Reduced capital costs, phased commissioning, fleet learning and factory fabrication make SMRs the future of clean, low-cost power
Other things being equal, an SMR should produce a unit of electricity at higher cost than a gigawatt-scale plant. That is the diseconomies of scale. Proponents accept the point but say other things are not equal. This briefing addresses those arguments and explains the four factors which SMR proponents advocate: lower cost of capital, phased commissioning, fleet learning and factory fabrication.
What are the implications for investors and policymakers?
Be realistic about the first units. The diseconomies of scale will not be overcome on a first-of-a-kind SMR. Fleet benefits, learning rates and even much of the cost-of-capital advantage from faster construction only fully crystallise with 'nth-of-a-kind' delivery.
To make SMRs investable, policymakers must distinguish between cost and time expectations for initial and nth-of-a-kind units. More fundamentally, investors need confidence in a pipeline, so that the units now being built read as part of a fleet, even if only on a multi-jurisdictional basis, rather than a series of one-offs. Capturing the benefits of modularisation requires as much of the following as possible:
- Streamlined licensing for modular designs: licensing and consenting pathways that reflect how SMRs are actually built and avoid duplicative review of identical units.
- Proportionate regulatory oversight of manufacturing: regulators engaging proportionately in the manufacture of modules and components, reflecting the modular approach.
- Reduced cross-jurisdictional regulatory overlap: tackling ‘double regulation’ of components and designs already approved by trusted foreign regulators, especially where SMRs are made in one country and deployed in another.
- Timely export control procedures: export controls that minimise duplication and delay while continuing to support safety and non-proliferation objectives.
The structural cost disadvantages
Let’s start with the economics of scale. Many fixed costs, from the containment structure to the turbine hall, do not fall linearly with rated output. Halve the megawatts and you do not halve the steel, concrete or control room. The International Energy Agency’s 2025 analysis starts from the assumption that overnight capital costs for a first-of-a-kind SMR are roughly double those of an efficiently delivered large reactor. A 2024 review of the published literature puts the average capital cost premium at around 41%. Treat these numbers with caution, since this is not real-world data yet, but they show what the diseconomies of scale imply.
The second issue is regulatory: the consenting envelope scales only weakly with rated output. Generic or pre-construction design review (eg, US NRC’s design certification, the UK Generic Design Assessment) typically takes three to five years, paid for by the vendor. On top of that sits a suite of site-specific consents that do not necessarily shrink for a smaller reactor: planning or development consent; full environmental impact assessment; an operating licence; environmental permits for radioactive discharges and cooling water; water abstraction and marine consents; security and safeguards arrangements; and decommissioning and waste-management plans, funding and interim storage. So, a 300 MWe reactor passes through substantially the same regime as a 1,200 MWe one. This loads a disproportionate fixed cost onto each megawatt-hour and a disproportionate lead time onto pre-construction development.
Even today’s most expensive nuclear builds offer good-value low-carbon energy once system-wide costs are counted. The promise of SMRs is that faster builds, lower cost of capital, fleet learning and factory fabrication will compound into something materially cheaper still. But that requires proper fleets.
Paul Butcher
Director of Public Policy
Why cost of capital is key
Arguably, the lever that matters most for SMR cost competitiveness is not overnight capital cost. It is the cost of capital. Nuclear projects are capital-intensive and front-loaded, and for a decade-long build at commercial rates, interest during construction can rival the civil and mechanical works in magnitude. Sizewell C (3.2 GW), which took its Final Investment Decision in July 2025, is expected to take between nine and twelve years to construct and commission. GE Vernova Hitachi targets 24 to 36 months from first nuclear concrete pour to ready for fuel load for nth-of-a-kind units of its BWRX-300 SMRs.
If this is achieved, it changes how you think about financing nuclear. GLOBSEC offers an illustrative comparison. At a 5% cost of capital, a conventional plant at US$6,600 per kW built over 15 years ends up costing around US$13,721 per kW once interest during construction is compounded. An SMR at US$10,000 per kW built over five years would work out at around US$12,763 per kW. Under this illustration, the SMR would be cheaper despite a 50% higher overnight capital cost. Phased commissioning reinforces the effect: modules can come online one by one, so revenue starts flowing before the full plant is complete. The IEA expects SMR projects to reach cash-flow break-even up to a decade earlier than large reactors.
Fleet effects, sooner
The next lever is repetition. South Korea shows that standardisation and serial construction can cut costs dramatically even with entirely conventional, stick-built methods.
You can see fleet effects in the UK right now. EDF reports that the second EPR unit at Hinkley Point C is going up 20 to 30% faster than the first, with prefabrication approaching 60%. Its reactor pressure vessel was installed at the end of May 2026, and the experience carries into the two identical units at Sizewell C. And because SMRs are quicker to build, their fleet effects should arrive sooner.
There is a further potential advantage: SMR designs are meant to be maximally site-agnostic, avoiding bespoke variants from one site to the next. Standardisation pays beyond design and supply chains too, in training programmes, operations, monitoring and other areas.
What does South Korea teach SMR policymakers?
A 2016 study of 349 reactors in seven countries named South Korea as the only one whose construction costs fell, in real terms, across its entire build-out.
Between its first reactor in 1971 and 2008 its costs fell by roughly 50%. This was achieved with conventional, on-site construction and no factory modularisation.
The OECD Nuclear Energy Agency’s 2020 comparison puts a more recent figure on it: Shin Kori 3 and 4, the first APR-1400s, came in at an ex-post overnight construction cost of around US$2,410 per kW, against roughly US$8,620 per kW for Flamanville 3 and US$8,600 per kW for Vogtle 3 and 4 (both pre-completion estimates). An SMR programme needs the same ingredients. Korean projects have typically started construction with 70 to 80% of the detailed design complete, against an estimated 40% for Flamanville. From its tenth reactor onwards, Korea froze a single domestic design (the OPR-1000), repeated it, and evolved it incrementally into the APR-1400. On 12 April 2026 Saeul-3 achieved first criticality, the ninth APR-1400 worldwide to start up (counting the four units at Barakah in the UAE); commercial operation is expected later in 2026.
Repetition did more than cut build costs. The pipeline cultivated a domestic supply chain, made long-term investment in improvement worthwhile, and avoided the lengthy delays commonplace elsewhere in the nuclear sector. That virtuous circle is a powerful reason for policymakers to look beyond any initial challenges.
Modularity and factory fabrication
Modularisation, where it can be delivered, sits on top of fleet effects as a further accelerator. Holtec says its SMR-300 will be substantially factory built and site assembled, with every component limited to 12 feet in diameter and to practical maximum weights, for manufacturing and transport flexibility. Co-siting reinforces the point: where several SMRs share one location, fixed siting costs such as grid connection get paid once, for the first unit, and spread across the rest.
The cost potential of standardisation and factory fabrication is well documented in other industries. Shipbuilding and aircraft manufacturing have cut unit costs by 10 to 20% with each doubling of cumulative output, the so-called learning rate. For SMRs, learning rates of around 5% per cumulative doubling have been called pessimistic and 15% more optimistic. The way to get the high rates is a standardised design with little or no variation between units.
None of this counts gigawatt-scale plants out: EDF Energy now calls Hinkley Point C a "large modular reactor", Sizewell C is priced around 20% lower as its virtual replica, and its EPR2 design has been designed for prefabrication and serial construction in pairs.
SMRs are less a new technology than a new way of delivering nuclear power. The real opportunity for policymakers is to apply the lessons of large-scale nuclear successes, particularly in France and South Korea, so that SMRs get past first-of-a-kind challenges and realise their full benefits.
Andrew Robinson
Senior Consultant
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
[View Source]
