Too Cheap to Meter Too Slow to Build
The promise of small modular reactors is plausible, well-funded, and still unproven.
The town of Wylfa sits on the northern tip of Anglesey, a Welsh island connected to the mainland by two bridges and a reputation for nuclear power. The original Wylfa power station opened in 1971 and ran for more than four decades, supplying electricity to households and businesses across the region. When it closed in 2015, the local economy lost something difficult to replace — not just jobs, but a particular kind of work: skilled, well-paid, long-term.
Plans for a new nuclear plant at Wylfa have circled since before the original station was switched off — Horizon Nuclear Power, then Hitachi, then a succession of other investors, none of whom made it past the preliminary stages. In June 2025, something different: the UK government selected Rolls-Royce SMR as its preferred partner to build the country’s first small modular reactors. By April 2026, a commercial contract had been signed and Wylfa confirmed as the site for up to three SMR units. If the timeline holds, the grid connection comes in the mid-2030s.
That “if” is doing a lot of work. The nuclear industry has a long and painful history of timelines that do not hold, costs that do not stay contained, and promises that outlast the political will required to keep them. The people of Wylfa — and the communities near every proposed SMR site — have reason to ask hard questions before deciding whether this time is genuinely different.
The honest answer is: it might be. And that is a more interesting answer than either the enthusiasts or the sceptics want to give.
Why nuclear has a second chance
To understand what small modular reactors are trying to solve, it helps to understand what broke the large ones.
Nuclear power has three advantages over almost every other energy source that are not marginal. It produces virtually no carbon emissions during operation. It generates electricity around the clock, regardless of whether the wind blows or the sun shines. And its energy density is extraordinary — a single kilogram of uranium fuel contains roughly two million times the energy of a kilogram of coal. These are the reasons nuclear provides about a tenth of the world’s electricity despite decades of political difficulty and public mistrust.
The problem is construction. Large nuclear plants take fifteen years to build in Western countries. The most recent to complete in the United States, the Vogtle plant in Georgia, came in seven years late and roughly $22 billion over its original $14 billion budget. The Hinkley Point C plant currently under construction in Somerset — the UK’s first new nuclear station in a generation — has seen its estimated cost rise from £18 billion when contracts were signed in 2016 to £35 billion in equivalent terms, approaching £49 billion at current prices. These are not outliers. They are the pattern.
The causes are structural: one-off bespoke designs that require learning everything from scratch on each new site; an atrophied supply chain that lost its skills during the decades-long pause in construction; and regulatory frameworks designed around 1970s technology that treat every new application as if nothing has been learned since. The result is that nuclear power, which ought to be the most reliable weapon in the decarbonisation arsenal, has spent thirty years being too expensive and too slow to be deployable at scale.
Nuclear’s advantages are extraordinary. Its construction record in the West is, on the evidence, one of the most expensive and difficult in modern engineering.
Small modular reactors are an attempt to escape that trap by changing the nature of the construction problem. Rather than building a bespoke cathedral on each site, the idea is to build a reactor in a factory — standardised, modular, replicable — and ship it to where it’s needed. A conventional reactor might generate 1,000 to 1,600 megawatts of electricity. An SMR typically produces between 50 and 470 megawatts per unit. You can add more units as demand grows, and the first unit can begin generating revenue while subsequent ones are still being assembled.
The factory model matters for cost. Nuclear’s expense problem stems partly from the fact that large reactors have never been built in sufficient numbers to drive down costs through learning and repetition. The steel and concrete and specialist components are expensive partly because the people and processes that handle them do so rarely. A factory model — producing the same design, repeatedly, under controlled conditions — is how aviation and automotive manufacturing drove down costs over decades. It is, at least in theory, how nuclear might do the same.
Where the theory meets the track record
There is a problem with the factory model, and it is not a small one: it has not yet produced a nuclear reactor on time or on budget, anywhere in the world.
As of early 2026, three SMRs have been built globally — one in China, two in Russia — plus one still under construction in Argentina. None were built on time. Russia’s floating plant, the Akademik Lomonosov, began construction in 2007 and reached operation only in 2019, at roughly three times its original budget. Argentina’s CAREM reactor broke ground in 2014 with a planned completion of 2017 and was still incomplete as of 2026, construction halted twice by funding crises, with costs significantly above initial estimates. China’s pebble-bed reactor at Shidaowan took approximately eleven years from construction start to commercial operation. These are first-of-a-kind projects, and some overrun is expected when building genuinely new technology. But the pattern of delay holds across different countries, different designs, and different political contexts.
The most instructive cautionary tale is NuScale Power. The American company was the first to receive US Nuclear Regulatory Commission design certification for an SMR, achieving that milestone in 2022 — a genuine achievement after years of rigorous review. NuScale appeared to be on the cusp of proving that the new model worked. Then, in 2023, its flagship project — a planned plant in Idaho — was cancelled. The project began with an estimated cost of around $3.6 billion for a 720-megawatt plant. Before cancellation, with the design rescaled and costs revised, that figure had risen to $9.3 billion. The customers, a consortium of public utilities, could not make the economics work. NuScale has since received regulatory approval for an uprated 77-megawatt design and is pursuing new projects, but the Idaho cancellation cast a long shadow over the industry’s cost claims.
Global SMR landscape in 2026
Operational
China (HTR-PM pebble-bed),
Russia (floating Akademik Lomonosov ×2)
Under construction
Argentina (CAREM; construction began 2014; multiple halts; still incomplete)
UK Rolls-Royce SMR contracted April 2026; Wylfa confirmed site; up to 3 units; grid target mid-2030s; £2.5bn govt commitment
USA NuScale NRC-approved (77 MWe); TVA + ENTRA1 Energy 6-GW programme announced; TerraPower (Natrium) and X-energy also advancing with DOE backing
Tech sector
Amazon ($500m+, X-energy); Google (Kairos Power, 500 MW by 2035; first unit targeting 2030)
Czech Republic Rolls-Royce SMR selected by CEZ; up to 3 GW planned
Summary
127 SMR designs globally; 51 in pre-licensing/licensing; only ~3 operational (OECD NEA, 2025)
The sceptical case against SMRs is not that the physics is wrong — it is that the economics may be structurally difficult. Smaller reactors lose the cost advantages that come with scale: building one 1,000-megawatt plant is cheaper per unit of output than building four 250-megawatt ones, because you only need one set of site preparation, one control room, one regulatory approval process. A 2025 study by Kim and Macfarlane — the latter a former chair of the US Nuclear Regulatory Commission — found that most cost projections for SMRs assumed high-volume serial manufacturing that has never materialised in nuclear, and that more realistic estimates placed the cost of electricity well above $100 per megawatt-hour, in some cases significantly more. For context, onshore wind in the UK currently costs around £40–50 per megawatt-hour.
There is also the waste question. Some analyses suggest that SMRs could produce greater volumes of radioactive waste per unit of energy generated than large conventional reactors — a consequence of their smaller scale and, in some advanced designs, the use of different fuel cycles. This is contested by the industry, but it is not a settled matter, and it matters to any community being asked to accept a reactor nearby.
None of this means SMRs cannot work. It means the confident promises being made — by governments, by technology companies, by investors — are running ahead of the evidence. The International Energy Agency’s scenarios have commercial SMR deployment beginning around 2030 — meaningful on paper, but, in the context of a global electricity system measured in thousands of gigawatts, still a rounding error in the near term.
What hangs on getting this right
The stakes are high enough that the question of whether SMRs work is not abstract.
Renewable energy — wind and solar — has become extraordinarily cheap and is now the cheapest source of new electricity generation in most of the world. But it is intermittent. The sun does not always shine; the wind does not always blow. Storing renewable energy at grid scale, for days or weeks at a time, remains an unsolved problem at the volumes the transition requires. This is sometimes called the baseload problem: the need for power sources that generate electricity continuously, regardless of weather, to underpin a grid that cannot afford to go dark.
Nuclear is one of the very few zero-carbon technologies that addresses the baseload problem directly. It generates power around the clock, in any weather, for decades. Hydropower does the same, but its geography is fixed. Large-scale geothermal is similarly constrained. If the world is serious about reaching net zero while keeping the lights on reliably, the honest arithmetic suggests that some form of nuclear will be part of the answer — and that SMRs, if they can deliver on their promise, could be a significant part.
The tech sector has reached a version of this conclusion independently. Amazon has committed over $500 million to SMR development. Google has partnered with Kairos Power to bring 500 megawatts of SMR capacity online by 2035, with the first unit targeting 2030, to power its data centres. Microsoft has signed agreements in the same space. The irony is pointed: AI’s insatiable demand for always-on power is one of the forces driving investment in nuclear energy’s revival.
The geopolitical dimension adds a further layer. Russia and China have both invested heavily in nuclear technology as a strategic export — Russia through Rosatom, which has construction contracts in over twenty countries; China through its state-owned enterprises. The UK-US Atlantic Partnership for Advanced Nuclear Energy, signed in September 2025, explicitly includes joint safety assessments and a shared commitment to eliminate dependence on Russian nuclear fuel by 2028. The race to build credible Western SMR supply chains is not purely a climate story. It is also a question of whose technology runs the world’s energy infrastructure in the decades ahead.
Against this backdrop, the communities living near proposed SMR sites find themselves in a familiar position — asked to accept proximity to technology that carries real and perceived risks, on the promise of jobs and cheap energy, with a track record that gives them every reason to ask for guarantees that are difficult to give. Research published in Energy Research & Social Science in 2025 put it plainly: the burdens and benefits of nuclear transitions are rarely distributed fairly, and existing frameworks to address this are inadequate. Wylfa’s community has been promised a new nuclear future before. The question is whether this version of the promise is better-designed, or just better-marketed.
What the technology still owes
At its best, an SMR helps communities, grids and industries access clean, reliable power that they cannot get from renewables alone. The UK government’s £2.5 billion commitment, the US government’s backing of multiple designs, and the private investment pouring in from the tech sector all reflect a real assessment that this technology could help solve a real problem. That is not spin. It is the outcome of serious people looking hard at the arithmetic of net zero.
Whether it adds something genuinely new is a sharper question. The factory model is the central claim — that modular construction changes the economics of nuclear in the same way that modular construction changed the economics of consumer electronics and automotive manufacturing. The honest answer is that this claim is plausible but unproven at commercial scale. The first Western SMR to be built on time and on budget will be the most important data point in energy policy this decade.
The question most consistently overlooked is whether this technology responds to the specific circumstances of the communities and contexts it serves — to their energy needs, their employment profiles, their risk tolerances, their histories with the industry. The SMR industry’s governance frameworks, on the evidence of current proposals, are not yet doing this. Regulatory processes are improving — the US ADVANCE Act of 2024 directed the NRC to streamline its approach, and by the end of 2025 it had met 30 of its 36 planned deliverables. But the frameworks for distributing the benefits and managing the disruptions of deployment — particularly for the communities most directly affected — remain underdeveloped.
The people of Wylfa have skills, knowledge and a relationship with nuclear power that spans generations. They are not a passive audience for a technology decision made elsewhere. Whether the SMR programme being planned for their island treats them as participants or subjects will say something important about whether the industry has learned anything from its history.
My opinion
I have worked inside the nuclear industry and I know what it delivers when it works. I also know what it costs when it goes wrong — not just financially, but in the lives of the communities closest to it. That makes it impossible for me to land cleanly on either side of this argument.
The NIMBY response to nuclear is rational, not irrational. If you live near a proposed site, you are being asked to accept proximity to technology whose consequences, in the worst case, outlast your lifetime and your children’s. That is a different category of ask than living near a wind farm.
But the power needs are also real. The grid cannot be decarbonised on intermittent sources alone. If small modular reactors can be built reliably, at reasonable cost, with genuine engagement of the communities they affect — that is worth pursuing. The question is whether the industry can meet that bar. It has not yet, but that is not the same as saying it cannot.
We also do not yet know if there will be any long-term consequences of renewable generation - what happens to wind patterns that are distributed by wind farms? what happens to ocean currents when hydro wave generation removes some of the kinetic energy; what happens to the earth’s core temperature when sunlight is blocked from hitting the ground by solar, or by heat extraction? We’ve made serious investments in all these technologies without the evidence of consequences - we know the consequences for nuclear.
What I am certain of is that complacency is the wrong response in either direction. The technology needs to prove itself commercially. The governance needs to prove itself ethically. Those are not the same challenge, and the industry tends to treat the first as if it solves the second.
The questions Wylfa should ask
The UK government has pledged £2.5 billion and named a preferred supplier. But the first SMR is a decade away from the grid. How much confidence is there that the political will, the budget and the regulatory process will all hold for ten years?
If renewables are now the cheapest source of new power, and storage technology is improving rapidly, is it possible that SMRs solve a problem that will largely have resolved itself by the time they arrive? Or is the baseload need real and irreducible?
The tech companies investing in SMRs are doing so to power data centres. If artificial intelligence is driving both the energy demand that makes nuclear necessary and the investment that makes nuclear viable, what does that say about who the energy system is being built for?
If you lived in Wylfa — or near any proposed SMR site — what would you need to know, and who would you need to hear from, before deciding whether to support the project?
The phrase “too cheap to meter” was coined in 1954 by Lewis Strauss, the chairman of the US Atomic Energy Commission, describing what nuclear power might one day become. It has since become one of the most famous over-promises in the history of technology. Small modular reactors are not making that promise. They are making a more modest one: that nuclear power can be built faster and cheaper than it has been. That is a lower bar. It is also one the industry has not yet cleared.
Whether it clears it this decade matters more than almost anything else in energy policy.
Sources & References
Wylfa power station: UK Government, GOV.UK: “Wylfa closes after almost 45 years” — https://www.gov.uk/government/news/wylfa-closes-after-almost-45-years
ANS Nuclear Newswire, 10 June 2025: “UK’s own Rolls-Royce wins SMR competition” — https://www.ans.org/news/2025-06-10/article-7102/uks-own-rollsroyce-wins-smr-competition/
NucNet: “UK Picks Rolls-Royce For Domestic Small Modular Reactor Rollout” — https://www.nucnet.org/news/uk-picks-rolls-royce-for-domestic-small-modular-reactor-rollout-6-2-2025
New Civil Engineer, 13 April 2026: “Rolls-Royce SMR secures Wylfa contract and £599M government loan” — https://www.newcivilengineer.com/latest/rolls-royce-smr-secures-wylfa-contract-and-599m-government-loan-13-04-2026/
Neutron Bytes, 19 April 2026: “UK & Rolls Royce Sign Deal for Three SMRs at Wylfa” — https://neutronbytes.com/2026/04/19/uk-rolls-royce-sign-deal-for-three-smrs-at-wylfa/
IEA, Nuclear Power page (nuclear generated 9% of global electricity in 2024): https://www.iea.org/energy-system/electricity/nuclear-power
IEA, Global Energy Review 2025: https://www.iea.org/reports/global-energy-review-2025/electricity
Utility Dive, 2026: “After 2 years, ratepayer pain and political fallout from Georgia’s nuclear plant Vogtle” — https://www.utilitydive.com/news/after-2-years-ratepayer-pain-political-fallout-georgia-nuclear-vogtle/817792/
The Current GA, May 2026: “Two Years After Completion, Plant Vogtle Still Looms Over the Nuclear Debate” — https://thecurrentga.org/2026/05/12/two-years-after-completion-plant-vogtle-still-looms-over-the-nuclear-debate/
EnergyTransition.org, April 2026: “The billion-dollar boondoggle: how Vogtle became the US’s monument to nuclear folly” — https://energytransition.org/2026/04/the-billion-dollar-boondoggle-how-vogtle-became-the-uss-monument-to-nuclear-folly/
New Civil Engineer, February 2026: “Hinkley Point C’s cost climbs to £35bn with confirmation Unit 1 will power up in 2030” — https://www.newcivilengineer.com/latest/hinkley-point-cs-cost-climbs-to-35bn-with-confirmation-unit-1-will-power-up-in-2030-20-02-2026/
Construction Wave, February 2026: “Hinkley Point C costs approach £49bn as project faces M&E delays” — https://constructionwave.co.uk/2026/02/23/hinkley-point-c-costs-approach-49bn-as-project-faces-me-delays/
US Department of Energy: “NRC Certifies First U.S. Small Modular Reactor Design” — https://www.energy.gov/ne/articles/nrc-certifies-first-us-small-modular-reactor-design
Utility Dive, 2023: “NuScale, UAMPS terminate small modular reactor project in Idaho” — https://www.utilitydive.com/news/nuscale-uamps-terminate-small-modular-nuclear-reactor-smr-project-idaho/699281/
Boise State Public Radio, November 2023: “NuScale nuclear reactor project in Idaho canceled” — https://www.boisestatepublicradio.org/news/2023-11-10/idaho-small-nuclear-reactor-project-canceled
E&E News: “NuScale cancels first-of-a-kind nuclear project as costs surge” — https://www.eenews.net/articles/nuscale-cancels-first-of-a-kind-nuclear-project-as-costs-surge/
US Department of Energy: “NRC Approves NuScale Power’s Uprated Small Modular Reactor Design” — https://www.energy.gov/ne/articles/nrc-approves-nuscale-powers-uprated-small-modular-reactor-design
World Nuclear News: “China’s demonstration HTR-PM enters commercial operation” — https://www.world-nuclear-news.org/Articles/Chinese-HTR-PM-Demo-begins-commercial-operation
NextBigFuture, December 2023: “China’s Pebble Bed Reactor Finally Starts Commercial Operation” — https://www.nextbigfuture.com/2023/12/chinas-pebble-bed-reactor-finally-starts-commercial-operation.html
Wikipedia: Akademik Lomonosov —https://en.wikipedia.org/wiki/Akademik_Lomonosov
Bellona.org, 2015: “New documents show cost of Russian floating nuclear power plant skyrockets” — https://bellona.org/news/nuclear-issues/2015-05-new-documents-show-cost-russian-nuclear-power-plant-skyrockets
Power Magazine: “Russia Sees Floating Nuclear Power Plant Costs Balloon” — https://www.powermag.com/russia-sees-floating-power-plant-costs-balloon/
Wikipedia: CAREM — https://en.wikipedia.org/wiki/CAREM
Nuclear Engineering International: “Argentina’s CAREM-25 SMR faces setbacks” — https://www.neimagazine.com/news/argentinas-carem-25-smr-faces-setbacks/
Buenos Aires Herald: “Construction of first Argentine-made nuclear power reactor halted amid layoffs” — https://buenosairesherald.com/business/construction-of-first-argentine-made-nuclear-reactor-halted-amid-layoffs
ScienceDirect, Progress in Nuclear Energy, 2025: “Challenges of small modular reactors: A comprehensive exploration of economic and waste uncertainties associated with U.S. small modular reactor designs” — https://www.sciencedirect.com/science/article/abs/pii/S0149197025003877
CleanTechnica summary, September 2025: “Small Modular Reactors and the Big Questions of Cost & Waste” — https://cleantechnica.com/2025/09/10/small-modular-reactors-and-the-big-questions-of-cost-waste/
UK Government / BEIS, 2024: “Onshore Wind and Solar PV: Cost of Electricity Report Update 2024” — https://assets.publishing.service.gov.uk/media/68ba91f411b4ded2da19fe92/onshore-wind-and-solar-pv-cost-electricity-report-update-2024.pdf
IEA: “The Path to a New Era for Nuclear Energy” — https://www.iea.org/reports/the-path-to-a-new-era-for-nuclear-energy
IEA data chart: “Small modular reactor global installed capacity by scenario and case, 2025–2050” — https://www.iea.org/data-and-statistics/charts/small-modular-reactor-global-installed-capacity-by-scenario-and-case-2025-2050
ESG Today: “Rolls-Royce Signs Deal with UK to for First Fleet of Small Modular Nuclear Reactors” (references Amazon/X-energy investment) — https://www.esgtoday.com/rolls-royce-signs-deal-with-uk-to-deliver-fleet-of-small-modular-nuclear-reactors/
Google/Kairos Power: “Google and Kairos Power partner to develop nuclear energy” — search Google’s official press release at blog.google for the specific announcement.
GOV.UK: “Golden age of nuclear delivers UK-US deal on energy security” — https://www.gov.uk/government/news/golden-age-of-nuclear-delivers-uk-us-deal-on-energy-security
Al Jazeera, 18 September 2025: “US and UK sign major nuclear power deal: What does it include?” — https://www.aljazeera.com/news/2025/9/18/us-and-uk-sign-major-nuclear-power-deal-what-does-it-include
Nuclear Innovation Alliance: “Regulatory Implementation Summary: NRC Progress Under the ADVANCE Act”(December 2025) — https://nuclearinnovationalliance.org/regulatory-implementation-summary-nrc-progress-under-advance-act
NRC ADVANCE Act dashboard: https://www.nrc.gov/about-nrc/governing-laws/advance-act/dashboard
NRC History: “Too Cheap to Meter: A History of the Phrase” — https://www.nrc.gov/reading-rm/basic-ref/students/history-101/too-cheap-to-meter
Wikipedia: Too cheap to meter — https://en.wikipedia.org/wiki/Too_cheap_to_meter
Neil Catton is the author of The Next Evolution, The Cognitive Crucible and The Shadow System - available on Amazon, and writes at the intersection of technology, ethics, and human purpose.