Global Race$40 billion. That’s what the United States and Japan committed in March 2026 to put GE Vernova Hitachi’s BWRX-300 reactors in Tennessee and Alabama. ₹20,000 crore. That’s India’s line item for five indigenous small modular reactors by 2033. SEK220 billion. That’s Sweden’s twelve-year lending framework for new nuclear, most of it aimed at SMRs. None of these numbers existed three years ago. All of them landed within the last twelve months. Here’s what’s actually driving the money.

AI data centers broke the old demand curve

Utilities used to plan for flat or slowly rising electricity demand. That model is dead, and hyperscalers killed it. The clearest evidence sits in Tennessee and Alabama, where the U.S. Department of Commerce confirmed a $40 billion commitment to deploy GE Vernova Hitachi’s 300-megawatt BWRX-300 design, totaling 3 gigawatts of new capacity, explicitly to stabilize power for the region’s growing tech and manufacturing corridor. It’s part of a broader $550 billion U.S.-Japan trade package, which tells you something about how central power generation has become to industrial policy, not just energy policy.

Tech companies aren’t waiting for governments to solve this either. Kairos Power’s Hermes 2 demonstration plant in Oak Ridge, built under a landmark deal with Google, is contracted to deliver 500 megawatts to the Tennessee Valley Authority grid by 2035 to help power Google’s regional data centers. A few data points on why this matters:

• U.S. power plant developers plan to add roughly 86 gigawatts of new utility-scale capacity to the grid in 2026 alone

• January 2026 saw a record $25.2 billion in U.S. data center construction spending in a single month

• Amazon has separately backed X-energy’s four-unit Xe-100 project with utility Energy Northwest in Washington state

None of that capacity comes from SMRs yet. That’s the point. Data center operators are locking in reactor capacity years before a single unit produces power, because the alternative, building enough gas peaker plants and hoping the grid holds, is starting to look like the riskier bet.

Governments are done being hostage to fuel prices

Sweden offers the cleanest example of energy security driving SMR money. Vattenfall’s project company, Videberg Kraft, selected Rolls-Royce SMR over GE Vernova in June 2026 to supply three reactors at the Värö Peninsula site, a deal worth several billion pounds, according to Euronews’ reporting on the selection. Prime Minister Ulf Kristersson announced the state would take a 60% stake in the project. That followed a national budget proposal earlier in the year that set up a SEK220 billion (roughly $23.4 billion) lending framework for new nuclear construction over twelve years, one of the largest single-country nuclear financing commitments anywhere right now.

The logic is blunt: Sweden’s electricity demand is projected to roughly double over the next two decades, and the government has said explicitly it wants fossil-free power at a stable price rather than exposure to the kind of gas price swings that hit Europe hard after 2022. A few reasons this framing keeps showing up across capitals:

• Gas price volatility punishes countries that lean on imported fuel for baseload power

• Nuclear fuel can be stockpiled for years, unlike a pipeline that can be shut off

• SMR capacity sited domestically reduces exposure to any single supplier or trade route

• Reliable, weather-independent output backstops variable renewables during low-wind, low-sun stretches

This isn’t unique to Sweden. It’s the same instinct behind the European Commission’s SMR strategy, published the same month, and it’s the same instinct pushing India’s Department of Atomic Energy to build reactors domestically rather than import large-scale plants it can’t fully control the supply chain for.

Net-zero targets need power that doesn’t quit when the wind does

The European Commission’s March 2026 strategy to get Europe’s first SMRs online by the early 2030s came bundled with a sobering number: its Nuclear Illustrative Programme estimates the EU needs roughly €241 billion in nuclear investment by 2050 to hit its own targets, covering both existing reactor lifetime extensions and new construction including SMRs. Energy Commissioner Dan Jørgensen framed it as a competitiveness issue as much as a climate one, tying SMR deployment to industrial security alongside decarbonization.

The reasoning shows up in capacity forecasts too. Under current policy, the International Energy Agency projects global SMR capacity will reach 40 gigawatts by 2050. With streamlined regulation and a fivefold jump in SMR investment this decade, that figure could hit 120 gigawatts, according to the Stanford Understand Energy program’s summary of IEA modeling. That’s the gap countries are racing to close:

• Solar and wind are now the cheapest new generation in most markets, but they’re intermittent by nature

• Grid-scale battery storage helps on the margins, not for weeks of low-wind winter stretches

• Firm, dispatchable, carbon-free generation is the missing piece, and large reactors take a decade-plus and tens of billions to build

• SMRs promise a faster, more modular path to that same firm capacity, if the economics hold up

Whether SMRs actually deliver on cost per megawatt is still an open question. What isn’t open to debate: without some new firm low-carbon source, most 2050 net-zero pledges don’t survive contact with a still, cloudy January.

Nuclear is now industrial policy, not just energy policy

Look closely at who’s writing the checks and you’ll notice they’re often not utilities. In Sweden, an industrial consortium called Industrikraft, formed by ABB, Volvo Group, SSAB, Saab, Alfa Laval, and five other manufacturers, took a 20% stake in Videberg Kraft and put up SEK400 million (about $42 million) specifically to keep Swedish heavy industry inside the supply chain for its own future power source. That’s steelmakers and carmakers betting on nuclear because they don’t want to depend on someone else’s grid for the electricity their factories will need.

India is running an even more explicit version of this play. Its ₹20,000 crore ($2.5 billion) Nuclear Energy Mission funds three homegrown designs, the 200-megawatt Bharat Small Modular Reactor, the 55-megawatt SMR-55, and a hydrogen-focused high-temperature gas reactor, with lead units slated for Tarapur and Andhra Pradesh’s Vizag campus. The explicit goal isn’t just energy security. It’s reducing import dependence and building export capacity in a market currently dominated by the U.S., China, and Russia. If you want to track how these industrial bets connect to actual project timelines and financing terms, SMRbrief Pro keeps the underlying deal data structured and searchable rather than scattered across press releases.

Manufacturing jobs, supply chain control, and export ambitions rarely show up in the climate-focused SMR coverage. They probably should.

The modular pitch is finally getting real-world proof points

Every argument for SMRs rests on one unproven claim: that factory-built, repeatable reactor designs will be cheaper and faster than one-off mega-projects. 2026 produced the first hard evidence either way. TerraPower’s Natrium reactor in Kemmerer, Wyoming received its NRC construction permit on March 4, 2026, the agency’s first approval for any commercial reactor in nearly a decade and the first for a non-light-water design in more than 40 years. The review came in ahead of schedule and 11% under budget, according to the Department of Energy’s account of the approval, with TerraPower crediting a risk-informed licensing basis it helped pioneer with regulators.

Ontario Power Generation got there first on the construction side, breaking ground on a BWRX-300 at Darlington in May 2025, the first SMR under construction in the Western world. Would a second, third, and fourth unit at the same site actually come in cheaper, the entire premise behind modular manufacturing? That’s the number every government listed above is implicitly betting on, and it’s still mostly a forecast rather than a track record. A few things worth watching before taking the modular cost curve on faith:

• Whether TerraPower’s under-budget review translates into an under-budget build, not just a fast license

• Whether Ontario’s second and third Darlington units actually get cheaper per megawatt than the first

• Whether Sweden and India’s projects, both first-of-a-kind for their respective countries, hit their own cost targets

So which of these five bets looks strongest to you: the demand story, the security story, or the industrial one? Worth revisiting this list in twelve months, because at least one of these numbers is going to look very different by then.

BasicsSmall modular reactors stopped being a slide-deck concept somewhere in the last eighteen months. TerraPower broke ground on an actual reactor vessel foundation in Wyoming. China’s CNNC is weeks from connecting a commercial SMR to the grid. Microsoft, Amazon, Google, and Meta have signed enough nuclear power purchase agreements to add up to nearly 10 gigawatts of committed capacity. And yet the public conversation about SMRs is still running on assumptions that were outdated the moment they were written down.

Some of that is genuine confusion between regulatory milestones that sound similar but mean very different things. Some of it is leftover skepticism from the industry’s most public failure. And some of it is just people repeating what they read in 2023 without checking whether it’s still true. Here are seven claims about SMRs that get repeated constantly, and what the record actually shows.

Myth 1: an NRC stamp means a reactor is basically getting built

NuScale Power is still the only company with a full NRC design certification for an SMR, and it holds two: the original 50 MWe module and, as of the agency’s February 2026 announcement, an uprated 77 MWe version, good for a 462 MW six-module plant. That’s a real, hard-won regulatory achievement. It is also not the same thing as a reactor under construction.

Design certification tells the NRC a blueprint is safe to reference. It says nothing about whether a specific utility will build it, finance it, or connect it to a grid. NuScale’s flagship project, the six-module Carbon Free Power Project in Idaho, was cancelled in November 2023 after utilities backed out over rising costs. Design approval didn’t save it.

Compare that with TerraPower, which has no NRC design certification at all for its Natrium reactor, a 345 MWe sodium-cooled fast reactor with molten salt storage. What it has instead is a construction permit, approved by NRC commissioners in March 2026, the first ever granted for a commercial non-light-water reactor, and workers are now pouring the actual foundation in Kemmerer, Wyoming.

The three-stage reality worth remembering:

• Design certification: the blueprint is cleared, no site required

• Construction permit: a specific plant at a specific site gets approved to build

• Operating license: required before the reactor can actually run

NuScale has stage one locked down twice over and nothing under construction. TerraPower skipped straight to stage two and has cranes on site. Neither company has cleared stage three yet.

Myth 2: small reactors are automatically cheap reactors

The pitch has always been that factory-built modules dodge the cost overruns that plague giant one-off plants. The Carbon Free Power Project tested that pitch in the real world, and the real world won. NuScale’s target price for power from the Idaho project climbed from $58 per megawatt-hour to $89 per megawatt-hour, a 53% jump, before the project collapsed entirely under a $1.4 billion DOE cost-share deal that never got spent as planned.

That doesn’t mean the underlying economic argument is dead. It means it’s unproven at scale. The industry’s actual bet isn’t that reactor one is cheap, it’s that reactor six through ten get cheap through repetition, the same learning curve that brought down costs in naval reactor manufacturing. Most developers cite a target of roughly 40% cost reduction once a design hits serial production. TerraPower’s Kemmerer plant alone runs up to $4 billion for a single unit. Whether that curve actually bends the way the slide decks promise won’t be knowable until several identical units are built and operating, not modeled.

If you’re pricing an SMR deal today, ask which number you’re actually being quoted: the first-of-a-kind price, or the promised sixth-unit price. Those are very different bets.

Myth 3: no SMR has ever produced a real watt of electricity

This one was defensible until roughly six months ago. It isn’t anymore.

China’s CNNC completed cold functional testing on Linglong One, a 125 MWe pressurized-water SMR (also called the ACP100) at the Changjiang Nuclear Power Plant on Hainan island, in October 2025, and is targeting grid connection in the first half of 2026. First concrete went down in July 2021, putting the project on a roughly 58-month construction timeline, a real data point for anyone trying to model how long these things actually take once permitting is done. CNNC expects the unit to produce about 1 billion kilowatt-hours a year, enough for roughly 526,000 households.

Russia’s floating Akademik Lomonosov plant, meanwhile, has been quietly supplying grid power to the remote town of Pevek since 2020. It’s not glamorous, and it’s not American, but it’s real electrons on a real grid.

Worth sitting with: the “it’s all still theoretical” argument has a shrinking shelf life. Every quarter that passes without an update on where Linglong One actually stands is a quarter where the myth gets a little more wrong.

Myth 4: every SMR needs the same exotic fuel

Not even close. HALEU, high-assay low-enriched uranium, enriched between 5% and 20% U-235 versus the roughly 3% to 5% used in today’s fleet, is required by most of the advanced non-light-water designs: TerraPower’s Natrium, X-energy’s Xe-100, Kairos Power’s Hermes, Oklo’s Aurora, and Radiant’s Kaleidos all need it. Conventional light-water SMRs, including NuScale’s design and GE Hitachi’s BWRX-300, run on the same standard low-enriched fuel already used across the existing US reactor fleet.

The HALEU side of that split is genuinely tight. Centrus Energy remains the only US commercial HALEU producer, and its demonstration cascade in Piketon, Ohio, turns out roughly 900 kilograms a year, a fraction of the multi-ton annual demand once Natrium and Xe-100 start pulling fuel. DOE is throwing real money at the gap: a $900 million task order to Centrus in January 2026, part of a $2.7 billion, ten-year commitment, on top of a separate NRC Part 70 license X-energy’s fuel subsidiary won in February 2026 for a facility rated at 5 metric tons of uranium annually, enough to feed up to 11 Xe-100 units. The urgency traces partly to the 2024 US ban on enriched uranium imports from Russia’s Rosatom, which had been the dominant commercial HALEU supplier.

So: some SMRs face a real fuel bottleneck. Others could be fueled tomorrow with supply chains that already exist. Lumping them together obscures which projects are actually at risk of delay and which aren’t.

Myth 5: Big Tech’s nuclear deals are just press releases

By mid-2026, tallying every publicly announced hyperscaler nuclear commitment gets you to roughly 13 deals and 9.8 gigawatts of committed capacity across Microsoft, Google, Amazon, and Meta. These aren’t handshake agreements sitting in a drawer. Microsoft’s $16 billion, 20-year power purchase agreement for the Crane Clean Energy Center, the restarted Three Mile Island Unit 1, covers 835 MW and just got accelerated to arrive in the second half of 2027 after FERC approved a transmission waiver on June 1, 2026 that moved 760 MW of grid-connection rights over from a neighboring site. DOE closed a $1 billion loan for that project in November 2025.

Amazon put roughly $700 million into X-energy for up to 12 Xe-100 units at a Pennsylvania campus, part of a broader $50 billion partnership targeting 960 MW. Google signed an order-book arrangement with Kairos Power for 500 MW. Meta’s January 2026 round of agreements, spanning Vistra, TerraPower, Oklo, and Constellation, targets up to 6.6 gigawatts, including funding for two Natrium units and a 1.2 GW campus with Oklo in Pike County, Ohio.

Keeping all of that straight, who signed what, which MW figures are contracted versus aspirational, which regulatory approvals are still pending, is genuinely a full-time job at this point. The context that makes news like this meaningful is exactly what SMRbrief Pro was built to provide.

Myth 6: SMRs are a niche play for small or remote grids

The IAEA still frames part of the case for SMRs around islands, remote communities, and grids too small for a traditional 1,000+ MW plant, and that use case is real. But the gigawatt-scale numbers above tell a different story about where the actual capital is going right now. Data center operators aren’t buying SMRs because their local grid is undersized. They’re buying them because AI training clusters need firm, continuous power at a 95%-plus capacity factor, something intermittent solar and wind can’t deliver on their own. IDTechEx currently forecasts the global SMR market growing to $53.8 billion by 2036 and toward $300 billion by 2046, driven substantially by exactly this demand curve.

If your mental model of an SMR customer is a mining camp or an Arctic town, it’s time to update it. The 2026 customer is increasingly a hyperscale data center campus with a power appetite measured in hundreds of megawatts.

Myth 7: the US has this race locked up

American headlines about TerraPower’s construction permit and NuScale’s design certifications can create the impression the US is comfortably ahead. China’s timeline argues otherwise. Linglong One reaching commercial operation in the first half of 2026 would make it the first land-based commercial SMR to actually generate power, years before NuScale, TerraPower, or the UK’s Rolls-Royce SMR reach the same milestone. China had roughly 29 reactors under construction nationwide as of early 2026, a construction pipeline far larger than anything currently underway in the US, UK, or Canada combined.

The US retains real strengths worth naming:

• The only construction permit ever issued for a commercial non-light-water reactor

• The deepest pool of private capital, courtesy of hyperscaler nuclear deals

• A dedicated, multi-billion-dollar federal push to solve the domestic HALEU bottleneck

But “ahead” isn’t a settled fact, it’s a contested claim that depends heavily on which milestone you’re measuring. Grid connection first, or commercial-scale advanced reactor technology first? Those might not be the same race.

Which of these seven gets corrected in the coverage you read next depends less on the technology and more on whether the reporting keeps pace with a sector that’s now moving on a quarterly, not yearly, clock.

Investments$141 to $220 per megawatt-hour. That’s where Lazard puts new-build nuclear generation in the US today, well above new-build natural gas combined cycle at $48 to $109/MWh. Anyone arguing SMRs win on cost alone is skipping half the spreadsheet. But cost is only one column, and the other two, safety and emissions, tell a very different story. Here’s what the actual data says across all three. ⚡

This isn’t a cheerleading exercise for nuclear, and it isn’t a hit piece on gas or coal either. It’s a straight comparison using Lazard’s 2025 LCOE+ report, Our World in Data’s mortality figures, and the IPCC’s lifecycle emissions analysis, the same three sources any serious analyst would reach for. 📊

Cost: the sticker shock and the subsidy math

Nuclear’s cost problem is real, and pretending otherwise does the sector no favors. Lazard’s 2025 LCOE+ report puts new-build nuclear at $141 to $220/MWh unsubsidized, against $48 to $109/MWh for new gas combined cycle and $71 to $173/MWh for new coal. Gas peaking plants are the one fossil category nuclear actually beats on paper, running $149 to $251/MWh. 💰

The gap gets worse before it gets better, because first-of-a-kind SMR economics are brutal. NuScale’s Carbon Free Power Project in Idaho saw its target price jump from $58/MWh to $89/MWh as construction costs rose 75%, from $5.3 billion to $9.3 billion, before the project’s utility customers walked away and NuScale cancelled it in 2023. Wood Mackenzie pegs current FOAK SMR costs around $180/MWh, with nth-of-a-kind production potentially bringing that down to roughly $100/MWh by 2030, a forecast that depends entirely on someone actually building enough reactors to reach that scale.

Two things are pulling nuclear’s real-world economics in the opposite direction from that Lazard chart, though. First, hyperscalers with balance sheets that dwarf any municipal utility are underwriting the FOAK premium directly: Meta’s agreement covering up to 8 TerraPower Natrium reactor plants and Amazon’s $50 billion partnership with X-energy for 960 MW of Xe-100 capacity both remove the financing risk that killed NuScale’s Idaho project. Second, gas isn’t standing still either. Lazard’s own 2025 report flags new combined-cycle gas turbine costs at a 10-year high thanks to turbine shortages and rising equipment costs, which narrows the gap from the other direction. Tracking which side of that gap closes faster, project by project, is exactly the kind of thing SMRbrief Pro members can search, filter, and export the full intelligence picture behind developments like this one.

• New-build nuclear (Lazard 2025): $141-220/MWh

• New-build gas combined cycle: $48-109/MWh

• New-build coal: $71-173/MWh

• Gas peaking: $149-251/MWh

• FOAK SMR (Wood Mackenzie estimate): ~$180/MWh, trending toward ~$100/MWh NOAK by 2030

Safety: the deaths-per-terawatt-hour gap fossil fuels can’t close

This is where the comparison stops being close. Nuclear power causes roughly 0.03 deaths per terawatt-hour of electricity generated, according to Our World in Data’s analysis, a figure that already includes Chernobyl and Fukushima. Coal causes 24.62 deaths per TWh. Oil causes 18.43. Even natural gas, the cleanest-burning fossil fuel, causes 2.82 deaths per TWh, almost entirely from air pollution rather than accidents. Run the math and nuclear results in 99.8% fewer deaths than coal per unit of electricity produced. 🔬

The regulatory system is starting to reflect that math. In 2023 the NRC finalized a performance-based emergency preparedness rule that lets SMR developers calculate a plant-specific emergency planning zone based on actual accident consequences, instead of defaulting to the 10-mile zone required for conventional reactors, on the argument that passive safety designs cap how much radioactive material could ever escape. The Union of Concerned Scientists opposed the rule, and that disagreement is legitimate: a statistically safe technology can still fail badly in a specific instance, and probability isn’t the same thing as certainty. Fair point. It doesn’t change the underlying mortality data, though, which fossil fuels have never come close to matching even before you count the roughly 1 million to 2.5 million annual deaths attributed globally to fossil-fuel air pollution.

• Nuclear: ~0.03 deaths/TWh (includes Chernobyl and Fukushima)

• Natural gas: 2.82 deaths/TWh

• Oil: 18.43 deaths/TWh

• Coal: 24.62 deaths/TWh

Emissions: the lifecycle numbers behind the “clean” label

Cost is nuclear’s weak spot and safety is its strongest argument. Emissions sit close to the strong end too, though not by quite the margin some advocates claim. 🌍 The IPCC’s median lifecycle estimate puts nuclear at roughly 12 grams of CO2-equivalent per kilowatt-hour, comparable to wind and lower than solar, against roughly 490 g/kWh for natural gas combined cycle and 820 g/kWh for coal. That’s a lifecycle number, covering mining, construction, and decommissioning, not just the zero direct emissions at the point of generation. The UN Economic Commission for Europe ran its own 2022 analysis and landed even lower, at 5.1 to 6.4 g/kWh for nuclear, the lowest figure among all low-carbon technologies it assessed.

There’s real spread in the academic literature worth acknowledging rather than glossing over. Some published life-cycle assessments put nuclear’s emissions as high as 100+ g/kWh depending on uranium enrichment method and ore grade, since gaseous diffusion enrichment runs roughly twice the emissions of centrifuge enrichment. That’s not nothing. But even the high end of that range sits an order of magnitude below coal, and centrifuge enrichment, the method Centrus Energy and most Western suppliers now use, sits at the low end.

Capacity factor matters here too, because a plant that runs constantly displaces more fossil generation than one that runs intermittently. US nuclear plants operated above 93% capacity factor in 2023. Natural gas combined-cycle plants averaged closer to 57%, and coal plants ran at just 42.5% in 2024, according to EIA data, meaning both fossil categories need standby backup capacity that nuclear generally doesn’t.

• Nuclear lifecycle emissions (IPCC median): ~12 g CO2eq/kWh

• Natural gas combined cycle: ~490 g CO2eq/kWh

• Coal: ~820 g CO2eq/kWh

• Nuclear capacity factor (US, 2023): 93%+

• Gas/coal capacity factor: 57% and 42.5%, respectively

Where fossil fuels still win, for now

None of this makes SMRs the obvious choice in every situation, and pretending otherwise would be exactly the kind of breathless optimism this publication tries to avoid. Gas remains cheaper to build, faster to permit, and vastly more flexible for utilities that need to add capacity in 18 months rather than a decade. GE Hitachi’s 300 MWe BWRX-300, now under construction at Darlington, Ontario, and TerraPower’s 345 MWe Natrium plant, which broke ground in Kemmerer, Wyoming this spring, are both multi-year projects competing against gas turbines that can be ordered, built, and running well before either reactor produces its first electron. 🚀

Have you run these numbers against a specific project you’re tracking, and did the comparison land where you expected? The honest read is that nuclear wins decisively on safety, wins comfortably on emissions, and still has real ground to make up on cost, ground that depends less on reactor physics than on whether Meta, Amazon, and Google keep writing the checks that get FOAK projects to NOAK pricing.

ConcernsKemmerer, Wyoming has 2,415 residents and, as of April 2026, an active nuclear construction site on its edge. TerraPower broke ground on its 345 MWe Natrium reactor there this spring, and the town’s mayor says a local poll found residents overwhelmingly in favor. That’s not the reaction most people expect when a nuclear plant moves in next door. 🏠

But Kemmerer isn’t unusual because its residents are unusually brave. It’s unusual because most people asking “should I worry about this” have never gotten a straight answer with numbers attached. Here are the five objections that come up in every SMR siting fight, and what the regulatory filings, peer-reviewed studies, and operating history actually say about each one.

Radiation exposure: what the dose numbers actually show

This is the first fear, and it’s the easiest to quantify. The IAEA puts the average annual radiation dose for someone living near an operating nuclear plant at roughly 0.0001 millisieverts, a fraction so small it’s routinely compared to eating a single banana. 🍌 Natural background radiation from soil, cosmic rays, and radon delivers 2 to 3 mSv per year to the average person, meaning routine plant emissions add almost nothing to that baseline.

Utah has actually turned this into a public argument. Pro-nuclear billboards there claim plants emit less radiation than bananas, and a University of Utah nuclear engineering professor confirmed the comparison holds for routine operation, while critics at HEAL Utah countered that the real issue is choice: nobody consents to living along a waste transport route the way they choose to eat fruit. That’s a fair distinction, and it’s worth sitting with rather than waving away.

One study complicates the clean narrative further. Germany’s KiKK study found children under five living within 5 kilometers of a nuclear plant had roughly double the leukemia rate of children farther away, even though emissions at that distance run at least 1,000 times below natural background levels. The researchers themselves called a direct causal link “implausible,” and the region’s overall leukemia rate matched the German national average. Nobody has explained the finding. It’s the honest asterisk on an otherwise reassuring dataset:

• Routine SMR emissions: roughly 0.0001 mSv/year at the site boundary

• Annual background radiation: 2 to 3 mSv/year

• A single chest CT scan: roughly 100,000x a banana-equivalent dose

• KiKK study finding: elevated childhood leukemia within 5 km, cause unconfirmed

Meltdown risk: why SMR safety cases lean on smaller emergency zones

Every conventional US reactor sits inside a 10-mile emergency planning zone, the area where local governments must maintain evacuation plans. That distance was set for large light-water reactors decades ago, and it’s the number most people picture when they imagine “danger radius.” SMRs are trying to shrink it. ⚡

In 2023 the NRC finalized a performance-based emergency preparedness rule that lets SMR and advanced reactor developers calculate a plant-specific emergency zone based on actual accident consequences, rather than defaulting to the 10-mile standard. NuScale’s methodology for doing this was accepted by the NRC’s Advisory Committee on Reactor Safeguards back in 2022, on the argument that its passive cooling design, no pumps, no operator action, no outside power needed, physically caps how much radioactive material could ever escape.

The Union of Concerned Scientists opposed the rule, arguing new technologies deserve more offsite planning scrutiny, not less, and that argument hasn’t gone away. It’s a real regulatory disagreement, not a settled question, and both sides are citing the same accident physics to reach opposite conclusions. What’s harder to dispute is the historical record: Three Mile Island’s 1979 partial meltdown, the worst commercial nuclear accident in US history, produced no measurable increase in cancer rates among nearby residents in the decades since, according to long-term epidemiological follow-up.

Property values: the data says the fear is priced wrong

This one has been studied more than almost any other nuclear worry, and the results consistently surprise people. 📊 A landmark analysis of four nuclear plants found no statistically significant drop in nearby home prices, and one dataset even showed values rising with proximity for the closest properties, plausibly because plant workers wanted to live near their jobs. A separate hedonic-pricing study of the Zion, Illinois site, examining sales more than a decade after the plant shut down with spent fuel still stored on-site, found the same thing: perceived stigma doesn’t show up in what buyers actually pay.

Kemmerer offers a live, opposite-direction version of this story. Local real estate agents describe a housing shortage, not a slump, as up to 1,600 peak construction workers move in ahead of the plant’s 2030 target date, pushing rents and home prices up rather than down. That’s obviously a different phenomenon than long-run stigma pricing, but it undercuts the assumption that reactors drain a town’s real estate market.

None of this means every site behaves identically. Distance, local economy, and whether the plant is operating or shut down all move the needle differently, which is exactly the kind of site-by-site nuance that’s easy to lose in a single averaged statistic. SMRbrief Pro members get the structured data layer that transforms individual news items into real market intelligence, which matters when the honest answer to “what happens to property values” is it depends where you’re asking about.

Water use: the real environmental trade-off

Unlike the first three concerns, this one holds up reasonably well under scrutiny, at least for some designs. 💧 Traditional light-water SMRs, the pressurized-water designs closest to today’s commercial fleet, still need substantial cooling water, and once-through systems that draw from a river or lake and discharge it warmer downstream can strain water-limited sites, especially when a data center customer is co-located and drawing its own supply.

The counterpoint is that SMRs aren’t one technology. Non-light-water designs, gas-cooled, sodium-cooled, molten-salt, can run on closed-loop or fully passive cooling systems that cut water consumption dramatically, and the Department of Energy notes that reactor size alone shapes how much cooling infrastructure a site needs. TerraPower’s sodium-cooled Natrium and Kairos Power’s molten-salt-cooled Hermes both fall into this lower-water category, while a conventional PWR-style SMR does not.

This is one concern where the honest answer is “ask which design”: the technology choice matters more than the fact that it’s nuclear.

• Once-through cooling: highest water draw, requires a large adjacent water body

• Closed-loop cooling towers: recirculates water, moderate consumption

• Air-cooled or hybrid systems: lowest water use, preferred for arid sites

• Non-light-water SMRs (sodium, gas, molten salt): generally lower water intensity than PWR designs

Nuclear waste: dry casks, and why Kemmerer says yes anyway

Spent fuel is the concern that ages best, in the sense that it never fully goes away, and that’s precisely what worries people. But the storage method used to manage it has an unusually clean track record. ♻️ The NRC’s own assessment states that dry cask storage has “an excellent safety record” across more than 30 years of US use, with steel-and-concrete casks engineered to withstand earthquakes, floods, and direct projectile impacts, and no credible accident scenario that produces consequences beyond the site boundary.

SMRs may actually improve this picture rather than worsen it. Several designs run on HALEU fuel, enriched between 5% and 20% U-235 instead of the roughly 3% to 5% used in today’s fleet, and the Energy Information Administration notes that higher burnup from HALEU can shrink the volume of spent fuel a reactor produces per unit of electricity generated. Smaller volume doesn’t solve the long-term disposal question, the US still has no permanent geologic repository, but it does mean fewer casks accumulating at each site over a plant’s operating life.

Back in Kemmerer, residents aren’t voting on abstractions. They watched their coal plant, the Naughton facility, head toward retirement, and chose a reactor that keeps roughly 250 permanent jobs in a town of 2,400 people over the alternative of watching the local economy hollow out. Which of these five concerns would change your own answer if a developer proposed a reactor in your town?

BasicsEvery SMR press release now arrives loaded with acronyms. NRC, GDA, HALEU, FOAK, LCOE, MWe. Skip past enough of them and you start nodding along without actually knowing what a company just promised, or how expensive that promise might get. 📚

That’s the problem this glossary solves. Ten terms, defined once, with the specific numbers and deals attached to each one so the definition sticks. Print this out. You will need it the next time a hyperscaler announces another gigawatt-scale nuclear deal.

The basics: SMR, MWe, and microreactor

Start with the term in the name. A small modular reactor is a nuclear plant that generates roughly 300 megawatts electric or less and is built largely from factory-fabricated modules, then shipped to site rather than poured and welded in place over a decade. 🏭 The modular part is the whole pitch: standardize the manufacturing, and the economics are supposed to improve with every unit built, the same way Boeing gets cheaper per plane after the first one. The International Atomic Energy Agency counts more than 90 distinct SMR designs in development worldwide, which tells you how crowded and unsettled this field still is.

MWe just means megawatts electric, the actual electricity output a plant delivers to the grid, as opposed to MWt (megawatts thermal), which measures heat output before conversion losses. Always ask which one a company is quoting. ⚡ The spread across current designs is wide:

• NuScale’s uprated US460 module: 77 MWe per unit

• GE Hitachi’s BWRX-300, now under construction at Darlington, Ontario: 300 MWe

• TerraPower’s Natrium, breaking ground in Kemmerer, Wyoming: 345 MWe

• Rolls-Royce SMR, headed toward deployment in Wales: 470 MWe

• Microreactors, the smallest subcategory: typically 1 to 20 MWe

That last line introduces microreactor, a reactor so small it can be transported by truck, rail, or cargo plane. The Pentagon’s Project Pele, built with the Defense Department’s Strategic Capabilities Office, is a 1-to-5 MWe TRISO-fueled design meant to power forward operating bases. Microreactors trade raw output for portability, which is a very different value proposition than a 470 MWe grid plant.

The paperwork: design certification, GDA, and construction permits

Nuclear regulators do not license a company. They license a design, then separately license the site and the construction, and the vocabulary around each step trips up more readers than any technical term on this list. 📋

In the US, the NRC’s design certification (or the newer standard design approval, a faster variant) confirms a reactor design meets federal safety requirements and can be referenced in future license applications. NuScale’s 50 MWe module became the first SMR ever certified in January 2023, and the company’s uprated 77 MWe US460 followed with a standard design approval in May 2025. As of mid-2026, NuScale remains the only company holding either.

The UK runs a parallel process called Generic Design Assessment, or GDA, jointly overseen by the Office for Nuclear Regulation, the Environment Agency, and Natural Resources Wales. It’s a three-step, multi-year review, and Rolls-Royce SMR’s 470 MWe design entered Step 3 in July 2024, with the full 53-month process expected to wrap by the end of 2026. Passing GDA doesn’t guarantee a build gets approved either; it just clears the design itself.

Then there’s the construction permit (CP), which authorizes actually pouring concrete for a specific reactor at a specific site, distinct from the combined license (COL) that bundles construction and operating authority into one approval. Kairos Power holds the first NRC construction permit ever issued for an advanced reactor, covering its Hermes test unit in Oak Ridge, Tennessee, and TerraPower’s construction permit application for Natrium is under active NRC review right now. Three separate approvals, three separate timelines, and none of them guarantees the next one comes easily.

The fuel bottleneck: HALEU

HALEU, high-assay low-enriched uranium, is fuel enriched to between 5% and 20% U-235, well above the roughly 3% to 5% used in today’s commercial reactor fleet. 🇺🇸 More than half of the SMR designs currently in development need it, and until recently the only reliable commercial supplier was the Russian state firm Tenex.

That’s the bottleneck everyone in this sector talks about. Centrus Energy operates the only domestic HALEU production line, a demonstration cascade at the American Centrifuge Plant in Piketon, Ohio, and delivered its contracted 900 kilograms to the Department of Energy by mid-2025. The DOE followed with a $900 million HALEU enrichment award in January 2026 to scale that up. On the customer side, Centrus signed a letter of intent to supply HALEU to Oklo for up to five Aurora powerhouses at a planned 1.2 GW campus in southern Ohio, with deliveries starting in 2029. ⛓️ Every HALEU supply deal, DOE contract, and enrichment milestone we cover ends up structured, sourced, and searchable in SMRbrief Pro, which is genuinely the fastest way to track a fuel chain this fragmented.

A federal ban on Russian uranium imports takes full effect in 2028, so the fuel question isn’t theoretical. It’s the single biggest constraint standing between a certified design and an operating reactor. Which of these fuel deals do you think closes first, Oklo’s or one of the hyperscaler-backed developers’? 🔬

The money terms: FOAK, NOAK, LCOE, and offtake agreements

FOAK means first-of-a-kind, the initial build of any new reactor design, when engineers are still finalizing details during construction and the supply chain doesn’t exist yet. NOAK, nth-of-a-kind, is what happens once a design has been built enough times that manufacturing, workforce training, and permitting all get repeatable. 💰 The gap between the two is enormous and it’s the reason SMR economics are so contested.

The clearest case study is NuScale’s own Carbon Free Power Project in Idaho. Target pricing started at $58 per megawatt-hour in 2021. By late 2023, construction cost estimates had jumped 75%, from $5.3 billion to $9.3 billion, pushing the target price to $89/MWh. The project’s utility customers walked, and NuScale cancelled it. Analysts at Wood Mackenzie peg FOAK SMR costs around $180/MWh today, possibly falling closer to $100/MWh by 2030 as NOAK production kicks in, though that forecast depends entirely on someone actually building enough units to get there.

LCOE, levelized cost of electricity, is the standardized metric analysts use to compare that cost per MWh across technologies over a plant’s lifetime. Lazard’s most recent unsubsidized estimate puts conventional nuclear at $141 to $221/MWh, well above utility-scale solar or onshore wind. It’s the number every SMR developer is racing to beat, and so far only aspirational NOAK projections get there.

Which brings us to offtake agreements, sometimes structured as power purchase agreements (PPAs), the contracts that actually make a reactor bankable by locking in a buyer before a shovel hits dirt. The last eighteen months have produced more of these than the prior decade combined:

• Meta and Constellation Energy: 1.1 GW, 20-year PPA, Clinton Clean Energy Center, Illinois (June 2025)

• Amazon and Talen Energy: 1,920 MW through 2042, Susquehanna Steam Electric Station, Pennsylvania

• Microsoft and Constellation: 835 MWe, 20-year PPA to restart Three Mile Island Unit 1 as the Crane Clean Energy Center, targeting H2 2027

• Meta and TerraPower: agreement covering up to 8 Natrium reactor plants, announced January 2026

• Amazon and X-energy: $50 billion partnership to deploy 960 MW of Xe-100 units in Texas

Big tech’s shift toward becoming anchor tenants for nuclear power, rather than just buying credits, is arguably the single biggest structural change in this industry’s financing since Vogtle. 📈 An offtake agreement from a company with Meta’s or Amazon’s balance sheet is exactly what lets a developer absorb FOAK cost risk that sank NuScale’s municipal utility customers.

So next time a headline drops with a number attached, you’ll know whether it’s describing thermal output or electrical, a design win or a construction win, and a first-of-its-kind gamble or a repeatable one. What’s the term you still see misused most often in nuclear coverage?

Every energy debate eventually collides with the same annoying word: reliable. 🔌 Solar and wind fans point to plunging costs and a decade of dominant growth. Nuclear fans point to uptime and shrug at the price tag, at least for now. Both sides have real data behind them, which is exactly why this fight refuses to resolve itself in a single tweet.

Small modular reactors, or SMRs, have entered this argument as nuclear’s answer to renewables’ speed and flexibility: factory-built, theoretically cheaper than a traditional plant, and small enough to sit next to the thing it’s powering. Whether that answer actually beats solar and wind depends entirely on which question you’re asking. Let’s work through the honest version. ⚡

What “reliable” actually means

This is where most of these arguments go sideways, because people use “reliable” to mean three different things at once. 🔍 A grid planner cares about capacity factor, the percentage of time a power source actually delivers its rated output. A homeowner cares about whether the lights stay on. A data center operator cares about something closer to both, at a punishing 99.999% uptime standard. ⏱️

By capacity factor, the gap is enormous:

• SMRs: roughly 90-95%, similar to today’s large nuclear fleet

• Onshore wind: roughly 35-45% in strong locations, lower elsewhere

• Utility-scale solar: roughly 20-25%, and zero at night by definition

• Offshore wind: up to 60% in the best sites, though at a steep cost premium

That’s the core of the nuclear argument: a reactor hums along at nearly full output around the clock, while solar and wind need batteries, backup gas plants, or a much larger buildout to cover the gaps. According to a piece hosted by Utility Dive, critics of this framing argue that a well-managed grid mixing wind, solar, and storage can already meet demand reliably around 95% of the time, which changes the calculus considerably once you stop treating any single source in isolation. So which number matters more to you: the 90%+ that a single reactor promises, or the 95% a well-managed mixed grid can already deliver? Neither side is lying to you here. They’re just measuring different things. 🎯

What the money says right now

Cost is where the story gets uncomfortable for SMR boosters. 📉 Lazard’s widely cited 2025 Levelized Cost of Energy+ report found that unsubsidized utility-scale solar and onshore wind remain the cheapest new-build power sources for the tenth straight year, with solar landing around $38 to $212 per megawatt-hour and onshore wind around $37 to $86, according to coverage from pv magazine USA. Nuclear, by comparison, landed at $141 to $220 per megawatt-hour in that same analysis. 💰

First-of-a-kind SMRs look even rougher up close:

• Current SMR power is estimated at $80 to $150 per megawatt-hour, well above solar and wind on a pure cost basis

• Solar runs roughly $30 to $50 per megawatt-hour unsubsidized

• Onshore wind runs roughly $25 to $50 per megawatt-hour

• SMR proponents are betting serial production eventually drops costs to $50 to $80 per megawatt-hour, a threshold nobody has actually hit yet

That last bullet is the whole industry’s bet, and it’s not a small one. NuScale’s cancelled Utah project, where costs ballooned from $5.3 billion to $9.2 billion before the plug got pulled, is the case study skeptics reach for constantly. Some analysts, including a sharply worded CleanTechnica critique, argue the learning curve that made solar panels and wind turbines cheap won’t transfer to nuclear, because each reactor design still carries its own licensing, fuel qualification, and safety case. That’s a fair point, and one worth sitting with even if you’re rooting for nuclear’s comeback. 🤨

The land and siting angle

Here’s where SMRs pull ahead on a metric that rarely makes headlines: footprint. 🏗️ A 300-megawatt SMR can fit on roughly 10 to 15 acres, while a solar farm producing the equivalent annual energy, once you account for that 20-25% capacity factor, needs something closer to 2,000 acres. That’s not a rounding error. It’s a genuine difference that matters enormously in places where land is scarce, expensive, or already spoken for. 📐

The siting flexibility matters just as much as the acreage:

• SMRs can be built close to the load they’re serving, including directly on a data center campus

• Solar and wind need favorable sun or wind resources, which don’t always line up with where power is actually needed

• Transmission buildout to move renewable power from good sites to demand centers is itself slow and expensive

• A reactor doesn’t care whether it’s cloudy or calm outside

This is precisely why companies building hyperscale AI data centers keep signing nuclear deals instead of just buying more solar panels. 🏢 It’s not that solar is a bad technology. It’s that a facility drawing 100 megawatts continuously can’t run on power that only shows up a quarter of the day. If you want the running tally of exactly which hyperscaler has signed which reactor deal and when, SMRbrief Pro keeps that database updated as the contracts land.

Why the smartest bet might be both, not either

Strip away the tribalism, and the two technologies are arguably solving different problems. 🧩 Solar and wind are cheap, fast to build, and excellent at bulk decarbonization when paired with enough storage and transmission. SMRs, if they can actually hit their cost targets, are better suited to the specific job of dense, always-on, weather-independent power in a small footprint.

A few honest caveats worth holding onto before picking a side: 🧭

• Not a single commercial SMR is currently operating at scale in the U.S., so every cost projection above 2027 is still a forecast, not a track record

• Solar and wind have a proven fifteen-year track record of hitting or beating their own cost-decline projections

• Reactor projects have a long history of running over budget and behind schedule, Vogtle 3 and 4 in Georgia being the most recent expensive example

• Grid operators increasingly treat “baseload” as an outdated concept, favoring flexible combinations of sources over any single dominant one

So which one actually wins? Neither, cleanly, at least not yet. Solar and wind win on cost and speed today. SMRs are making a real, if unproven, case on reliability and land use for the specific jobs that need it. The honest move for now is watching which SMR developer actually connects a reactor to a real customer’s meter first. What would change your own mind on this: a single SMR hitting its cost target, or wind and solar closing the reliability gap with cheaper storage? ⚖️

Nuclear stocks used to be the kind of thing your uncle owned three shares of and never mentioned. Not anymore. 📈 Small modular reactors, or SMRs, have gone from a niche engineering concept to one of the most talked-about themes in energy investing, and a lot of that noise is coming from AI data centers that need round-the-clock, carbon-free power and can’t wait a decade for a traditional plant to get built.

I’m not a financial advisor, and nothing here is a recommendation to buy anything. What I can do is walk you through how SMR investing is actually structured, so you’re not just throwing money at a ticker because it showed up in your feed. ⚛️ Let’s get into it.

What you’re actually buying when you buy “SMR”

Here’s the first thing beginners get wrong: they treat “nuclear stock” as one category. It isn’t, not even close. 🎲 The SMR world splits into a handful of distinct businesses, each with a different risk profile, and lumping them together is how people end up disappointed. Here’s the breakdown: 🔗

• Pure-play reactor developers: companies designing and trying to commercialize SMRs or microreactors, with little or no operating revenue yet

• Fuel and enrichment suppliers: companies that mine uranium or produce the specialized fuel reactors need to run

• Diversified industrial giants: massive companies where SMRs are one growth bet among many other business lines

• Nuclear utilities: companies that already operate reactors and are exploring SMRs as an add-on

• Funds: baskets that hold a mix of the above so you’re not betting on a single name

According to smrintel.com’s investor guide, there were roughly 16 publicly traded nuclear and SMR-related stocks as of May 2026, spanning reactor developers, uranium miners, enrichment firms, services companies, and utilities. 🔍 That’s a genuinely small universe, small enough that a beginner can get their arms around it in an afternoon, which is more than you can say for most sectors.

The direct route: individual SMR and nuclear stocks

If you want concentrated exposure, individual stocks are the obvious path, but “concentrated” cuts both ways. 🎢 NuScale Power (NYSE: SMR) is often cited as the furthest along on paper, and per the Motley Fool, it’s the only SMR design to receive U.S. regulatory certification, though the stock has also swung wildly, including a roughly 73% drop over one recent stretch amid delays. Oklo (NYSE: OKLO) takes a different approach, pitching smaller reactors directly to data center operators and defense customers rather than the grid at large. ⚡

Beyond the two most-discussed names, a few other categories are worth knowing:

• BWX Technologies (NYSE: BWXT): a “picks-and-shovels” play that builds reactor components and fuel regardless of which developer wins

• Cameco (NYSE: CCJ): one of the world’s largest uranium miners, giving exposure to the fuel side rather than reactor technology itself

• Centrus Energy (NYSE: LEU): currently the only licensed U.S. producer of HALEU, the specialized fuel most next-generation reactors need

• GE Vernova (NYSE: GEV): a diversified energy giant whose SMR business, through its GE Hitachi joint venture, is one growth line among many

• Constellation Energy (NASDAQ: CEG): an established nuclear operator dipping into SMRs through partnerships rather than building its own

Worth sitting with for a second: none of the pure-play developers, Oklo and NuScale included, currently generate meaningful revenue from an operating commercial reactor. You’re betting on execution and timelines, not existing cash flow. That’s a very different kind of bet than buying a utility that’s been cashing power bills for fifty years. 💭

The diversified route: nuclear and uranium ETFs

If picking individual reactor companies feels like guessing which teenager grows up to be a CEO, funds are the more forgiving option. 🧺 As of April 2026, there were six U.S.-listed nuclear and uranium ETFs, according to an ETF comparison from Green Stocks Research, with roughly $15.2 billion in combined assets.

• Global X Uranium ETF (URA): the largest of the group, focused on uranium miners and nuclear component makers, and includes names like Cameco and Oklo

• Sprott Uranium Miners ETF (URNM): a purer bet on the mining side of the fuel cycle

• VanEck Uranium and Nuclear ETF (NLR): spreads across the entire value chain, from miners to utilities to reactor developers, and pays a small dividend

• Range Nuclear Renaissance ETF (NUKZ): weighted toward companies whose revenue is directly tied to building, operating, and servicing reactors

None of these funds are pure SMR plays. 🌊 They mix in uranium miners, established utilities, and legacy nuclear infrastructure alongside the newer reactor companies, which dilutes your SMR-specific exposure but also softens the blow if any single developer stumbles. If you’re the type who wants to track how these holdings shift and which developers are actually gaining ground versus just gaining headlines, SMRbrief Pro gives you the structured database to go deeper than any single article can. 🗂️

Risks a beginner shouldn’t skip past

This is the section people scroll past, and it’s exactly the one you shouldn’t. 🚨 SMR investing carries real, specific risks that go beyond the usual “stocks can go down” disclaimer. ⚠️

• Pre-revenue status: most pure-play developers are funded by government contracts, grants, and stock sales, not product sales, which means dilution is a constant risk

• Regulatory timelines: nuclear projects are uniquely sensitive to permitting, and a single delay at the NRC can push a company’s revenue timeline out by years

• Construction and cost overruns: traditional nuclear has a rough track record here, and it’s unproven whether SMRs actually solve it at commercial scale

• Uranium price cycles: fuel and mining stocks rise and fall with commodity prices, independent of how well any single reactor design is doing

• Political and public sentiment risk: nuclear currently enjoys broad bipartisan support, but that can shift quickly after a high-profile incident anywhere in the world

Have you actually sat down and calculated what percentage of your total portfolio you’d be comfortable losing entirely if a developer’s licensing application gets denied? That’s not a rhetorical question. It’s the number you should know before you place a single order. 🎯

Actually getting started

Once you understand the categories and the risk, the mechanics are pretty ordinary. 🧭 Open a brokerage account if you don’t already have one, decide how you want your exposure split between individual stocks and diversified funds, and size your position based on money you can genuinely afford to see cut in half. Nuclear stocks, especially the speculative developers, are not where you park an emergency fund. 💼

• Start with a fund like NLR or URA if you want broad exposure without picking winners

• Add individual names only once you understand each company’s specific business model and revenue stage

• Consider dollar-cost averaging into positions rather than buying all at once, given how volatile this sector has been

• Revisit your thesis every time a company reports earnings or hits a regulatory milestone, not just when the stock price moves

The SMR sector is still writing its first chapters. Which part of the value chain makes more sense for your own risk tolerance: the developers betting everything on one reactor design, or the fuel and services companies that get paid regardless of who wins? 🔋

Ask five people in the nuclear industry to define a microreactor, and you’ll get five confident, slightly different answers. That’s not because anyone is being sloppy. It’s because small modular reactor and microreactor grew up together, got used interchangeably for a decade, and only recently split into their own lanes as actual steel started showing up in Idaho, Alaska, and Tennessee. ⚛️

If you’ve followed the nuclear revival even loosely, you’ve probably seen both terms in the same headline, sometimes the same sentence. That’s fair, since a microreactor technically is a type of SMR. But treating them as synonyms glosses over what actually separates them: where these things go, how fast they show up, and who’s writing the checks. Let’s sort it out.

So what actually makes a reactor micro

The simplest way in is size, even though size alone doesn’t tell the whole story. Idaho National Laboratory’s Gateway for Accelerated Innovation in Nuclear program draws its line at 50 megawatts electric (MWe): anything below that, and often below 10 MWe, counts as a microreactor, while SMRs run roughly 50 to 300 MWe. The World Nuclear Association frames it a bit differently, defining SMRs generally as 300 MWe or less and reserving “microreactor” for the smallest designs, typically under 20 MWe. Wikipedia’s own breakdown lands somewhere in between: commercial SMRs deliver anywhere from 10 up to 300 MWe per module, and anything under 10 MWe usually earns the micro label.

None of these numbers match exactly, and that’s honestly the point. 🔬 Nobody has agreed on one hard cutoff, so the same reactor gets called a microreactor in one report and a “very small SMR” in another. What everyone does agree on is the scale difference from what’s already running. According to Idaho National Laboratory, microreactors are 100 to 1,000 times smaller than a conventional commercial reactor in electrical output, small enough to sit on a few acres and power a factory or a few thousand homes instead of an entire city. That’s not a rounding error, it’s a different category of machine. 📏

• SMRs: roughly 50-300 MWe, with some designs stretching toward 600 MWe

• Microreactors: generally under 20 MWe, frequently under 10 MWe

• Today’s average U.S. reactor: around 1,000 MWe

• Demonstration microreactors like Project Pele: as little as 1.5 MWe

So next time someone throws around either term, ask them what output they’re actually talking about, in MWe, not vibes. It clears up more confusion than any dictionary definition. Curious how your own back-of-napkin guess compares to the industry’s? 💡

It’s not just about megawatts

Here’s where the distinction gets more interesting than a spreadsheet of wattages. 🧭 According to Wikipedia’s entry on nuclear microreactors, microreactors are built specifically to be transportable, meaning they can be moved by road, rail, or air, and they typically fall in the 1 to 20 MWe range compared to the 20 to 300 MWe window for SMRs. That mobility is a design goal, not an afterthought. An SMR is still, in most cases, a permanent power plant, just a smaller and more factory-built one than the 1,000-plus-MWe giants utilities have built for decades. A microreactor is closer to a shippable appliance. 🚚

Most microreactor designs share a few traits that set them apart operationally, not just numerically:

• They ship as a single, pre-assembled unit rather than being built piece by piece on site

• Many run on high-assay low-enriched uranium (HALEU), fuel enriched higher than what today’s commercial fleet uses

• Several use TRISO fuel, tiny ceramic-coated uranium particles that are famously hard to melt down

• They’re designed to run for years, sometimes a full decade or more, without refueling

• Many can island themselves entirely off the electric grid, powering a single facility on their own

That last point matters more than it sounds. Per the U.S. Energy Information Administration, microreactors, as a subset of SMRs, generally sit at 20 MW or less and can run tied to the grid, disconnected from it entirely, or feeding a tiny local grid of their own. That flexibility is exactly why the Pentagon, remote mining operations, and increasingly data center operators have started paying attention. Nobody’s proposing to truck a 300-MWe SMR into a forward operating base or a hospital parking lot. A microreactor, at least on paper, could actually get there. 🌍

Who’s actually building these things

The company rosters for SMRs and microreactors overlap, but they’re not the same list, and it’s worth knowing who’s playing in which category. On the SMR side, you’ve got designs like X-energy’s Xe-100, already slated for a demonstration at Dow’s Seadrift site in Texas, GE Vernova Hitachi’s BWRX-300, Holtec’s SMR-300, and Westinghouse’s AP300, all sized in the tens to hundreds of megawatts and built for permanent grid connection.

Microreactor developers are a different crowd, and a scrappier one. 🚀 Oklo is developing its liquid-metal-cooled Aurora powerhouse and, according to Utility Dive, expects its combined licensing approach with the Nuclear Regulatory Commission to cut initial licensing timelines by 50 to 85 percent compared with a typical application. That’s not a small claim in an industry famous for permitting delays. Radiant Industries is building Kaleidos, a roughly 1-MWe reactor sized to fit inside a single shipping container, with testing at Idaho National Laboratory’s DOME facility targeted for this year. Westinghouse’s eVinci and BWXT’s Project Pele round out the field, alongside newer entrants like NANO Nuclear’s Kronos and Valar Atomics. If you’re tracking the nuclear market professionally, SMRbrief Pro gives you the structured database to go deeper than any single article can. Who do you think crosses the finish line to actual commercial operation first? 📈

• Oklo: Aurora, a liquid-metal-cooled fast reactor targeting commercial operation at Idaho National Laboratory

• Radiant: Kaleidos, 1 MWe, shipping-container form factor, aimed at diesel generator replacement

• Westinghouse: eVinci, the first microreactor with an NRC-approved instrumentation and control system

• BWXT: Project Pele, 1.5 MWe, built for the Department of Defense

• NANO Nuclear: Kronos micro modular reactor, with a proposed unit at the University of Illinois

Worth noting: none of these companies are generating meaningful revenue yet from an operating reactor. Every one of them is still pre-revenue on the hardware itself, funded by federal contracts, venture capital, and increasingly by hyperscalers hungry for power. That’s not a knock, it’s just where the industry actually stands in mid-2026. Skepticism and optimism can coexist here. 🤔

The military is writing the microreactor playbook first

If you want to see where microreactors are actually closest to running, skip the utility press releases and look at the Pentagon. 🪖 Project Pele, run by the Department of Defense’s Strategic Capabilities Office, broke ground at Idaho National Laboratory in 2024 to build a transportable 1 to 5 MWe reactor designed to fit inside four 20-foot shipping containers and travel by truck or cargo plane, according to the Department of Energy. BWXT delivered the initial HALEU fuel load to the lab in December 2025, with testing scheduled for 2027.

That’s just one program among several. The Air Force picked Eielson Air Force Base in Alaska to pilot a fixed microreactor and issued a notice of intent to award the contract to Oklo for a liquid-metal-cooled design. The Army’s newer Janus Program has already identified nine candidate installations for reactor siting. And the Advanced Nuclear Power for Installations program has cleared eight companies, including Oklo, BWXT, General Atomics, Kairos, Radiant, Westinghouse, and X-energy, to compete for base-power contracts. In February 2026, a Valar Atomics microreactor even made the trip from California to Utah aboard Air Force C-17s, proving the transport concept works in practice and not just in a slide deck. ✈️

None of these are commercial power plants yet in the way a utility-scale SMR eventually will be. They’re demonstrations, prototypes, and pilot programs, funded because the military needs resilient power at remote bases faster than a decade-long licensing process can normally deliver. 🔋 But that urgency is exactly why microreactors are likely to hit commercial reality before most SMRs do. Which of these programs do you think actually reaches full operation first: Pele, the Eielson pilot, or one of the Janus sites? 🎯

Every SMR pitch deck has the same slide. Factory-built modules, shorter construction timelines, lower financing risk, cheaper power. It sounds like common sense: smaller things generally cost less. 💰 But nuclear power has never played by common sense economics, and the honest answer to “which is cheaper” is genuinely more complicated than either side of this debate wants to admit. Let’s actually run the numbers. 🔬

The economies-of-scale problem nobody can escape

Here’s the physics-meets-economics reality that every SMR company has to fight against: bigger reactors have always been cheaper per unit of power, because a lot of the cost, containment structures, control systems, safety equipment, doesn’t shrink proportionally with reactor size. This is why the industry spent seventy years building reactors bigger, not smaller. 📉

The current cost estimates make that gap pretty explicit. <cite index=”65-1”>The EIA estimates overnight capital costs for SMRs at around $9,500 per kilowatt, actually higher than large reactors on a per-kilowatt basis</cite>, while <cite index=”65-1”>a large light-water reactor comes in at about $7,800 per kilowatt, or roughly $8,100 once regional cost differences get factored in</cite>. That is not a small gap, and it runs directly against the “smaller is cheaper” pitch.

• South Korea’s SMART SMR: <cite index=”66-1”>projected at $10,000 per kWe, about four times higher than the APR1400, a large reactor from the same country</cite>

• EU comparison: <cite index=”67-1”>the IEA estimates SMR overnight costs around $10,000 per kW in Europe, versus $6,600 per kW for traditional nuclear</cite>

• Academic bottom-up modeling: a detailed light-water SMR study found <cite index=”63-1”>an overnight cost of $4,844 per kW, with a levelized cost of $89.6 per megawatt-hour</cite>, closer to parity but still not obviously cheaper

The pattern across nearly every independent estimate is the same. On a straight per-kilowatt basis, first-of-a-kind SMRs are not cheaper than large reactors. Some estimates put them meaningfully more expensive. 🚨

NuScale’s cautionary tale

If you want a real-world case study instead of a spreadsheet projection, look at what happened to the industry’s most advanced SMR project. It is not pretty, and it is the example every skeptic reaches for first. ⚡

<cite index=”62-1”>NuScale and the Utah Associated Municipal Power Systems announced that costs for their 462-megawatt SMR project had risen dramatically, with the target power price climbing from $58 to $89 per megawatt-hour, a 53% increase</cite>. The construction estimate jumped even harder. <cite index=”62-1”>Estimated construction costs rose 75%, from $5.3 billion to $9.3 billion</cite>, which <cite index=”62-1”>put the project at roughly $20,139 per kilowatt, about as expensive as the Vogtle mega-reactor project in Georgia</cite>. The project was eventually cancelled outright. Vogtle itself is no poster child either. <cite index=”65-1”>Its two new reactors came online in 2023 and 2024 at a total cost exceeding $30 billion, against an original budget of roughly $14 billion, working out to about $13,600 per kilowatt</cite>.

So the uncomfortable truth is this: both SMRs and giant reactors have a track record of blowing their budgets. Small reactors were supposed to be immune to this because factory production avoids the on-site chaos that sank Vogtle. NuScale’s collapse suggests that promise hasn’t been proven yet, at least not at first-of-a-kind scale. 🌱

Where SMRs might actually win

Okay, so the pessimistic case is loud and well documented. But it’s not the whole picture, and dismissing SMR economics entirely would be its own kind of oversimplification. There are a few places where the math genuinely favors small reactors. 📈

• Financing risk: <cite index=”61-1”>capital costs on a dollar-per-kilowatt basis may be higher for smaller units due to lost economies of scale, but there may be real advantages in financing costs from shorter project durations and lower interest exposure</cite>

• Construction speed: one bottom-up analysis found <cite index=”63-1”>a light-water SMR construction duration averaging 4.5 years with a 90% probability of landing between 3.4 and 6.0 years</cite>, versus five-plus years typical for large plants

• Learning curve potential: <cite index=”59-1”>Idaho National Laboratory projects a high-case scenario where deploying 32 SMR units drives a 55.6% reduction in overnight construction cost through manufacturing learning effects</cite>

• Reduced overrun risk: <cite index=”68-1”>a single-unit SMR requires significantly less total on-site labor, and if built by an experienced workforce, it could avoid the cost-overrun risks tied to megaprojects</cite>

That learning curve is really the whole ballgame. Nobody disputes that a tenth reactor off an assembly line should cost less than the first one. The argument is entirely about how fast that curve bends, and whether investors are willing to eat the expensive early losses to get there. <cite index=”60-1”>Wood Mackenzie expects first-of-a-kind SMR costs around $180 per megawatt-hour, dropping 40% to about $100 by 2030 as manufacturing scales up</cite>. That’s a big bet on a curve that, right now, exists mostly on paper.

The FOAK versus NOAK trap

This distinction, first-of-a-kind versus nth-of-a-kind, is where most of the confusion in this whole debate actually lives, and it’s worth being blunt about it. Nearly every rosy SMR cost projection you’ll see is an NOAK number, not a FOAK number. Nearly every real-world cost overrun story, NuScale included, is a FOAK story. Comparing the two is comparing a company’s five-year sales forecast to its first quarter of actual revenue. 🧬

• FOAK reality: high costs, unproven supply chains, first-time regulatory approval, workforce retraining, all the expenses of doing something for the first time

• NOAK promise: standardized designs, established suppliers, trained crews, manufacturing efficiencies compounding across dozens of units

• The gap between them: <cite index=”67-1”>even under the IEA’s optimistic Announced Pledges Scenario, cost parity between SMRs and conventional nuclear might not arrive until mid-century, and under the more conservative Stated Policies Scenario, SMR capital costs stay roughly one-third higher even by 2050</cite>

<cite index=”67-1”>As of 2024, only three SMRs had actually been constructed globally, each in a different country, which is nowhere near enough real-world experience to prove the learning curve theory one way or the other</cite>. Everything past that is projection, and projections in this industry have a rough track record. If you’re trying to separate genuine NOAK progress from optimistic marketing across the dozens of SMR companies now raising capital, that’s precisely the kind of deal-by-deal tracking SMRbrief Pro exists to sort through.

So which one actually wins?

Right now, on hard numbers, large reactors hold a real edge per kilowatt, and anyone telling you SMRs are simply cheaper today is skipping past a lot of inconvenient data. But “today” is doing a lot of work in that sentence. The entire SMR bet is that serial manufacturing, shorter build times, and lower financing risk eventually flip the equation, the same way factory production flipped costs in solar panels and batteries. Whether nuclear behaves like those industries, or whether reactors are just too complex and too regulated to follow the same curve, is genuinely unresolved. 🚀

I lean skeptical of near-term parity, mostly because the industry has promised this before and Vogtle happened anyway. But I also think dismissing the whole SMR category because of one cancelled Utah project is too easy. What would actually change your mind here, a handful of on-time, on-budget NOAK deployments, or is the learning curve theory itself the thing that needs more scrutiny before another dollar of public money goes toward proving it?

Nuclear power has an image problem, and it is not really about meltdowns anymore. Ask most people what worries them about a reactor in their backyard, and the answer usually circles back to one word: waste. ☢️ Small modular reactors get pitched as the fix for almost everything wrong with nuclear power, cheaper, faster to build, safer by design. What gets talked about far less is what comes out the other end. So let’s actually walk through it: what SMR waste is, how much of it there is, where it sits today, and why the honest answer to “where does it go” is still, mostly, nowhere permanent. 🔬

Not all nuclear waste is the same thing

The word “waste” hides a lot of variety, and that matters because each type gets handled completely differently. Regulators sort it into three broad buckets, and understanding them clears up most of the confusion people have about nuclear cleanup. 🧬

• Low-level waste: contaminated gloves, tools, filters, and protective gear. Decays fast and typically goes to a near-surface disposal facility

• Intermediate-level waste: reactor components and resins with higher radioactivity that usually need some shielding but not deep burial

• High-level waste: the spent fuel itself, along with reprocessing byproducts, and this is the stuff that actually keeps regulators up at night

<cite index=”50-1”>Most low-level waste is typically sent to land-based disposal facilities relatively soon after packaging</cite>, and that part of the system genuinely works. The trouble starts with high-level waste, the spent fuel rods pulled straight out of the reactor core, still hot, still radioactive, and still dangerous for a very long time. That fuel <cite index=”48-1”>contains a combination of uranium, plutonium, and fission products that can take somewhere between 10,000 and 100,000 years of decay to match the radioactivity of natural uranium</cite>. Numbers like that are hard to hold in your head, but they are the entire reason this is a policy problem and not just an engineering one. 📊

Do smaller reactors actually make less waste?

Here’s the part that surprises a lot of people, and honestly it surprised me too the first time I read the research. Smaller does not automatically mean less waste. In fact, some of the leading studies say the opposite. 🚨

A widely cited PNAS analysis looked closely at three advanced SMR concepts and found that <cite index=”47-1”>SMRs will likely produce more voluminous and chemically or physically reactive waste than large light water reactors, largely because their smaller cores leak more neutrons</cite>. That is a real physics problem, not marketing spin. Smaller reactor cores lose more neutrons out the edges instead of using them to split fuel, which means less efficient burnup and, in some designs, waste that is trickier to package and store per unit of electricity generated.

• iPWR designs (like NuScale’s): waste volume and chemistry closer to conventional plants, but still scaled down per unit

• Sodium-cooled fast reactors: more reactive waste streams that can complicate long-term packaging

• Molten salt reactors: liquid fuel adds handling complexity, since spent fuel isn’t a neat solid rod you can just pull out

Not everyone agrees this is a fatal flaw. Advocates point out, reasonably, that <cite index=”56-1”>recent studies from Argonne National Laboratory and the National Academies found the amount of spent fuel produced by SMRs and microreactors will be comparable to waste from conventional reactors on a fleet-wide basis</cite>, and that some advanced designs are built specifically to burn down existing waste stockpiles rather than add to them. Oklo is probably the loudest example, marketing its Aurora design as running partly on recycled material. Whether that pans out at commercial scale is still an open question, and it’s one worth watching closely if you’re tracking which designs actually deliver on their waste claims versus which ones are still theoretical. 💡

Where does the waste actually sit right now?

This is the part that tends to shock people who assume there’s a “nuclear waste facility” somewhere quietly handling all of this. There isn’t. Not in the United States, anyway. 🏭

<cite index=”49-1”>The nation has over 90,000 metric tons of spent nuclear fuel from commercial power plants, and DOE is responsible for disposing of it in a permanent geologic repository but has yet to build one, because policymakers have been at an impasse over what to do with this fuel since 2010</cite>. That backlog isn’t shrinking either. <cite index=”49-1”>The amount of spent fuel stored at power plants keeps growing by about 2,000 metric tons a year</cite>, sitting in cooling pools and steel-and-concrete dry casks at reactor sites across the country. <cite index=”51-1”>Those 77 sites span 35 states, and they are increasingly becoming de facto permanent disposal facilities by default, not by design</cite>.

Why? The short version is Yucca Mountain. Congress picked the Nevada site back in 1987, Nevada fought it for decades on safety and political grounds, and the project effectively died under the Obama administration. Nobody has replaced it with a real alternative since. 🌍 A few private companies have tried to fill the gap with consolidated interim storage facilities in Texas and New Mexico, but <cite index=”53-1”>a federal appeals court ruled the NRC lacked authority to license a temporary storage facility not located at a nuclear plant or federal site, nullifying one of those licenses</cite>. So even the stopgap solutions are stuck in court.

How other countries are actually solving this

It’s worth zooming out, because the U.S. isn’t the only country wrestling with this, and some are genuinely further along. This is where the international comparisons get interesting. 🌎

• Finland: the Onkalo repository near Eurajoki is the furthest along of any project worldwide, built 430 meters deep in bedrock roughly 1.8 billion years old

• Sweden: SKB received government approval to construct a deep repository at Östhammar and began surface preparation work in 2025

• Canada: selected the Wabigoon Lake Ojibway Nation-Ignace area in late 2024 as the site for its national repository

• France, Switzerland: both pushing toward license applications for their own deep geologic sites

<cite index=”55-1”>These countries typically work through a stepwise process: a siting decision, then a license to construct, then a license to operate, and history shows the whole sequence can take decades</cite>. The common thread across the successful programs is patience and consent. Sweden and Finland spent years building trust with host communities before construction even started, which is basically the opposite of how the Yucca Mountain fight played out. If you want a running comparison of which countries are actually shipping fuel underground versus just talking about it, SMRbrief Pro members can search and filter that regulatory pipeline directly.

Is there a faster fix coming?

A few paths could ease the pressure sooner than a mined repository ever will, and none of them require waiting fifty years for consensus politics to sort itself out. ⚡

• Deep borehole disposal: companies like Deep Isolation propose lowering waste canisters into narrow, directional boreholes drilled thousands of feet down, avoiding the enormous cost of excavating a mined repository

• Fuel recycling and reprocessing: a recent executive order pushed to restart commercial reprocessing, aiming to shrink both the volume and radioactive lifespan of leftover material

• Transmutation: converting some long-lived radioactive elements into shorter-lived ones, which could shrink the timescale a repository needs to hold waste safely

• Consolidated interim storage: centralizing waste at one or two sites instead of 77 scattered ones, even without a permanent repository yet

None of these fully solves the problem on their own. <cite index=”51-1”>Storing spent fuel in pools and dry casks is safe for decades but not for the millennia needed to truly isolate it from the environment</cite>, so eventually something permanent has to get built, somewhere, with somebody’s consent. That is genuinely the hard part, and it has almost nothing to do with the reactor technology itself. 🔋

So here’s the real question worth sitting with: as SMRs multiply across more sites in more states, does that make the waste problem easier to solve, because smaller, more efficient designs eventually replace an aging fleet, or harder, because now there are more scattered locations each generating their own stockpile with nowhere permanent to send it? I don’t think the industry has a clean answer yet, and I’d bet the states currently courting SMR projects for their data center power demand haven’t fully reckoned with that tradeoff either. Worth asking your local utility commission before the ribbon-cutting ceremony, not after.

Ask five nuclear engineers to name their favorite reactor and you will start a genuinely heated argument. 🔥 That is not a metaphor problem, it is the whole industry in miniature. Small modular reactors are not one technology wearing five outfits. They are five fundamentally different machines competing for the same job: making steady, carbon-free electricity in a box small enough to ship on a truck. Some use water. Some use molten salt. One uses liquid metal that would look at home in a blacksmith’s forge. Understanding which is which matters more than ever, because in 2026 this stopped being a science-fair debate and became a multibillion-dollar race with tech giants, utilities, and governments all picking sides.

The stakes are real. <cite index=”2-1”>The U.S. is leading the world in small modular nuclear reactor development, with 28 siting announcements as of 2026, more than the next four countries combined</cite>. So let’s break down the five reactor families actually competing for that ground, what makes each one tick, and who is betting big on them. ⚡

Light water reactors: the safe, boring, winning bet

Light water reactors, or LWRs, are the reactor world’s equivalent of a Toyota Camry. Not flashy, but everyone trusts them because they have logged decades on the road. 🚗 These designs use ordinary water as both coolant and moderator, the same basic physics running inside nearly every large power plant on Earth today.

That familiarity is the whole selling point. <cite index=”19-1,10-1”>Light water-cooled SMR designs are typically smaller versions of existing large reactor designs that use hydrogen in water as a moderator to slow neutrons and increase the odds of a fission event, and in most cases they run on the same low-enriched uranium fuel already used in U.S. reactors</cite>. Regulators already know how to evaluate this physics, which shaves years off the approval process compared to anything more exotic.

• NuScale Power is the clearest example, with its VOYGR design already carrying NRC design approval, the first SMR to hit that milestone

• Deep Fission is pursuing a borehole-based light water design meant to skip most above-ground construction entirely

• China’s Linglong One (ACP100), a 125 MWe demonstration unit, is on track to become the world’s first land-based commercial SMR

The tradeoff is that light water designs cap out around 300 MWe, and they still need active cooling systems and pressurized components, which is exactly the complexity SMRs were supposed to shrink. Still, being first through licensing counts for a lot, and NuScale is already doing front-end engineering work with Romania’s RoPower on that country’s first SMR plant. 📈 If you want the full breakdown of who’s actually shipping product versus who’s still in the slide-deck phase, SMRbrief Pro tracks every licensing milestone as it lands.

High temperature gas-cooled reactors: heat that does double duty

Swap water for helium and you get a completely different animal. High temperature gas-cooled reactors (HTGRs) run hot, really hot, often above 750°C, which opens doors that water-cooled designs simply cannot reach. 🌡️

These reactors typically use TRISO fuel, tiny uranium kernels wrapped in ceramic and carbon layers that are individually almost indestructible, paired with graphite as a moderator. The payoff is versatility. Because HTGRs can deliver both electricity and industrial-grade process heat, they are a natural fit for the unglamorous but massive markets of steel, cement, hydrogen production, and desalination, not just the power grid. 🏭

• X-energy’s Xe-100 delivers 80 MWe per module and is headed for a four-unit demonstration at Dow’s Seadrift plant in Texas, with construction expected to begin in 2026

• Nano Nuclear Energy is developing a 15 MWe HTGR under a build-own-operate model

• China’s HTR-PM, a pebble-bed design, is one of only two SMRs currently operational anywhere in the world

Amazon has also put money behind X-energy, including backing a four-unit Xe-100 project with Energy Northwest in Washington state. That is not a small tech company hedging its bets, that is a hyperscaler betting on process heat and grid power from the same reactor. The industrial-heat angle is probably the most underrated part of this whole category, honestly. Most coverage fixates on electricity, but decarbonizing a cement kiln is arguably the harder, more valuable problem. 💡

Molten salt reactors: liquid fuel, low pressure

This is where things get genuinely strange, in a good way. Molten salt reactors (MSRs) use, well, molten salt, either as a coolant around solid fuel or, in some designs, as the fuel itself, dissolved directly into the liquid. <cite index=”16-1”>This distinguishes molten salt reactors from designs that use liquid metal, gas, or water as coolants</cite>.

Why bother with something this unconventional? <cite index=”17-1”>Molten salt reactors use molten fluoride salts as primary coolant, at low pressure</cite>, which sidesteps a lot of the engineering headaches that come with keeping water pressurized at high temperatures. And because the fuel can be liquid, <cite index=”16-1”>MSRs can be refueled while operating, essentially performing online reprocessing, while conventional reactors have to shut down for refueling</cite>.

• Terrestrial Energy’s Integral Molten Salt Reactor (IMSR) is a 195 MWe design moving through Canadian regulatory pre-licensing, with roughly $950 million committed

• Moltex is pursuing the Stable Salt Reactor with a partnership tied to New Brunswick Power and Point Lepreau

• Kairos Power, technically a fluoride salt-cooled variant, has a 500 MW deal with Google, one of the biggest single commitments in the sector

Canada in particular has become an MSR hub. It has the regulatory infrastructure, the workforce, and the political will, and it is not chasing headlines the way some U.S. players are. 🌍 The catch with MSRs is that almost none of them have run commercially at scale yet, so cost and reliability numbers are still mostly projections rather than track record.

Fast neutron reactors: skipping the slow-down step

Every reactor discussed so far relies on a moderator to slow neutrons down, because slow neutrons split uranium atoms more reliably. Fast neutron reactors throw that assumption out. <cite index=”13-1”>They use high-energy neutrons to split atoms rather than the slower thermal neutrons used in most commercial power plants, and they’re typically cooled with liquid metal, usually sodium or lead</cite>.

Ditching the moderator has real advantages. <cite index=”13-1”>Liquid metal-cooled fast reactors extract more energy from their fuel and reduce the amount of long-lived waste generated</cite>, which means these designs can burn fuel types, including some reprocessed waste, that conventional reactors cannot touch. ♻️

• TerraPower’s Natrium pairs a 345 MWe sodium-cooled reactor with molten salt energy storage, is under construction in Kemmerer, Wyoming, and has an 8-plant, 2.8 GW deal with Meta

• Oklo’s Aurora uses liquid-metal cooling in a compact 15-75 MWe design, holds a site use permit from the Department of Energy, and just locked in domestic HALEU fuel supply from Centrus for up to five units in southern Ohio

• Newcleo is pursuing a lead-cooled fast reactor, having relocated its headquarters from the UK to Paris while chasing an Italian demonstration by the end of 2026

Oklo, in particular, has become the poster child for this category on Wall Street. Its market capitalization sits near $11 billion even before its first reactor achieves criticality, which tells you how much investors are pricing in future data-center demand rather than current output. 🚀 Whether that bet pays off depends heavily on whether the domestic HALEU supply chain, still dominated by companies like Centrus, can scale fast enough to feed all these fast reactors at once.

Heavy water and other niche designs

Beyond the big four sit a handful of other approaches worth a mention, mostly variants and hybrids rather than wholly separate categories. <cite index=”11-1”>The broader universe of SMR types also includes heavy water reactors</cite>, which use deuterium oxide instead of ordinary water as a moderator, a lineage that traces back to Canada’s CANDU reactors. Microreactors, generally defined as anything under 10 MWe, blur the line further, since some use light water internals while others borrow gas or metal cooling from the categories above. 🔬

• Heavy water designs: better neutron economy, can run on natural uranium, but face a smaller commercial pipeline today

• Microreactors: sub-10 MWe units aimed at remote sites, military bases, and mining operations rather than grid-scale power

• Hybrid concepts: some developers are blending fast-spectrum physics with molten salt coolant, chasing benefits from both categories at once

None of these is likely to dominate the SMR market outright, but they matter because they show the field is still genuinely experimental. Nobody has locked in a winning formula yet. 🧬

So which type actually wins?

Honestly? Probably more than one. Light water designs will likely dominate the first wave of deployments simply because regulators already understand them. Gas-cooled and molten salt designs make more sense for industrial customers who need heat, not just electrons. Fast reactors are the ones chasing the biggest, splashiest data-center deals right now, Meta, Google, Amazon are all in somewhere on this list, which says a lot about where near-term capital is flowing. 📊

What is your read? Are you betting on the tortoise, light water’s regulatory head start, or the hare, fast reactors and their eye-popping hyperscaler contracts? I would genuinely like to hear pushback on this, because the honest answer is nobody, including people running these companies, knows for certain which technology wins the next decade. If you want to go deeper than any single article can, including deal-by-deal financing data and regulatory timelines across all five categories, that is precisely what SMRbrief Pro was built to provide.

One more wrinkle worth watching: fuel. Nearly every advanced design above, gas-cooled, molten salt, fast, leans on HALEU, and <cite index=”30-1”>at present only Russia and China have infrastructure to produce it at scale</cite>, with domestic U.S. supply still ramping. Track that bottleneck, via sources like the U.S. Energy Information Administration or the World Nuclear Association, and you will have a pretty good early signal for which reactor types actually get built this decade versus which stay stuck on the drawing board. For deeper background on the underlying physics, Wikipedia’s SMR overview is a solid starting point.

Every SMR pitch deck opens the same way: smaller reactors, built in factories, stamped out like Toyotas instead of hand-carved like cathedrals. 💰 Cheaper by design, faster to build, easier to finance. It’s a great story, and parts of it are even true. But the nuclear industry has a long, expensive habit of promising one number and delivering a very different one, and the SMR wave so far hasn’t broken that habit. 📉 Here are five real dollar figures, not projections, not vendor slide-deck math, that show what building one of these things actually costs right now.

The number that punctured the hype

$9.3 billion. That’s what NuScale’s flagship US project, the Carbon Free Power Project in Idaho, was projected to cost by the time it collapsed in November 2023, up from an original $5.3 billion estimate just two years earlier. A 75 percent jump in construction cost pushed the target power price from $58 to $89 per megawatt-hour, and that’s before accounting for inflation between now and whenever the plant would have actually opened. 📊 The Utah utilities that had signed up for the power backed out, subscriber by subscriber, until there weren’t enough left to justify finishing it.

Here’s the number that really stings: once you divide that final $9.3 billion by the plant’s planned 462 megawatts, NuScale’s project worked out to roughly $20,139 per kilowatt, according to IEEFA’s analysis of the cost estimate. That’s essentially the same price per kilowatt as the infamously over-budget Vogtle expansion in Georgia, the very poster child for large nuclear’s cost problems that SMRs were supposed to fix. If the factory-built promise can’t beat the thing it’s replacing on a dollar-per-kilowatt basis, what exactly is the pitch? 🤔

A few reasons that estimate ballooned so fast:

• Commodity and steel prices climbed sharply between 2021 and 2023

• Interest rates used in the project’s financing model rose roughly 200 basis points

• The project shrank from twelve modules to six, losing economies of scale it was counting on

• Licensing, safety review, and quality assurance costs kept adding years and dollars

The baseline these projects are trying to beat

To be fair to SMR developers, large nuclear’s own track record isn’t exactly a high bar to clear. 📉 Vogtle Units 3 and 4 in Georgia, the only new nuclear reactors completed in the US in the past three decades, were originally priced at $14 billion. By the time Unit 4 finally came online in 2024, seven years late, the total had climbed past $34 billion, according to Wikipedia’s detailed construction history. That’s more than double the original number, on a project built by an experienced utility with federal loan guarantees behind it. 🏗️

The Vogtle math, in the numbers that actually moved:

• $14 billion: the original 2008 estimate for both new reactors

• $25 billion: the revised total by 2018, after Westinghouse’s bankruptcy

• $34 billion: the final tally once Unit 4 entered service in 2024

• 7 years: how far behind schedule the project ultimately ran

This is the real comparison SMRs need to win, not against solar or wind, but against the nuclear industry’s own history of blown budgets. And so far the early data suggests SMRs are inheriting the same disease rather than curing it. 💸 Have you noticed how every big infrastructure story eventually turns into a story about who pays for the overrun?

The one real number instead of a promise

Not every SMR cost figure is a cautionary tale. 🏗️ GE Hitachi’s BWRX-300 is actually under construction right now at Ontario’s Darlington site, and its current price tag is a real, audited estimate rather than a vendor projection. As of May 2025, the project was estimated to cost at least CAD $7.7 billion, working out to roughly €15,870 per kilowatt, per analysis from the Heinrich Böll Foundation’s Brussels office. That’s still expensive by any normal standard, but it’s a number attached to concrete actually being poured, not a slide in an investor deck. 💵

That distinction matters more than it sounds. A project with:

• A signed construction contract and a utility actually pouring foundations

• Regulatory approval already secured rather than pending

• A single owner (Ontario Power Generation) absorbing the financial risk

• A real, if painful, price tag instead of an aspirational one

is simply a different category of investment than one still selling subscriptions to future power. 📋 If you want the underlying financials tracked deal by deal instead of piecing them together from press releases, SMRbrief Pro keeps a running ledger of what every active project actually costs versus what it originally promised.

The gap that supposedly closes with scale

Here’s the number every vendor is betting the whole business model on: today’s first-of-a-kind (FOAK) SMRs run somewhere between $80 and $150 per megawatt-hour, while vendors promise that Nth-of-a-kind (NOAK) production, once a design has been built dozens of times, drops that to $50 to $80 per megawatt-hour. ⚡ The Department of Energy has floated an even more optimistic long-run figure of around $60/MWh if designs get standardized and mass-produced. 📉

The catch is how many units it actually takes to get there. Industry estimates suggest the cost curve doesn’t flatten out until somewhere between 5 and 7 identical units, or 10 to 20 gigawatts of installed capacity, based on techno-economic modeling published through the Department of Energy. South Korea needed to build 24 nearly identical reactors before it saw meaningful cost reductions on its fleet. 🔁

Right now, the field looks like this:

• NuScale: light-water design, one canceled US project, pursuing Romania and Poland instead

• TerraPower: sodium-cooled Natrium, one unit under construction in Wyoming

• X-energy: high-temperature gas design, backed by Amazon, still pre-construction

• GE Hitachi: BWRX-300, the only design actually pouring concrete, in Canada

Five different designs, five separate learning curves, and not one of them has built a second unit yet, let alone a fifth. 🔢

So the real number that matters isn’t $60 or $80 or $150 per megawatt-hour. It’s the number of identical reactors any single company needs to actually build before its own promised price shows up on a customer’s bill, and right now that number is closer to zero than anyone in the industry likes to admit. Which of these five numbers would you want written into a power contract before signing it?

“Are small modular reactors safe?” is one of those questions that sounds simple and isn’t. ⚛️ Ask a developer and you get a confident yes, backed by physics and diagrams of water flowing downhill on its own. Ask the Union of Concerned Scientists and you get a much more skeptical answer, backed by a different set of numbers. Both sides are working from real data. 🔬 The truth sits in between, and it’s more interesting than either pitch. Here are six specific facts, the reassuring ones and the uncomfortable ones, that actually answer the question instead of just restating it.

The physics behind the safety pitch

Fact 1: Most SMR designs lean on passive safety, meaning the reactor cools itself using gravity and natural convection instead of pumps, motors, and backup generators. NuScale’s version eliminates so many of those active components that the company says it cuts the potential for component failure by several orders of magnitude, according to its own explanation of emergency planning zones. Water heats up, rises, and pulls cooler water in behind it, and that cycle keeps running without anyone touching a switch. 💧

Fact 2: A smaller core genuinely means less radioactive material to begin with, which changes the math on worst-case accidents. Research on postulated accidents across four different SMR technologies found that for designs claiming strong chemical retention of fission products, the criteria used to trigger evacuation weren’t exceeded past about a kilometer from the plant, compared to the miles-wide zones drawn around conventional reactors. That’s not marketing spin. It’s a direct consequence of having a smaller source term, the technical term for how much radioactive material a plant could actually release. ☢️

Those two facts explain almost the entire safety pitch you’ll hear from any SMR vendor, and they’re both grounded in real engineering, not hope. Still, “walk-away safe” is a phrase worth being a little suspicious of, which brings us to the next section.

What the numbers actually show

Fact 3: When researchers actually run the probabilistic models, SMRs do come out ahead. One comparative study puts core damage frequency for SMR designs somewhere between 1×10⁻⁸ and 4.47×10⁻⁷ per reactor-year, against 1×10⁻⁵ to 1×10⁻⁴ for a typical conventional reactor. 📊 That’s roughly one to three orders of magnitude better on paper, which lines up with what the Nuclear Regulatory Commission requires of the existing US fleet: a core damage frequency goal below 1×10⁻⁴ per year. Worth remembering, though, that this metric is a model output, not a measurement, so treat the exact exponent with a healthy grain of salt rather than gospel.

Fact 4: The industry’s push for smaller Emergency Planning Zones (EPZs) isn’t just a safety claim, it’s also a cost argument, and being honest about that matters. Cutting an EPZ down to five miles instead of the standard ten can save developers as much as $50 million per project. That doesn’t make the underlying physics wrong. It does mean the incentive to argue for a smaller zone isn’t purely altruistic, and regulators are still working out exactly how small is actually justified case by case.

A quick way to see how the safety case actually breaks down by design:

• Light-water SMRs (NuScale, Holtec): rely on the same physics as today’s reactors, just with fewer moving parts

• High-temperature gas reactors (X-energy): use fuel pebbles designed to retain fission products even at extreme temperatures

• Sodium-cooled and molten salt designs (TerraPower, Kairos): operate at low pressure, which removes one whole category of accident

• All of them: still need real-world operating history before anyone can say the models were right

If you want to compare those probabilistic numbers design by design instead of taking any single vendor’s word for it, SMRbrief Pro tracks the safety filings and licensing data across every active project. Which of these designs would you trust with a construction permit in your own backyard? 🏗️

Where the skeptics have a real point

Fact 5: Proliferation concerns aren’t fringe griping, they’re coming from credentialed nuclear physicists. Edwin Lyman, the Union of Concerned Scientists’ director of nuclear power safety, has argued alongside researchers from MIT, the Colorado School of Mines, and Princeton that certain high-assay low-enriched uranium fuels used in some SMR designs could shorten the path to a weapon without further enrichment. TerraPower’s Natrium reactor, for example, uses uranium enriched to around 19 percent, well above the roughly 5 percent used in conventional light-water fuel. That’s a real technical debate among people who build this stuff for a living, not internet noise. ☢️

Fact 6: The waste doesn’t disappear, and neither does the need to store it somewhere. Lyman’s research points out that per unit of heat generated, small reactors produce just as much highly radioactive waste as large ones, and some HALEU-fueled designs actually require more mined uranium per kilowatt-hour, not less. Vendor promises to eventually haul spent reactors away are, in his words, not realistic given the current absence of any licensed centralized storage site. Any community hosting an SMR should plan on being a long-term steward of that waste, full stop. 🌍

That’s a genuinely uncomfortable pair of facts sitting right next to two genuinely reassuring ones, and both sets are true at the same time. Nuclear safety debates rarely allow for that kind of nuance, but this one demands it.

So, are SMRs safe

Here’s the honest verdict: the engineering behind passive safety is real, the accident math looks meaningfully better on paper, and regulators are holding these designs to the same or tougher standards than the existing fleet. At the same time, the fuel and waste tradeoffs are unresolved, the proliferation questions are being raised by serious scientists rather than activists with an agenda, and not one commercial SMR has years of operating history to actually validate any of these models yet. NuScale is still the only design with full NRC certification as of 2026. Everything else is a well-reasoned bet.

Safer than yesterday’s nuclear plants on paper? Probably. Risk-free? No technology gets to claim that, and any vendor who tells you otherwise is skipping the parts of the story that don’t fit on a slide. What would actually change your mind on this, a decade of clean operating data, or a specific incident?

Forget the cooling towers. Forget the sprawling site the size of a small town. A small modular reactor, or SMR, is closer in scale to something you’d see parked at a shipyard: a sealed vessel, often built to be shipped by truck or rail, quietly making heat. ⚛️ That image has powered a wall of investor decks and government press releases for years, and in 2026 it’s finally getting tested against something less forgiving than a slide: actual concrete, actual regulators, and actual electricity bills. So what is this thing, really, and how does it turn uranium into power without the footprint of a conventional plant? Here’s the five-minute version. 🔬

What “small” and “modular” actually mean

The two words in the name are doing separate jobs, and mixing them up is where most confusion starts. Small refers to output. ⚡ The International Atomic Energy Agency puts the ceiling at 300 megawatts of electricity per module, roughly a third of what a single conventional reactor puts out. Modular refers to how the thing gets built: in a factory, in standardized pieces, rather than poured and welded together on-site over a decade. 🏗️ The World Nuclear Association frames SMRs around factory fabrication, chasing the economies of series production and shorter construction times, which is genuinely the whole economic bet in a sentence. Build the same box a hundred times in a controlled factory, the theory goes, and you dodge the cost overruns that have plagued one-off mega-projects for decades.

Not every “small” reactor is modular, and the category actually splits into a few tiers:

• Microreactors: under 10 MWe, small enough for a military base or a remote mine

• SMRs proper: 10 to 300 MWe, the sweet spot most companies are chasing

• Large conventional reactors: 1,000+ MWe, the traditional utility workhorse

• Process-heat SMRs: sized and rated in megawatts thermal instead of electric, built to heat industrial facilities rather than push power onto a grid

That last category matters more than people think. Developers are increasingly engineering SMRs to integrate directly with industrial energy systems, acting as nuclear boilers and behind-the-meter generators rather than utility-scale power stations feeding a transmission grid. Refineries, aluminum smelters, and steel plants need heat that solar and wind simply can’t deliver on demand, and that gap is a big part of why SMRs are having a moment. 🏭

The physics: same fission, smarter plumbing

Strip away the marketing and an SMR runs on the exact same principle Enrico Fermi proved out under a football stadium in 1942: split uranium atoms, capture the heat, boil water, spin a turbine. Just like a full-size reactor, the fission heat boils water into steam, and that steam turns a turbine and generator to make electricity. Nothing exotic there. The real engineering happens in what carries that heat away from the core, and this is where SMR designs actually start to diverge from each other and from their bigger cousins. 🔥

Most reactors currently working through licensing are still light-water reactors (LWRs), the same coolant technology that has run the global nuclear fleet for seventy years, just shrunk and simplified. 🌡️ But the pipeline behind them looks noticeably more varied:

• Light-water SMRs (NuScale, Holtec’s SMR-300): water-cooled and water-moderated, the safest bet regulators already understand

• High-temperature gas reactors (X-energy’s Xe-100): helium-cooled, pebble-bed fuel, designed for industrial process heat

• Sodium-cooled fast reactors (TerraPower’s Natrium): liquid metal coolant paired with molten-salt energy storage for load-following power

• Molten salt reactors (Kairos Power, Terrestrial Energy): fuel dissolved directly in the coolant salt, which changes the safety math entirely

Each of these is chasing a genuinely different customer. A utility wants something that slots into the existing grid without drama. A chemical plant wants steady, very hot process heat. A data center operator wants firm power that doesn’t blink when clouds roll in. One reactor type won’t satisfy all three, which is exactly why so many competing designs are still in the running instead of the field narrowing to a winner. 📈 Which of those customers do you think ends up driving the most actual construction by 2030?

Passive safety: why these things are built to babysit themselves

Ask any SMR developer what separates their design from a 1970s-era plant and you’ll get the same answer within thirty seconds: passive safety. 🛡️ The idea is that if something goes wrong, physics handles the emergency instead of a control room full of operators scrambling for backup power. Passive safety systems lean on natural phenomena instead of powered equipment: gravity-fed cooling lines that open automatically during an outage, or natural convection, where a heated fluid rises and pulls cooler fluid in behind it, keeping a cooling cycle running on its own. ⚙️ No diesel generators. No frantic pump restarts. Just water finding its way downhill. 💧

NuScale’s version of this is a good concrete example. According to the Department of Energy, its emergency core cooling valves are simpler than those in traditional reactor designs and open automatically without extra pumps, power, or operator action, which lowers both the failure points and the capital cost. Holtec makes an even bolder claim about its SMR-300, describing it as “walk-away safe” because the fuel stays cooled indefinitely without anyone touching a switch.

That said, “walk-away safe” deserves a raised eyebrow, not a standing ovation. A closer read of the safety literature is more careful:

• Passive systems reduce, but don’t eliminate, the need for human oversight

• Operators still monitor plant status and manage non-safety functions during an event

• Materials used in passive systems need long-term performance validation, since some designs are genuinely new

• Simulator training has to account for how passive systems behave differently than active ones, which regulators are still working through

These designs aim to prevent fuel damage for extended periods without anyone stepping in, but that doesn’t eliminate the need for human response entirely. It’s a real safety improvement, not a magic trick, and the honest version of the pitch says so. If you want the full regulatory paper trail behind each design’s safety case rather than the marketing summary, SMRbrief Pro is where that gets tracked design by design.

Where things actually stand right now

Here’s the part most explainers skip: how far along is any of this, actually? As of mid-2026, NuScale remains the only SMR with full NRC design certification, for its uprated 77 MWe US460 design approved in May 2025, and that certification process itself is worth understanding if you’re wondering how long it actually takes to get a new reactor design blessed by regulators. Behind NuScale, the field is crowded but moving. TerraPower’s Natrium design has a construction permit application working through NRC review for its Kemmerer, Wyoming demonstration plant, and Meta has already signed on for up to eight Natrium plants down the line. X-energy’s Xe-100 is working through review for a Dow Chemical facility in Texas. Kairos Power holds a construction permit for its Hermes test reactor in Oak Ridge, Tennessee. Holtec’s dual-unit SMR-300 project, called Pioneer, cleared an early NRC docketing step in February 2026.

A few data points worth sitting with 📊:

• BWRX-300 (GE Hitachi) is already under construction at Ontario’s Darlington site, ahead of most US competitors

• NANO Nuclear’s microreactor cleared NRC construction-permit acceptance for a University of Illinois deployment in May 2026

• The EU’s SMR Strategy, adopted in March 2026, formally targets SMRs at hard-to-decarbonize industries like petrochemicals and steel

• None of this is fast: TerraPower’s own timeline points to years of review before turbines spin

I’ll be honest about the tension here: the industrial logic for SMRs is strong, the safety engineering is genuinely improved over older designs, and yet the actual build record so far is thin. ⚡ One certified design, a handful of permits, and a lot of promised timelines that keep sliding to the right. That gap between the pitch and the concrete is exactly what to watch over the next two years. 🚀 So the real question isn’t whether SMRs can work. It’s whether any developer can actually ship one on time, on budget, and at the price they promised, before investor patience runs out. What would it take to convince you either way?

Nuclear bans used to feel permanent 🔒. Countries wrote them into law after Chernobyl, after Fukushima, after Three Mile Island, and for decades those laws sat there like settled history. Not anymore. Belgium repealed its phase-out law. Italy is rewriting its constitution-adjacent referendum consensus. Germany’s chancellor calls his own country’s shutdown a strategic mistake, out loud, on the record. The reversal wave sweeping Europe right now isn’t a fringe movement, it’s happening in parliaments, cabinet meetings, and government coalition agreements. Here’s who’s changing course, and why the timing lines up so neatly with rising electricity prices and a war that keeps reminding everyone what energy dependence actually costs 💡.

Belgium: from phase-out to full nationalization in under two years

Belgium might be the fastest, most dramatic reversal on the list. The country’s 2003 phase-out law banned new reactor construction and mandated that all seven of its plants close after 40 years of operation. For two decades that law sat untouched. Then, in a single parliamentary session on May 15, 2025, Belgium’s federal parliament voted 102 to 8, with 31 abstentions, to repeal it entirely, removing both the mandated closure dates and the ban on new construction.

That vote wasn’t the end of the story, it was the opening move. By spring 2026, Prime Minister Bart De Wever announced Belgium intended to nationalize its nuclear fleet outright, negotiating a full takeover of the country’s reactors from French energy company Engie and halting all decommissioning activities immediately. Energy Minister Mathieu Bihet has set an explicit target: 8 gigawatts of nuclear capacity by 2035, split evenly between extending existing reactors and building new ones, including Belgium’s first small modular reactor.

What makes Belgium’s case worth studying:

• Public opinion never actually turned against nuclear the way the 2003 law implied. Polling found 71% of Belgians support continuing to use nuclear power, with only 14% favoring a full phase-out

• Nuclear still supplied over 55% of Belgium’s power mix even as the phase-out clock ran down

• The reversal happened under a five-party “Arizona coalition” government that formed specifically around energy security concerns

• Engie itself resisted the plan, with its CEO calling further extensions “unthinkable,” which is exactly why the government moved toward nationalization instead of negotiation

Have you noticed how often the “public turned against nuclear” narrative doesn’t actually hold up once you look at the polling? Belgium is a pretty clean example of a government’s own law lagging years behind what voters actually wanted.

Italy: reversing a referendum that’s been law since Chernobyl

Italy’s case is arguably the most symbolically loaded reversal on this list, because Italy didn’t just phase out nuclear administratively. Italians voted to ban it, twice, first after Chernobyl in 1987 and again after Fukushima in 2011. Those referendums gave the public direct control over nuclear policy, which is part of why Italy’s about-face feels so significant now.

In October 2025, Italy’s cabinet approved a draft enabling law to begin unwinding the ban. That proposal moved into committee hearings at the Chamber of Deputies starting January 21, 2026, with testimony ranging from small modular reactor designs to the country’s waste-management backlog. Italy currently sits on more than 32,000 cubic meters of radioactive waste in temporary storage, a legacy nobody has fully resolved even as the political mood shifts. At the March 2026 Nuclear Energy Summit in Paris, Environment and Energy Security Minister Gilberto Pichetto Fratin formally committed Italy to a global pledge to triple nuclear capacity by 2050, joining 37 other signatory nations.

The realistic timeline matters here, because it undercuts any sense that this is happening overnight:

• Italy’s national energy plan doesn’t expect nuclear to contribute meaningfully until after 2040

• Site selection would take another 2 to 3 years once legislation passes

• Licensing and construction stretch well into the 2030s before residents see any real infrastructure

• Italy is explicitly skipping the large pressurized water reactors of the 1960s in favor of advanced SMRs and fourth-generation lead-cooled fast reactors

Italy technically bypassed direct public consultation this time around, relying on parliamentary legislation rather than another referendum, which is its own quietly significant detail given how the original bans came about.

Germany: the phase-out its own chancellor calls a mistake

Germany closed its last three nuclear plants in April 2023, completing a phase-out that traces back to a 2011 decision under Angela Merkel following Fukushima. Fast forward to today, and Chancellor Friedrich Merz has publicly called that phase-out a “serious strategic error.” That’s not an opposition talking point, that’s the sitting head of government describing his own nation’s flagship energy policy in those terms.

And yet Merz has also called the phase-out “irreversible,” which sounds contradictory until you look at the coalition math. Merz’s CDU wanted to explore reactivating decommissioned plants before the 2025 election. Once the CDU formed a coalition with the center-left SPD, that plan didn’t survive negotiations. Merz has been blunt about it: <em>”The SPD did not want that and we had to accept that.”</em> Nuclear experts like Rainer Klute dispute the “irreversible” framing on technical grounds, arguing decommissioning is not the same as permanent destruction and that some plants could theoretically return to service more cheaply than new construction.

What’s actually moving in Germany right now, short of a full reactor restart:

• Merz has signaled he’ll withdraw Germany’s opposition to classifying nuclear as “renewable” under EU legislation, a largely symbolic but telling shift

• Germany posted the highest household electricity prices in the entire European Union in the first half of 2025, according to Eurostat, a fact that keeps getting cited in domestic nuclear debates

• Small modular reactors are increasingly floated as the realistic path forward, since traditional large-reactor economics remain politically toxic

• Germany’s only path to a fast reversal would likely require cooperation with the pro-nuclear AfD, which Merz has explicitly ruled out

Germany is the clearest case of a country where elite opinion has shifted well ahead of what the current governing coalition can actually deliver.

Denmark and the quieter reversals

Not every reversal comes with parliamentary drama. Denmark banned nuclear construction back in 1985, and that ban sat almost entirely unchallenged for forty years. Then, in the same week Belgium repealed its phase-out law, two-thirds of Danish MPs voted to approve a formal government analysis into the potential use of nuclear power for energy security. Climate and Energy Minister Lars Aagaard put it plainly: Denmark has essentially no living institutional memory of nuclear power, which is exactly why an analysis needs to happen before anything more concrete can move forward.

Spain offers a different flavor of second-guessing. The government confirmed a full phase-out policy in December 2023, with the first reactor closing in 2027 and the last by 2035. By April 2025, though, the same government indicated it would consider proposals from plant operators to extend those closure dates if any were submitted, a soft walk-back that hasn’t become a formal reversal but keeps the door open. Add in Greece opening public debate on SMRs despite historical seismic concerns, and the pattern across Europe becomes hard to miss: nuclear opposition that once felt like settled consensus is now openly contested inside government, not just outside it.

For anyone trying to track which of these reversals are actually moving through legislation versus which ones are still just political signaling, SMRbrief Pro members can search, filter, and export the full intelligence picture behind developments like this one, which matters a lot when the difference between “committee hearing” and “operational reactor” can span a full decade.

Nuclear energy provides around 23% of the EU’s electricity and roughly half its low-carbon electricity, and European Commission President Ursula von der Leyen has now called the continent’s earlier retreat from nuclear a “strategic mistake.” That’s a remarkable sentence for an EU official to say out loud in 2026. So here’s the real question worth sitting with: are these reversals driven by a genuine reassessment of nuclear’s safety and economics, or mostly by an energy price shock and a war that made everyone nervous about depending on someone else’s gas? Probably some of both, and which one dominates will shape how fast, and how far, this reversal wave actually goes.

If you still think of nuclear power as a Republican cause and renewables as a Democratic one, the data hasn’t agreed with you for a while now 📊. Both parties have shifted, sometimes loudly, sometimes quietly, and the old script barely applies anymore. That doesn’t mean everyone agrees. It means the disagreements have moved to new ground: how fast to build, who pays, how much regulatory risk is acceptable, and whether speed and safety can really coexist on the timelines being promised. Here’s an honest look at where each side actually stands, not where the stereotypes say they should.

The bipartisan baseline nobody talks about enough

Start with the number that surprises people most: support for nuclear power has grown in both parties over the past several years. Pew Research found that roughly seven in ten Republicans and Republican leaners favor expanding nuclear power, compared to about half of Democrats and Democratic leaners, a 17-point gap that is actually smaller than the partisan splits on offshore drilling or coal mining. Support has grown by double digits on both sides since 2020, with Pew’s polling showing Democratic favorability up 15 points and Republican favorability up 16 points over that stretch.

A few facts worth sitting with:

• Democrats officially endorsed nuclear energy in their party platform for the first time since the Nixon administration, a shift that took roughly 48 years to happen

• The bipartisan ARC Act, reintroduced in the Senate by Jim Risch (R-Idaho) and Ruben Gallego (D-Ariz.), aims to speed up advanced reactor deployment through joint DOE and Nuclear Regulatory Commission action

• Younger Republicans are actually less supportive of expanding nuclear power than their older peers, a reversal of the usual pattern on energy issues 🤔

• Men remain far more likely than women to favor expanding nuclear power in both parties, a gap that has nothing to do with partisanship at all

None of this means the parties are indistinguishable. It means the fight isn’t really “nuclear yes or no” anymore. It’s about how.

Where Republicans have put their weight

The current Republican approach is best described as deregulate first, build fast, and worry about the details later 🚀. President Trump signed four executive orders on nuclear energy in May 2025, aiming to quadruple U.S. nuclear capacity from around 100 gigawatts today to 400 gigawatts by 2050. That target alone would require building more nuclear capacity in the next 25 years than the country built in the previous 70.

The specifics get aggressive fast. The orders directed the Department of Energy to create a pilot program letting advanced reactor designs bypass standard NRC review entirely, with a target of at least three reactors reaching criticality outside national laboratories by July 4, 2026. Eleven reactor designs from ten companies entered that pilot, and Valar Atomics reportedly reached cold criticality as an early milestone. The administration also directed DOE to designate AI data centers as critical defense facilities, a framing that opens the door to faster siting and permitting for reactors built specifically to power them.

Other threads in the Republican playbook:

• An “all-of-the-above” energy framing that pairs nuclear expansion with continued support for natural gas and fossil fuels

• Legislation aimed at building a domestic uranium supply chain and banning Russian uranium imports, tying nuclear policy explicitly to energy independence

• A push for at least 20 new international nuclear cooperation agreements to help U.S. reactor exporters compete globally

• Consistent skepticism toward wind and solar subsidies, framed as picking “winners and losers” in the energy market

The catch, and it’s a real one, is that speed and safety don’t automatically travel together. NPR reported in early 2026 that the Trump administration quietly rewrote nuclear safety rules, including changes to long-standing radiation exposure principles, to keep pace with the July 2026 criticality deadline. Career staff at DOE’s nuclear office reportedly asked universities for volunteers just to keep up with the review workload. That’s not a partisan talking point, it’s a documented operational strain, and it’s worth taking seriously regardless of which side of the aisle you’re on ⚠️.

Where Democrats have actually moved

The Democratic shift is less flashy but arguably more structural. Energy Secretary under the Biden administration, Jennifer Granholm, called for building a new wave of conventional nuclear reactors, a more aggressive stance than anything the Obama administration proposed. Billions in federal support flowed toward both conventional nuclear and the advanced reactor industry during that period, sometimes with Republican cooperation and sometimes without it.

The tension inside the party is real, though, and it hasn’t gone away. Traditional environmental groups remain firmly opposed. The Sierra Club still calls nuclear a “uniquely dangerous energy technology for humanity” on its own website, and that opposition has shaped which Democratic officials feel comfortable championing nuclear loudly versus quietly. Nuclear Regulatory Commission nominees under Democratic administrations have drawn criticism from pro-nuclear advocates for maintaining the status quo rather than pushing advanced reactor licensing forward more aggressively.

A few things define the current Democratic position:

• Framing nuclear primarily as a climate solution rather than an energy independence or national security one

• Continued internal split between labor and industrial Democrats who back nuclear jobs and environmental-wing Democrats who don’t

• Support for nuclear tends to pair with, not replace, continued investment in wind, solar, and grid modernization

• Less appetite for the kind of regulatory bypass the current administration has pursued, with more emphasis on reforming the NRC process itself rather than routing around it

It’s worth being honest here: the Democratic coalition hasn’t fully resolved its own internal disagreement about nuclear the way Republicans have converged on “build faster.” That ambiguity shows up in funding decisions, in NRC appointments, and in how loudly individual Democratic lawmakers are willing to advocate for it.

The real dividing lines in 2026

If you strip away the rhetoric, the actual disagreements between the parties land in three places. First, pace versus caution. Republicans have generally accepted more regulatory risk in exchange for speed, betting that faster deployment matters more than getting every safety review exhaustive on the first pass. Democrats, even the pro-nuclear wing, tend to want deployment to move through more traditional oversight channels, even if that means a slower buildout.

Second, market structure. Republicans generally favor letting nuclear compete without picking winners through subsidy, while still backing federal loans and fuel supply chain investment when framed as national security. Democrats are more comfortable with direct government support and see it as one tool among several, alongside renewables, rather than a replacement for them.

Third, and this one gets underdiscussed, is waste and disposal policy. Neither party has meaningfully resolved what happens to spent nuclear fuel long-term, and that unresolved question sits quietly underneath every optimistic capacity target either side proposes. If you’re trying to track how these policy currents actually translate into specific project approvals, financing deals, and regulatory decisions rather than just campaign rhetoric, SMRbrief Pro turns the nuclear intelligence in this article into a searchable, filterable, always-updated resource you can act on.

What does bipartisan support actually buy the SMR industry if the two parties can’t agree on how fast is too fast? That’s the question worth watching heading into the next election cycle, more than any platform plank either party puts on paper.

Every splashy SMR announcement, the groundbreaking, the site permit, the reactor render with clouds behind it, rests on a document most people never read: the power purchase agreement. Strip away the ribbon-cuttings and the deal is simple. Somebody has to promise to buy the electricity before anybody will spend billions building the plant that makes it. ⚡ That promise is the PPA, and in the small modular reactor world right now, it is doing more work than almost anything else in the industry.

What a power purchase agreement actually is

A PPA is a contract between a power generator and a buyer that locks in the price, volume, and duration for electricity sold over a fixed period, often 15 to 25 years. That’s the whole concept. No mystery, no fancy structure required. But in nuclear, and especially in SMRs, the PPA carries weight that goes well beyond a simple sales contract.

Here’s why: SMRs are capital-intensive up front and cheap to run once built 🔋. A developer might spend billions on construction and licensing, then earn the money back slowly over decades of operation. Banks and private equity won’t fund that kind of bet on a hope. They want a buyer locked in before ground breaks. A signed PPA is the piece of paper that turns “we think someone will buy this power” into “here is a legally binding commitment for the next 20 years.” That’s the difference between a reactor that gets financed and one that stays a rendering on a website.

Features that show up in almost every SMR-related PPA:

• A fixed or escalating price per megawatt-hour, sometimes tied to inflation

• A committed volume of power, often ramping up over several years as the plant reaches full output

• A duration long enough to match the debt repayment schedule, typically 15 to 20+ years

• Provisions for what happens if the plant is delayed, underperforms, or shuts down early

• Options to extend, expand, or adjust volume as the buyer’s needs change

Why AI and data centers turned PPAs into front-page news

For decades, PPAs were a fairly boring instrument used by utilities to buy wind and solar output. Then the AI boom hit, and suddenly the biggest buyers of nuclear electricity weren’t utilities at all. They were hyperscalers 🖥️ desperate for round-the-clock, carbon-free power to run data centers that never sleep.

The clearest example is the restructured agreement Talen Energy signed with Amazon Web Services in June 2025, a 17-year, roughly $18 billion power purchase agreement covering up to 1,920 megawatts of carbon-free electricity from the Susquehanna nuclear plant. What makes this deal worth studying isn’t just the size, it’s the structure. The contract runs through 2042 with a staged ramp-up: deliveries are expected to hit somewhere between 840 and 1,200 megawatts by 2029, climbing to the full 1,680 to 1,920 megawatts by 2032. That ramping schedule matters because it lets Amazon match its power draw to its actual data center build-out, while giving Talen a revenue floor to plan around.

There’s an SMR angle buried in the fine print too. Talen and Amazon agreed to jointly explore building new small modular reactors within Talen’s Pennsylvania footprint, on top of pursuing capacity uprates at the existing plant. So the PPA isn’t just paying for today’s electrons, it’s laying groundwork for tomorrow’s reactors 🏗️.

A few other things worth flagging about this shift:

• Deals like this moved from behind-the-meter setups, where power flows directly to the data center, to front-of-the-meter arrangements that route through the grid

• Meta signed a 20-year nuclear PPA with Constellation Energy in mid-2025 to keep Illinois nuclear capacity running for its AI operations

• Constellation separately signed a 20-year PPA with Microsoft covering the entire output of the revived Three Mile Island plant, expected back online in 2028

• Hyperscalers are increasingly willing to underwrite risk that utilities used to shoulder alone

PPAs versus contracts for difference: two ways to de-risk a reactor

The PPA isn’t the only tool in the de-risking toolbox, and it helps to know the difference. A contract for difference, or CfD, works almost like an insurance policy layered on top of a market sale. Under a CfD, the electricity gets sold on the open power exchange, but a government or counterparty agrees to cover the gap between the market price and an agreed floor price if the market price falls short. It’s a government-flavored cousin of the PPA, and the UK has leaned on it heavily.

The UK government has said it will offer a similar mechanism to stabilize revenue for new nuclear projects and pull in private capital, with Rolls-Royce SMR’s Wylfa site already selected to host the country’s first small modular reactor. The logic is straightforward. Nuclear construction risk is enormous and unpredictable 📉, so governments step in with a revenue guarantee that private lenders alone won’t provide.

A quick comparison of the two structures:

• PPA: a direct contract between generator and specific buyer, price and volume fixed by negotiation

• CfD: the sale happens on the open market, with a third party (often government) topping up or clawing back the difference against a strike price

• PPAs dominate in the U.S. and among corporate buyers like Amazon and Microsoft

• CfDs show up more where governments are directly financing new nuclear buildout, as in the UK

Cost certainty is often worth more to nuclear developers than raw cost competitiveness, which is why mechanisms like power price escalation, contracts for difference, and tax credit monetization all exist to rebalance risk and make projects financeable in the first place. The PPA is simply the version of that logic showing up most often in American SMR deals right now.

Have you noticed how often “financing” ends up being the real story behind a reactor announcement, more than the technology itself? It usually is.

Why lenders treat the PPA as the linchpin

Ask anyone who finances power plants for a living and they’ll tell you the same thing: revenue certainty beats almost everything else on the checklist 💰. There’s still very little real-world validation of SMR cost estimates, so lenders and equity investors remain wary of the technology’s economic viability. That skepticism doesn’t disappear because a company has a slick reactor design. It softens, though, when there’s a signed PPA on the table showing a creditworthy buyer committed to purchasing output for two decades.

Think of it from a bank’s perspective. A reactor is a 20-year bet with a construction phase full of things that can go wrong. If data center demand shifts geographically, or if cheaper renewables paired with battery storage undercut PPA pricing, the whole business case for an SMR project can weaken fast. A strong PPA doesn’t eliminate that risk, but it puts a floor under it. That’s why:

• Project financing almost never closes before a PPA (or an equivalent mechanism) is signed

• The creditworthiness of the buyer, not just the generator, gets scrutinized heavily

• Lenders often require minimum contract lengths that match debt amortization schedules

• Early termination and force majeure clauses get negotiated line by line, because they determine who eats the loss if things go sideways

This is also where the market has gotten genuinely interesting to watch. Traditional utilities used to be the only credible long-term buyers. Now Amazon, Microsoft, and Meta are essentially acting as sovereign-grade counterparties, and that’s changing what lenders consider “safe” for a nuclear project. Whether that shift holds up if AI capital spending ever cools is one of the more honest open questions in the sector right now, and I don’t think anyone has a confident answer yet.

What this means for the next wave of SMR deals

Every company chasing SMR deployment, whether it’s Rolls-Royce SMR, NuScale, X-energy, Kairos, or Oklo, is ultimately chasing the same prize: a buyer willing to sign a long enough, big enough PPA to unlock financing. SMR development in Western countries is proceeding largely through private investment now, reflecting a broader shift from government-funded nuclear research toward private-sector-led deployment aimed at affordable, carbon-free energy, according to the World Nuclear Association. That shift only works if the revenue side of the equation gets locked down first.

Watch for a few patterns going forward:

• More staged, ramping PPA structures like the Talen-Amazon deal, rather than single lump-sum commitments

• Growing use of hybrid structures blending PPA elements with CfD-style price floors, especially outside the U.S.

• Corporate buyers negotiating SMR-specific clauses that fund new reactor construction, not just purchase existing output

• Utilities re-entering the space as intermediaries between hyperscalers and generators, since not every tech company wants to become its own retail electricity provider

For anyone trying to separate genuine momentum from press-release noise in this sector, the fastest filter is simple: is there a real PPA behind the announcement, and who actually signed it? SMRbrief Pro members can search, filter, and export the full intelligence picture behind developments like this one, which makes that filtering job a lot faster than combing through SEC filings yourself.

So next time a headline touts a new SMR project, skip past the rendering and go looking for the contract. What does the PPA actually say about price, volume, and who’s on the hook if it slips? That’s where the real story lives.