BasicsConcernsEvery meltdown argument ends up as a fight over three place names. Three Mile Island in 1979 was an equipment fault plus confused operators. Chernobyl in 1986 was a reactor with a physics flaw. Fukushima Daiichi in 2011 was a tsunami that drowned the backup power. Different failures need different fixes, and lumping them together as “nuclear accidents” is the first myth, the one that makes every later debate worse.
SMRs do not remove risk. As SMRbrief has reported on the data center energy push, no commercial SMR has yet sold a watt to a paying customer in the United States. The designs below are approved or under review, not proven over decades of operation. But five common meltdown assumptions do break down when applied to designs from NuScale, X-energy and Rolls-Royce SMR. Here is where each one stands as of late September 2026.
Myth 1: a power failure means a meltdown
The three Fukushima Daiichi units that melted were not undone by the earthquake. They shut down on schedule. Then a 15-metre tsunami flooded the site, disabled 12 of 13 backup generators and knocked out the heat exchangers that dump decay heat to the sea, according to the World Nuclear Association’s account. Fission stopped. Heat did not. Without pumps, the cores boiled dry.
Large plants need electricity to survive their own shutdown. The NuScale Power Module is built to need none. Each 77 MWe module sits in a water-filled pool, and NuScale says it shuts down and cools itself with no operator action, no AC or DC power and no added water, as described in a design paper in Frontiers in Energy Research. The NRC approved that uprated design in May 2025, per the Department of Energy. SMRbrief’s glossary of SMR terms notes NuScale is still the only company holding an NRC design certification or standard design approval.
Reactor: NuScale Power Module, pressurized water, 250 MWt (77 MWe) per module
Plant configuration: VOYGR-6, six modules, 462 MWe
Regulator: US NRC, standard design approval, May 2025
Most advanced overseas project: RoPower in Romania, up to six modules at the former Doicesti coal site with Nuclearelectrica (June 2025 reporting, so check for updates)
Passive is not the same as invulnerable. Gravity and density differences are weak forces, and the passive safety literature flags that weak driving forces can make these systems harder to prove in the first hours of an accident. That is why regulators test them instead of taking vendors at their word. Insurers are paying attention too, and SMRbrief covered how passive designs change the tail risk that pushed underwriters out of nuclear in the first place.
Myth 2: fuel always melts when cooling fails
Conventional fuel is uranium oxide in metal tubes. Overheat it and the cladding reacts with steam to make hydrogen, which is what blew apart the outer buildings at Fukushima. X-energy’s Xe-100 attacks the fuel itself. It runs on TRISO particles: a uranium kernel wrapped in three layers of carbon and ceramic, packed into billiard-ball-sized pebbles and cooled by helium. The Department of Energy says TRISO fuel cannot melt in a commercial high-temperature reactor. Idaho National Laboratory backed that with a test that held irradiated TRISO at up to 1,800 degrees Celsius for more than 300 hours, beyond predicted worst-case accident conditions, with full retention of fission products.
The flagship project is Long Mott Generating Station: four Xe-100 modules, 800 MWth and 320 MWe, at Dow’s Seadrift Operations in Calhoun County, Texas. The NRC finished its environmental assessment in May 2026 with a finding of no significant impact, and the final safety evaluation is targeted for November 2026, per POWER’s reporting on the NRC dashboard. That date may have moved, so check the NRC docket.
Reactor: Xe-100, helium-cooled high-temperature gas reactor, 80 MWe per module
Fuel plant: TX-1 in Oak Ridge, Tennessee, which received the first-ever NRC Part 70 HALEU fuel fabrication license in February 2026
Fuel timeline: TX-1 operations expected in the first half of 2028, per X-energy’s July 2026 NRC briefing
Open item: the 13-month irradiation test of TRISO-X pebbles at INL started in November 2025, so qualification data for this exact fuel is still coming
The safest fuel is also the priciest. X-energy’s illustrative lifetime cost for a four-reactor Xe-100 plant runs $1.66 billion to $2.18 billion, with $1.095 billion to $1.25 billion of it fuel, per a February 2026 amended registration statement. Would you pay that premium for a core that does not melt?
Myth 3: operators have to save the day
Three Mile Island is remembered as a technology failure, but the Department of Energy’s summary pins it on equipment failure and operator error together. A relief valve stuck open. The control room indicator said it had closed. Operators, unaware they were losing coolant, made choices that made it worse. About one-third of the core melted, yet the vessel and containment held and the accident left no detectable health impacts, according to the World Nuclear Association.
A feedwater failure tripped the turbine and the reactor, as designed
The relief valve opened as designed, then stuck open for more than two hours
Instruments told operators the valve had reclosed
Confused decisions in the control room turned a manageable transient into a partial meltdown
The design goal of passive systems is to take those decisions out of the first hours. Cooling that runs on natural circulation does not wait for an operator to diagnose anything. The worst thing a confused operator can do is nothing, and in a passive design that is the correct response. Modern designs have added passive systems specifically to cut the risk of compounding human error, a trend the same passive safety literature documents.
Operators still matter for security, maintenance and everything outside the emergency window. The claim is narrow: the first response does not depend on someone reading a misleading gauge correctly at 4 a.m.
Myth 4: any reactor can run away like Chernobyl
Chernobyl’s RBMK pairs a graphite moderator with boiling water coolant. When steam forms, less neutron-absorbing water remains, and reactivity rises. That is a positive void coefficient, and at Chernobyl it pushed power to around 100 times the rated capacity, per the World Nuclear Association’s RBMK appendix. Light-water designs work the opposite way. Boil the water and the reaction slows, because the water is the moderator.
Here the honest answer is that SMRs inherit this fix rather than invent it. Every Western light-water reactor already has a negative void coefficient. What SMRs add is variety that removes the mechanism entirely.
NuScale module: light-water PWR, water is both moderator and coolant
Rolls-Royce SMR: 470 MWe PWR, in Step 3 of the UK’s Generic Design Assessment
Xe-100: helium coolant, so there are no steam voids to form at all
RBMK: graphite plus boiling water, a combination found in no other power reactor
The myth still earns a place on this list because it drives more public opinion than any engineering paper. Next time someone calls an SMR a smaller Chernobyl, ask which moderator they mean. Most conversations end there.
Myth 5: every plant needs a 10-mile evacuation zone
Existing US plants plan around a plume exposure zone of about 10 miles and an ingestion zone of about 50 miles, per an NRC emergency preparedness presentation. In November 2023 the NRC finalized 10 CFR 50.160, effective December 18, 2023. It lets SMR and advanced reactor applicants size the plume zone to the accident consequences and prove readiness through performance-based drills. Where the zone stays on site, local first responders are not required to join radiological exercises.
NuScale went first. The NRC’s Advisory Committee on Reactor Safeguards concurred with staff that NuScale’s methodology gives the same protection as a 10-mile zone and can produce a zone at the site boundary, per Nuclear Engineering International. The approval covers only the NuScale design.
Legacy rule: roughly 10-mile plume zone for large light-water plants
New rule: scalable zone, effective December 18, 2023
Payoff: plants can sit next to industrial customers, housing or a former coal site
Open question: a 2023 NRC rulemaking document noted the agency has never licensed a commercial plant with a site-boundary zone, and FEMA and state agencies raised concerns
This is the myth with the most work left. A smaller zone needs a site-specific case, and neighbors will want to see it. Every project, deal, company, and regulatory event we cover is logged, verified, and searchable in SMRbrief Pro, including each of these zone decisions as they land.
The next checkpoints are concrete. The NRC’s final safety evaluation for Long Mott’s four Xe-100 modules is targeted for November 2026, and the UK’s assessment of Rolls-Royce SMR’s 470 MWe design is expected to finish by the end of 2026, per the Office for Nuclear Regulation. Both dates may have slipped, so check the regulators’ pages. When those decisions land, which of the five myths do you expect to die first, and which will survive anyway?