Too Small to be Viable, Small Modular Nuclear Reactors (SMRs) Part 2 - From the Past Until Today - Cost & Risk Analysis
Smith Wattanasoponvanij
Independent Technology Consultant - l Process & Technology Development l Technology Research l Decarbonization l Green Technologies l
1. Capital Expenditure (CAPEX)
Both conventional large reactor and SMR suffer from high overnight costs. SMR designs aim for lower unit costs through modularity. First-of-a-kind (FOAK) SMRs, however, are coming in high. NuScale saw its construction budget revised to $9.3?billion (a 75% increase).
Licensing and Regulatory
Large reactor projects incur substantial up-front regulatory costs for design certification, site licensing, and safety analysis. SMRs face the same rigorous safety reviews, but regulators are adapting processes for smaller designs. While SMR designs are smaller, licensing is not proportional to reactor size, and each new design must undergo independent review. It means initial SMR developers bear high regulatory costs similar to large reactors. Efforts are underway to standardize and streamline SMR licensing, which could reduce costs in the long term. Until regulators harmonize approvals internationally, SMR vendors pursuing multiple markets must pay separate licensing fees and comply with multiple regulatory regimes – a costly barrier.
Site Preparation and Infrastructure
Large plants require extensive site works – heavy foundations for massive containment buildings, large cooling systems (cooling towers or water intake/outfall structures), and robust grid interconnections. Site-specific infrastructure costs can be billions (land, earthworks, transmission upgrades). SMRs by design have a smaller footprint and often simpler balance-of-plant requirements. Many SMR concepts (e.g. NuScale) use underground or compact containment, reducing land use. They also target existing brownfield sites (retired coal plants or existing nuclear sites) to leverage infrastructure and grid connections, cutting prep costs. However, multiple SMR units may require duplication of some infrastructure (each module needs its pad, cooling interface, etc., though multi-module plants can share facilities).
Economies of scale favor large reactors for certain site costs (one large cooling system vs. several smaller ones). Still, factory-built modules can mitigate on-site work, minimizing weather delays and on-site labor. The shift from field construction to module assembly is expected to reduce on-site construction time and cost uncertainty.
Fabrication/Construction
The modular construction approach of SMRs is a key differentiator in CAPEX. Large reactors are essentially one-off projects – custom built on-site, which historically leads to complex project management and risk of cost overruns. SMRs are intended to be manufactured in series, with parts or entire modules built in controlled factory settings. Factory fabrication can improve quality control and productivity (analogous to shipbuilding or aircraft industries). It also enables a manufacturing learning curve – unit costs decline as workers gain experience and tooling is optimized.
However, realizing cost savings requires volume: a factory production line is expensive to set up and only pays off if dozens of units are ordered. Early SMR deployments thus carry the cost of establishing production facilities. Without a large order book, the first few SMRs might not see significant savings over on-site builds. Additionally, transportation of large modules has logistics costs (e.g. special barges or rail transport for oversized components). SMRs trade the economies of scale of a large unit for economies of series – if demand scales up, factory production can dramatically lower CAPEX per unit.
Cost Overruns
Large reactor projects have a notorious record of budget overruns and schedule slips. The overruns drive up financing costs and ultimately the delivered cost of energy. SMRs are not immune to overruns. FOAK SMR projects can suffer similar overruns as large reactors, often due to design changes, supply chain hurdles, and the challenge of first-time construction. The hope is that after these FOAK experiences, SMRs will enter a stable learning curve. In contrast, each large reactor tends to be a bespoke project, limiting learning rate. Historical data suggests learning-by-doing in nuclear construction is possible only with standardization
2. Operational Expenditure (OPEX)
Fuel Cost per MWh
Nuclear fuel costs are a relatively small fraction of nuclear generation costs, but SMRs typically have slightly higher fuel cost per unit of electricity. The lower thermal efficiency (due to smaller core and simpler steam cycles) means more fuel is burned per MWh. Additionally, some advanced SMRs plan to use high-assay low-enriched uranium (HALEU, ~15–20% U-235) instead of standard 3–5% enrichment, which currently costs more due to limited supply. Overall, in the near-term, SMRs are expected to have a fuel cost penalty relative to large reactors. In the long term, if SMRs achieve higher burnup fuels or advanced fuel cycles, the gap could narrow.
Maintenance and Staffing
A major OPEX component is operation & maintenance (O&M), including staffing, routine maintenance, and security. Large nuclear plants benefit from scale but require a sizable workforce for operation. SMRs are designed to use smaller crews and more automation. A multi-module SMR plant could share staff across reactors, reducing per-MW staffing. SMR vendors claim even greater savings by using passive safety (needing fewer operators per reactor) and less complex equipment.
However, some costs do not scale down linearly. Regulatory requirements for staffing (e.g. minimum control room operators, security personnel per site) impose a baseline staffing level even for a smaller plant. For instance, an SMR might need a security force not much smaller than a large plant, due to similar security rules, which means more security cost per MW. Maintenance tasks like refueling or component replacement may be simpler for a small unit (with many systems modularized), but if you have many modules, total maintenance effort could be higher in aggregate.
Long-term O&M trends, Nuclear plants generally have rising O&M costs as they age (equipment replacements, regulatory upgrades). With SMRs, the hope is that modular designs simplify maintenance – for example, being able to swap out modular components or even entire reactor modules for off-site servicing (in some concepts). Additionally, having multiple smaller reactors gives operational flexibility. Maintenance can be staggered so the plant as a whole never fully shuts down (e.g. refueling one SMR module at a time while others operate). This improves capacity factor for the site, spreading fixed O&M over more MWh.
Security and Emergency Planning
Large reactors have stringent security and emergency planning requirements. SMRs are aiming for reduced emergency planning zones (perhaps site-boundary EPZ) and intrinsic safety that could justify smaller security per unit. If regulators accept a smaller EPZ for SMRs (as proposed for some designs with lower accident source term), this could shrink off-site emergency costs. Security staffing might also be optimized if the SMR is underground or has fewer access points to guard. However, until regulations change, an SMR site with several reactors will be treated similarly to a large plant for security – meaning the cost advantage might only materialize for very small or remote SMRs
Impact of Modular Design
Modular design can streamline operations. Training can be standardized for identical units, and spare parts inventories can be shared. There is a potential economy of multiples in operations. The 10th identical SMR a company operates will be handled more efficiently than the first, as operators gain experience and procedures are optimized (learning curve in O&M). Also, having many small reactors allows load-following by unit shuffling – turning off some modules during low demand – but this can hurt economics if not managed.
Overall, SMRs may achieve slightly lower O&M per MWh than large plants once matured, but in early years any savings might be offset by FOAK inefficiencies or conservative staffing to ensure safety.
3. Other Costs (Financing, Insurance, Waste, Decommissioning)
Financing Models and Interest During Construction
Because nuclear projects are capital-intensive and have long lead times, the cost of financing (interest) is a crucial part of total cost. Large reactors often require financing for billions over a decade. SMRs aim to reduce financing risk by shorter build times. A shorter construction schedule means interest accrues for fewer years. There’s also more flexibility in financing modular deployment. A utility could build a few SMR modules at a time, using revenue from the first modules to help finance later ones – a pay-as-you-go approach impossible with a single large unit. The incremental investment model could attract more investors or make use of cash flow recycling, mitigating the debt burden. Additionally, different financing models are being tried for SMRs as public-private partnerships, vendor financing, and fleet orders. SMR proponents argue that smaller projects are more “bite-sized” for investors, widening the pool of potential financiers compared to mega-projects that only governments or large utilities can fund.
However, if SMR costs per kW remain high, the financing challenge doesn’t disappear – multiple small reactors still need capital. In sum, SMRs improve financing feasibility via smaller scale and faster timelines, but strong government support (loan guarantees, tax credits, or assured offtake contracts) is still essential.
Insurance and Liability
All nuclear operators must cover accident liability, typically through a combination of operator insurance and government-backed indemnification. Insurance premiums for nuclear plants can be on the order of a few $ million per year (pooled insurance systems), which is a small fraction of overall costs. SMRs are unlikely to drastically change insurance costs initially – insurers will still require coverage for any potential accident. If an SMR has a smaller radioactive inventory and lower potential off-site impact, in theory the liability risk is lower, which could eventually lead to reduced insurance premiums or smaller required liability pools. Some SMR developers hope for proportional or reduced insurance requirements. Additionally, novel deployment models raise new liability questions. It might necessitate case-by-case arrangements and potentially higher insurance due to untested risk profiles.
For now, one should assume insurance and liability coverage costs are roughly comparable per reactor-site for SMRs and large plants, until frameworks adapt to account for reduced risk profiles.
Waste Management and Spent Fuel Disposal
Nuclear waste management is a long-term cost that must be planned for each reactor. Most countries require plant operators to pay into waste funds or handle interim spent fuel storage. SMRs will incur similar per-MWh of large reactor fees if instituted. One difference is spent fuel volume per energy output, because of lower efficiency, an SMR might discharge more spent fuel per MWh.
The handling of spent fuel might be simpler for SMRs if smaller fuel assemblies or modular removal is used. Some designs propose sending back entire fuel modules to a central facility, which could centralize and streamline waste handling (the operator of the power plant might not need a full on-site pool for long-term cooling if fuel is shipped out). This approach (used in research reactors or proposed for some microreactors) could shift waste management cost from the plant operator to a fuel service provider (for a fee). Ultimately, disposal in a geological repository will charge by volume or heat load, so more waste means higher cost.
Advanced SMRs that plan to reprocess or recycle fuel have different waste profiles, potentially less long-lived waste but more complex handling (mostly long-term possibilities).
For today’s analysis, assume each SMR bears comparable waste management costs per MWh as large reactors, with possibly a slight penalty for inefficiency. All operators must factor in spent fuel storage costs (on-site dry casks, etc.) until a final disposal solution is available. SMRs won’t magically eliminate these costs that they can only reduce them if they achieve higher burnup or implement take-back programs.
Decommissioning Costs and Long-Term Liabilities
Decommissioning a nuclear plant at end-of-life is a major cost item. Operators typically build a decommissioning fund (sinking fund from revenues set aside). An SMR (or an SMR plant with multiple modules) will also require a decommissioning fund.
The cost per reactor is lower, but the cost per MW may be higher because one doesn’t get scale economies in dismantling large components. However, there are potential advantages. Smaller reactors could be decommissioned faster and with less radiation dose to workers, and in some cases entire modules could be removed and shipped to a specialist facility for disassembly, rather than deconstructed in situ. This “return to vendor” model (if deployed) could simplify the process for the plant owner. There may be economies of performing the same task repeatedly at the same site.
The overall decommissioning cost for an SMR plant might be comparable in total to a large plant, just distributed over multiple small units. It’s likely each SMR unit’s decommissioning will be planned during design.
One also must consider long-term liabilities such as site contamination risk, spent fuel remaining on site, etc., which are similar challenges regardless of reactor size. Until real SMRs undergo decommissioning (the first won’t retire until mid-century), estimates are speculative. Regulators will not license SMRs without proof of a decommissioning funding plan, so investors should treat decommissioning as an unavoidable cost
4. Levelized Cost of Energy (LCOE)
LCOE combines all the above cost factors (CAPEX, OPEX, fuel, decommissioning, financing) into a single measure of overall cost per MWh generated, spread over a plant’s life. It’s a key metric for comparing economic competitiveness.
Current LCOE Estimates (2020s)
A 2018 Lazard analysis estimated large nuclear LCOE in the range $112–$189 per MWh. Similarly, OECD Nuclear Energy Agency predicted new SMRs’ LCOE is higher than large reactors. One study found SMR electricity costs 50–100% higher than current large reactors. For instance, South Australian Nuclear Fuel Cycle Royal Commission’s economic assessment calculated SMR LCOE ~22% above large reactor LCOE – about A$198–225/MWh for SMRs vs A$180–184/MWh for large PWR/BWR (equivalent to $140–159 vs $127–130 USD). NuScale’s own target of $65/MWh for its first plant was viewed as highly optimistic; independent analysis put it closer to $159/MWh. As noted, NuScale’s updated projection is $89/MWh with subsidies. For context, China’s large reactor LCOE is much lower – around $70/MWh – thanks to low construction costs and cheap capital (affordable financing and subsidies). But Russia’s floating SMR produces extremely expensive power ($200/MWh) with the high cost due to large staffing requirements, high fuel costs, and resources required to maintain the barge and coastal infrastructure.
In summary, present-day SMR LCOEs are generally not yet beating large reactor LCOEs; they often need subsidies or regulated pricing to be viable.
Future Cost Trends (2030, 2040, 2050)
SMR advocates project that costs will decline with volume and learning, improving LCOE in the medium to long term. Wood Mackenzie estimates the LCOE for a new SMR in the 2020s is “upwards of $120/MWh” in developed markets. By the 2030s, with technology learning, SMR LCOE could fall below $80/MWh given strong government support, innovation, and multiple deployments. For the long term, some industry targets are even lower. Rolls-Royce aims for ￡50/MWh ($65) for its SMR after several units, and the Energy Impact Center’s scenario suggested an SMR LCOE of only $36/MWh (though this is a highly optimistic outlier). A more conservative view from NREL’s Advanced Reactor analysis shows that by mid-century, with modest deployment, SMR capital costs might drop ~30–40%, while large reactor costs drop ~20%. It’s important to emphasize these are projections with uncertainty. Additionally, the learning rate (cost reduction per doubling of capacity) for SMRs is still uncertain.
The 2030s outlook: a few SMRs will be in operation, providing real cost data. Many countries plan pilot SMRs by early 2030s. If those succeed, by 2040s we could see dozens of SMRs built and a clearer cost trend that possibly standard designs benefiting from serial production.
By 2050, under an aggressive climate scenario, hundreds of SMRs globally could bring substantial economies of scale. Conversely, large reactors might remain niche (except in some countries), so their costs might stay high due to bespoke builds and supply chain attrition in the West.
In short, SMR LCOE is expected to trend downward 2025→2050, potentially reaching competitiveness with large reactors.
Sensitivity Analysis – Financing & Deployment Factors
LCOE for nuclear is highly sensitive to cost of capital (discount rate). If SMRs can tap low-interest loans (through government support or climate finance), their LCOE drops significantly. Thus, policy measures like loan guarantees, regulated asset base (RAB) models, or long-term power contracts improve economics.
Another sensitivity is capacity factor: nuclear plants are best run at high capacity factors (~90%). If an SMR is used in load-following mode and achieves only 50% capacity factor, its LCOE could double (since capital costs are fixed). IEEFA analysis of NuScale shows if capacity factor dropped from 95% to 65%, the power cost would rise from $90 to $129/MWh, and at 45% it would skyrocket to ~$183/MWh (SMRs need to run near full output).
Deployment scale is another factor: if only a handful of SMRs are built, there’s little cost improvement; if dozens are ordered, then economies of series kick in.
Also, costs will come down faster if the industry focuses on a few proven SMR designs worldwide, rather than 50 different boutique designs each built once (standardization vs. multiple designs). Harmonizing designs and pooling orders (like buying 10 identical SMRs) would spread development costs and enable bulk procurement of components.
5. Additional Cost Metrics and Decision Factors
Cost per MWh and Payback Period
Beyond LCOE, investors look at simple payback and cash flows. A large reactor typically has very long payback – often >15–20 years – because the initial investment is so high and the first revenue comes only when the plant is operational (which could be 7–10 years after breaking ground). SMRs could improve payback dynamics by phasing investment and revenue. Additionally, smaller units reduce the risk of never reaching operation (as happened with some large projects that were abandoned mid-way, resulting in total loss). Thus, from an investment perspective, SMRs offer modular return on investment – each unit is an incremental investment with its own returns, potentially more palatable than one enormous bet.
However, one must also consider that many SMR business cases assume follow-on units; a first unit alone could have a very long payback if it’s burdened with FOAK costs. For large reactors in regulated markets, payback is often managed via rate recovery (utilities start recouping costs during construction in some cases, or immediately upon operation via rate base increases). In competitive markets, payback depends on wholesale electricity prices. If prices are $50/MWh and your nuclear LCOE is $100/MWh, you’ll never pay back without subsidies or carbon pricing. So market conditions and policies (like carbon credits, clean energy payments) strongly affect payback.
Return on Investment (ROI)
ROI for nuclear projects tends to be relatively low and long-term, which is why private investors are cautious. Large reactor projects often require government guarantees to achieve even single-digit percentage returns, given the risks. But initial investors in new nuclear face a long wait. The competitive landscape matters if SMRs can fill niches that others can’t (remote areas, replacing coal one chunk at a time, providing high-temperature heat), they might command premium revenue, boosting ROI.
However, if they are competing in the general electricity market, their ROI will be challenged by cheap renewables. Generally, nuclear ROI is more about strategic investment (for reliability or climate goals) than high profits. We can compare qualitatively that a wind or solar farm might have an ROI in the teens (with favorable policies), whereas a nuclear plant might be single-digit.
Cost Competitiveness vs Other Energy Sources
In today’s terms, new nuclear (large or small) struggles to compete on pure cost per MWh with gas or renewables in many regions. However, comparing LCOE alone can be misleading for dispatchable vs variable sources. Wind and solar require backup or storage for equivalent reliability, which adds system costs not captured in their standalone LCOE. Nuclear provides steady baseload (or flexible output) and has a high capacity factor. From a system cost perspective, a certain amount of firm capacity (like nuclear or hydro or gas) is needed to maintain reliability in a renewables-heavy grid. In essence, SMRs need to leverage their niche strengths to justify their costs.
Nonetheless, from an investor lens, if costs don’t come down, SMRs will remain a tough sell purely against cheap renewables or gas.
Scalability and Learning Curve (Cost Reduction Potential)
One of the strongest arguments for SMRs is the learning curve effect – as more units are built, costs per unit should drop, following a learning rate observed in other industries (renewables, shipbuilding, aircraft). Traditional large nuclear never achieved a consistent learning curve globally due to stop-start build programs and design changes. SMRs, if ordered in dozens, could finally unlock a nuclear learning curve. The scalability of production also means output can ramp up once the first hurdles are passed. If a factory can produce 5 SMR modules a year, a second identical factory could double that, etc.
By contrast, building more large reactors faster is constrained by the limited number of companies and sites able to handle such mega-projects. Thus, if demand materializes, SMRs could scale more like a product than a project, with modular deployment across many sites. This could lead to mass production economies by 2040s: lower labor hours per unit, bulk material purchases, standardized training and O&M, and perhaps a global supply chain that drives down component costs.
The key is whether the nuclear sector can truly standardize – historically, changes are often forced by regulatory updates or technology tweaks. But if an SMR design can freeze its design for a fleet of 50 units, the savings could be immense. Therefore, investors see a potential upside in SMRs: if you invest in an early SMR company and it manages to capture a large market with its standard design, later units could have very high profit margins due to manufacturing efficiency – a payoff similar to early investors in a successful tech product. Of course, this comes with high risk if the market doesn’t scale or if a design gets leapfrogged by a competitor before reaching economy of scale.
6. Risk Analysis
Supply Chain Risks
Nuclear-grade components and materials are specialized. Large reactors require huge forgings (e.g. reactor pressure vessels) which only a few suppliers worldwide can make. Any bottleneck or quality issue there can delay a project.
SMRs use smaller components that more suppliers could fabricate, potentially easing this bottleneck. However, SMRs introduce new supply chain needs. Modular fabrication facilities must be established, and skilled labor must be trained for factory assembly of reactor modules. If demand is uncertain, suppliers may be hesitant to invest in capacity, leading to potential shortages or higher prices for SMR components initially.
Another specific risk is fuel supply for advanced SMRs. Several SMR designs require HALEU fuel or TRISO fuel, which currently have very limited commercial production. Traditional 5% enriched fuel supply is robust globally (many suppliers), so near-term LWR-based SMRs are safe on fuel, but novel fuel SMRs carry this risk.
Additionally, quality control in the supply chain is crucial: a substandard component can cause major delays. SMR factories will need nuclear-grade QA on all modules – a challenge if trying to produce at scale. If supply chain issues cause delays, that erodes the very cost advantage SMRs seek (time savings). Thus, investors will watch whether a reliable supply network (for reactor vessels, steam generators, control systems, etc.) forms around SMRs.
Financing Risks
Nuclear projects require massive upfront capital and long horizons, which can scare off investors without guarantees. A key risk is that investors or lenders demand a high risk premium, making the cost of capital so high that the project becomes uneconomical. the risk remains if an SMR project runs over-budget or if electricity market prices don’t pan out, who absorbs the loss? In many cases, taxpayers or taxpayers have shouldered cost overruns. For wider deployment, financiers will look for risk reduction mechanisms: e.g. DOE loan guarantees, insurance against regulatory delays, or construction contingency funds.
Another risk is market risk is if power demand or prices drop in the future (say, due to economic downturn or a flood of cheap renewables), the revenue for a nuclear plant might fall short, affecting its ability to service debt.
There’s also currency and country risk for international projects. Emerging market adopters of SMRs might face currency devaluation or political instability that could derail financing.
In summary, while SMRs offer a chance for more flexible financing, they won’t be entirely immune to the classic nuclear finance challenge: high capital risk. The willingness of private capital to engage will depend on early success cases proving on-time, on-budget delivery.
Regulatory and Policy Risks
Nuclear energy is heavily influenced by government policy, which can change. A major risk is regulatory approval delays. Each country’s regulator might have different requirements; lack of international regulatory harmonization means a design could be certified in one country but face years of review elsewhere. If one SMR design becomes the first to clear several major regulators, that will reduce uncertainty for others.
Political risk is also significant: public acceptance of nuclear varies, and a change in government can lead to policy reversals (as seen in some countries phasing out nuclear). An SMR program could be cancelled or delayed if public opinion turns or if a promised subsidy is withdrawn. On the positive side, many governments are now supportive of SMRs on paper, passing enabling legislation and funding R&D. But these supportive policies need to endure for decades to see projects through.
Permitting and local opposition could pose risks too. While SMRs are smaller, they still are nuclear facilities requiring local consent. Delays in local approvals or legal challenges (common for large reactors) could also affect SMRs, albeit one selling point is to use existing nuclear or industrial sites to minimize opposition.
In sum, regulatory risk remains a big unknown for first movers: until a few SMRs are fully licensed and operating, there will be caution. Political support (or lack thereof) for waste repository solutions also indirectly affects SMR success, since a country unwilling to handle nuclear waste will be hesitant to deploy reactors big or small.
Market and Demand Risks
The business case for SMRs assumes that there will be customers ready to buy them or the power they produce. A risk is overestimation of market size. If the anticipated SMR “boom” doesn’t materialize because alternatives (e.g. renewables + storage, or improved large reactor designs, or even fusion in decades ahead) take the spotlight, then manufacturers won’t achieve the volume needed to lower costs.
Market risk also includes future electricity price uncertainty: an investor in an SMR needs confidence that there will be a long-term revenue stream. This often means either regulated markets or long contracts. A merchant SMR selling into an open market would face significant price risk.
We should mention customer adoption. Utilities tend to be conservative – getting them to embrace a new reactor design will require proof of reliability and cost. The cancellation of some planned SMR projects shows that customers can pull back if economics don’t justify it.
On the flip side, if global decarbonization efforts intensify and countries put a premium on firm, clean power, the demand for SMRs could outstrip supply mid-century.
Key takeaways
Disclaimer:
The thoughts shared in this my personal LinkedIn article are entirely my own. They don't reflect the views of my current employer, and I haven't included any proprietary or confidential information from my company here. It's just me reflecting, and if anything seems similar to my company’s work, it's purely a coincidence. And, I'm human - mistakes happen. If you have any questions or want to dive deeper into these ideas, feel free to discuss it with appropriate discussion.