modular reactor capacity in active construction 18.0% a year, from about 1,600 megawatts electric in 2025 to about 8,400 in 2035 on the Douglas Insights model: data centre offtake, government funding, energy security policy and coal plant replacement. Spend per megawatt falling 1.8% a year trims value growth to 15.88%.

Data centre offtake is the first and newest driver. The Google and Kairos Power agreement covers 500 megawatts, and the Amazon and X-energy plan starts with a four-unit, 320 megawatt project with Energy Northwest in central Washington, expandable to 12 units and 960 megawatts. Those two named deals alone add up to more than half of the 1,600 megawatts in active construction in 2025, which shows how much of the forward SMR pipeline technology buyers now anchor. Grid access is the parallel constraint for the same buyers, covered in our Data Center as a Bottleneck Market report.

Government support is the second driver. Canada’s nuclear regulator authorised construction of the first BWRX-300 at Darlington on 4 April 2025 (CNSC decision), and Ontario Power Generation budgets CAD 20.9 billion for four units, including CAD 6.1 billion for the first (OPG release, 13 May 2025). In the United Kingdom, Great British Energy – Nuclear signed its contract with Rolls-Royce SMR on 13 April 2026 for a three-unit project of at least 1.4 gigawatts, backed by GBP 2.6 billion from the 2025 Spending Review (UK government release).

Energy security is the third driver. The United States banned imports of Russian low enriched uranium from 11 August 2024 under a law signed on 13 May 2024, with any waivers ending no later than 1 January 2028 (US Department of Energy waiver procedures). The ban pushes reactor buyers toward domestic fuel and component supply, which is why fuel supply and fabrication, 10% of 2025 SMR spending or USD 480 million, grows as first cores approach loading.

Coal plant replacement is the fourth driver. A retiring coal station already holds a grid connection, cooling water and a trained workforce, and a 300 megawatt BWRX-300 or a 345 megawatt Natrium unit is sized to reuse them. TerraPower’s first Natrium plant sits beside a retiring coal plant at Kemmerer, Wyoming, which makes it the clearest test of the coal-to-nuclear model.

What could stall SMRs, first of a kind cost, HALEU fuel or licensing queues?

Three restraints could hold small modular reactor capacity to the 8.0% a year path of the cost overrun scenario, which ends near USD 9.96 billion in 2035 instead of USD 21.0 billion: first of a kind cost, high assay low enriched uranium supply and licensing capacity. Cost is the largest, as the cancelled Utah project and Darlington’s CAD 6.1 billion first unit show.

Each restraint is modelled. First of a kind cost and schedule risk is the most important: new designs built for the first time routinely cost more and take longer than planned, and until early units demonstrate real costs, investors and utilities remain cautious. Fuel supply is second: many advanced designs require high assay low enriched uranium, whose commercial supply has depended largely on Russia, and with US import waivers for Russian fuel ending by 1 January 2028, building alternative supply is urgent but slow. Licensing and supply chain are third: nuclear licensing is lengthy, and the specialised nuclear supply chain has limited capacity after decades of little new construction in many countries.

Does nuclear island equipment or site construction take more of SMR spending?

Nuclear island equipment takes more: reactor modules and nuclear island equipment hold 40% of 2025 small modular reactor spending, USD 1.92 billion, against 28% or USD 1.34 billion for construction and balance of plant. The lead comes from the core nuclear equipment every module needs, from the pressure vessel to the safety systems.

Reactor module and nuclear island equipment covers the reactor pressure vessel, steam generation and safety systems. Construction and balance of plant, USD 1.34 billion, covers civil works and conventional plant, including the step-up equipment at the grid connection tracked in our Liquid Filled Power Transformer Market study. Engineering, design and licensing accounts for 22%, USD 1.06 billion, a large share at this early stage as SMR designs are completed and licensed. Fuel supply and fabrication contributes 10%, USD 480 million, and grows as projects approach fuel loading and new HALEU supply is built. Each category is modelled through 2035 by region.

Where are small modular reactors being built, and which region grows fastest?

North America leads small modular reactor spending with 40% of 2025, USD 1.92 billion, while the Middle East grows fastest at 18.0% a year, from USD 192 million to about USD 1.00 billion in 2035, because it starts from a small base. Canada’s Darlington build and US technology-backed projects keep North America ahead.

North America grows 16.8% a year to about USD 9.05 billion in 2035, driven by Canada’s first small modular reactor under construction at the existing Darlington nuclear site in Ontario, and by United States projects supported by technology company agreements and federal programmes. Asia Pacific holds 34%, USD 1.63 billion, at 14.2%, led by China, which has a land based small modular reactor nearing operation, together with programmes in South Korea, Japan and India. Europe holds 18%, USD 864 million, at 16.8%, with the United Kingdom’s Rolls-Royce SMR contract and programmes in Poland, the Czech Republic, Romania, Sweden and elsewhere. The Middle East contributes USD 192 million at 18.0%, Latin America USD 96 million at 12.0% and Africa USD 96 million at 14.0%. Six regional models sum to the global figure, with country tables in the Excel model.

Which companies lead small modular reactor development, and whose designs are being built?

Nine developers named in this report hold the designs that matter, and GE Vernova Hitachi leads on build progress: its 300 megawatt BWRX-300 at Darlington received Canada’s first licence to construct a small modular reactor on 4 April 2025. Rolls-Royce SMR, NuScale, TerraPower and X-energy follow with licences, contracts or technology backers.

GE Vernova Hitachi’s BWRX-300 is under construction in Canada and planned elsewhere, and Rolls-Royce SMR has been selected in the United Kingdom. NuScale holds an approved design in the United States despite the cancelled first project. Advanced reactor developers include TerraPower with its sodium cooled Natrium design, X-energy with its high temperature gas reactor, Kairos Power with molten salt cooled reactors, and Holtec, Westinghouse and others. China National Nuclear Corporation’s Linglong One is among the most advanced land based projects, and Russia has deployed floating small reactors. The competitive chapter profiles each developer’s design, licensing status, projects, customer agreements and fuel requirements. No leader share is published here: with only a handful of SMR units in active construction worldwide, a company share of spending would rest on one or two projects and would mislead.

What does a megawatt of SMR cost, from Darlington’s first unit to a series fleet?

Ontario Power Generation’s first Darlington unit, budgeted at CAD 6.1 billion for 300 megawatts, works out at about CAD 20.3 million per megawatt including interest, escalation and contingency, while the four-unit total of CAD 20.9 billion averages about CAD 17.4 million. That 14% drop from first unit to fleet average is the series saving SMR economics depend on.

Across the small modular reactor market, average spending is USD 3.0 million per megawatt of capacity in active construction per year in 2025, reflecting annual outlays on projects that take several years to build. Total overnight costs for first of a kind small modular reactors are uncertain and have been estimated at levels well above those of later units, with published estimates rising in some cases due to inflation and design maturity. Developers expect costs to fall substantially as designs are standardised and built in series, which is the basis for the slightly negative spend per megawatt leg, but this depends on building many units. Projects are financed through combinations of utility investment, government funding and loans, and customer agreements. The pricing chapter publishes cost ranges by design type and project stage, with the caveat that first of a kind costs remain highly uncertain.

How far apart do the SMR cost overrun and fleet deployment cases land by 2035?

By 2035 the small modular reactor market lands between about USD 9.96 billion and USD 32.3 billion, around a USD 21.0 billion base case, depending on whether first units such as Darlington hold their budgets. Almost all of the spread comes from capacity growth, which runs from 8.0% to 25.0% a year across the three cases.

The base case carries 18.0% growth in capacity in active construction and a 1.8% annual decline in spend per megawatt for a 15.88% revenue CAGR and USD 21.0 billion in 2035. The cost-overrun scenario, in which early units exceed budgets, customers withdraw and projects are deferred, sets the legs at 8.0% and minus 0.4%, landing near USD 9.96 billion. The fleet-deployment scenario, in which early units succeed, series production begins and technology company demand scales, sets them at 25.0% and minus 3.2%, carrying the market to about USD 32.3 billion. Each 1-point change in capacity growth moves the 2035 figure by roughly USD 1.76 billion. Confidence is low given first of a kind uncertainty.

Which approvals must a small modular reactor clear, from design licence to HALEU safeguards?

A small modular reactor clears three sets of approval before it earns revenue: a design and site licence from the nuclear regulator, fuel cleared under non-proliferation and import rules, and the funding terms of national energy policy. Canada’s Darlington licence to construct, issued on 4 April 2025 and valid to 31 March 2035, shows the first layer in practice.

Three layers matter. Nuclear licensing comes first: reactors require design certification or approval and site specific construction and operating licences from nuclear regulators, and licensing timelines strongly affect project schedules, with regulators working to modernise processes for new designs. Fuel and non-proliferation rules are second: production and use of high assay low enriched uranium are governed by non-proliferation safeguards and national security controls, and restrictions on Russian nuclear fuel imports affect supply. Energy policy and funding are third: government support, tax credits, loan programmes and national nuclear targets shape project economics and financing. The regulatory chapter maps these requirements by jurisdiction.

What would it take for SMRs to deliver on their promise?

The case for small modular reactors rests on a promise that has not yet been proven: that building many small, identical reactors in factories will be cheaper and faster than building a few large ones on site. Realising that promise requires several things to go right. The first units must be completed close to their budgets and schedules, to give investors and customers confidence. Developers must then secure orders for many units of the same design, since the cost reductions come from repetition, standardised supply chains and experienced workforces rather than from small size itself. Supply chains for nuclear components and fuel, particularly high assay low enriched uranium, must be built out. And regulators must be able to license multiple units efficiently. There is a real risk that the industry fragments across too many competing designs, each built too few times to achieve cost reductions. The model reflects steady growth as early projects proceed, with the fleet deployment scenario capturing the case where series production takes hold and the cost overrun scenario the case where early units disappoint.

Douglas Exclusive: the first of a kind cost and project tracker

The first of a kind cost and project tracker is a Douglas Insights model and estimate, not an official register. The dated SMR milestones quoted on this page, drawn from the seven regulator, government and company releases linked above, are the kind of public record it starts from, and cost lines that developers have not published are Douglas Insights estimates.

The tracker follows, by project and design, licensing status, construction milestones, estimated and reported costs, customer agreements, fuel arrangements and schedule changes, converting the project pipeline into capacity in active construction and spending by category and region. Licence holders receive it as a maintained tab in the Excel model.

Methodology and receipts

The small modular reactor model is built bottom-up in numbers: about 1,600 megawatts electric in active construction in 2025, multiplied by USD 3.0 million of annual spend per megawatt, gives USD 4.80 billion, split into four spending categories and six regional models that sum to the global figure. Capacity then grows 18.0% a year to about 8,400 megawatts while spend per megawatt falls 1.8% a year, giving USD 21.0 billion in 2035.

Inputs are small modular reactor projects by region, design and stage, construction timelines, spending per megawatt by stage, fuel requirements, and project and developer disclosures, with large conventional reactors, microreactors, research reactors, uranium mining and plant operations excluded. Every figure carries a numbered source and a confidence grade in the fact sheet above, and the working model ships with every licence. The next scheduled review of this study is September 2027.