Background: nuclear marine propulsion history
Nuclear marine propulsion uses the controlled fission of uranium-235 (U-235) or plutonium-239 (Pu-239) atoms in a moderated reactor core to generate heat, which is then used to raise steam in a steam generator and drive a steam turbine, geared to the propeller shaft or coupled to an electric generator for an integrated electric propulsion plant. The energy density of the fission process is approximately 80 million MJ/kg of U-235 fully fissioned, against 42 MJ/kg for VLSFO on a chemical-energy basis. The factor of approximately two million between fission and chemical energy density is the structural argument for nuclear marine propulsion: a reactor core of 50 to 200 kg of U-235 equivalent supports 5 to 12 years of full-power operation between refuelling, against 7 to 10 days for a conventional bunkering cycle on a deep-sea diesel vessel.
The first nuclear-powered ship was the United States Navy submarine USS Nautilus (SSN-571), commissioned in 1954 with a Westinghouse S2W pressurised-water reactor. The military fleet has expanded since then to include approximately 160 nuclear-powered submarines and aircraft carriers in service across the United States, Russia, the United Kingdom, France, China and India. The military reactor fleet operates under each navy’s national regulatory regime and is not subject to the IMO civilian framework that this article examines.
The civilian merchant nuclear fleet has been much smaller and has progressed in fits and starts. The four historic civilian newbuilds (NS Savannah, Otto Hahn, Mutsu, Sevmorput) demonstrated the technology in the 1960s through the 1990s with mixed economic and political outcomes. The Russian (and prior Soviet) state has operated a continuous civilian nuclear-icebreaker fleet under Atomflot since the launch of Lenin in 1959, with the current fleet comprising six operational nuclear icebreakers including the new Project 22220 Arktika class. The 2024-2025 wave of commercial SMR-marine programmes is the first credible attempt to revive civilian nuclear-merchant shipping at deep-sea scale since the decommissioning of Otto Hahn in 1979, with the principal commercial drivers being the IMO net-zero framework, the FuelEU Maritime trajectory, and the cost trajectory of competing zero-carbon fuels.
The case for nuclear marine propulsion rests on four structural advantages. First, the fission process produces zero TtW carbon dioxide because the energy source is a nuclear reaction rather than a chemical combustion. Second, the energy density of fission fuel is six orders of magnitude higher than chemical fuels, which eliminates the bunker-tank size penalty that affects hydrogen, ammonia and bio-LNG lifecycle pathways. Third, the refuelling interval of 5 to 12 years eliminates the bunker-supply chain dependency and the geographical constraint on bunker availability. Fourth, the lifecycle WtW intensity of centrifuge-enriched LEU marine fuel is among the lowest of any propulsion option, comparable to the best-case green-hydrogen pathway.
The case against rests on five considerations. The capital cost per installed kilowatt is several times higher than a conventional propulsion plant, with an Aframax-equivalent SMR estimated at USD 200 to 500 million against USD 80 to 100 million for a conventional VLSFO/methanol newbuild. The regulatory framework is incomplete: the IMO civilian regime is provided by Resolution A.491(XII) of 1981, which is widely regarded as outdated, and the IGF Code is silent on nuclear. Most flag states have no domestic regulation for civilian nuclear-powered merchant ships beyond the warship exemption framework. Port-state acceptance of nuclear-powered merchant calls is limited to a small set of ports globally (estimated at approximately 15 ports, principally Russian Murmansk plus a handful of others). The nuclear-waste disposal infrastructure for spent fuel and decommissioned reactor pressure vessels is operated by the producing state and adds a long-tail liability that is difficult to internalise in commercial shipping economics.
Historic civilian nuclear-merchant ships: NS Savannah, Otto Hahn, Mutsu, Sevmorput
Four civilian nuclear-merchant ships have operated to date, with three out of service and one in continuous operation since 1988.
NS Savannah was the first civilian nuclear-powered ship, commissioned in 1962 by the United States Maritime Administration as a demonstration project of the Atoms for Peace programme. The vessel was a 596 foot (182 metre) cargo-passenger ship of approximately 22,000 tonnes displacement, with a Babcock and Wilcox pressurised-water reactor of 74 MWt thermal output driving a steam turbine of approximately 22,000 shaft horsepower. The reactor used 4.4 percent enriched uranium dioxide fuel in a single core, with the original core supporting approximately 300,000 nautical miles of operation between refuelling cycles. NS Savannah was operated as a demonstration vessel until 1972 (with cargo service from 1965), and was retired from service due to the high crewing cost (a dual nuclear-and-marine crew was required) and the limited cargo earnings of the demonstration cargo configuration. The vessel is preserved at the Port of Baltimore and is one of three United States National Historic Landmark ships.
Otto Hahn was the German civilian nuclear-merchant ship commissioned in 1968 by the German Atomic Research Society (GKSS) and the Hapag-Lloyd shipping line. The vessel was a 172 metre, 16,800 tonne ore carrier with a 38 MWt advanced pressurised-water reactor of approximately 11,000 shaft horsepower. Otto Hahn operated commercial ore-carrying voyages between Northern Europe, West Africa, South America and the United States from 1968 to 1979, accumulating approximately 650,000 nautical miles. The vessel was decommissioned in 1979 and converted to a conventional diesel-engine ore carrier, primarily because of the operational cost and the port-acceptance constraint: many West African and South American ports refused entry to a nuclear-powered ship, which forced operational compromises that eroded the commercial case.
Mutsu was the Japanese civilian nuclear-merchant ship commissioned in 1974 by the Japan Nuclear Ship Development Agency. The vessel was a 130 metre, 8,200 tonne research and demonstration ship with a 36 MWt pressurised-water reactor. Mutsu had a troubled operational history: the reactor leaked radiation on its initial sea trial in 1974, and the vessel was unable to enter port for several months due to local opposition before reaching Mutsu port for repair. The ship returned to service in 1990 after extensive modifications, completed four research voyages between 1990 and 1992, and was retired in 1995. Mutsu is the cautionary tale in the civilian nuclear-merchant fleet history, illustrating the fragility of public acceptance and the regulatory complexity of civilian nuclear-marine operation.
Sevmorput is the Russian civilian nuclear-powered LASH (Lighter Aboard Ship) carrier, commissioned in 1988 and the world’s only continuously operating nuclear-powered commercial cargo vessel as of 2026. The vessel is 260 metres long, 32,000 tonnes deadweight, with a single KLT-40 pressurised-water reactor of approximately 135 MWt thermal output and 39,500 shaft horsepower. Sevmorput is operated by Rosatomflot for Arctic supply runs and occasional commercial voyages, with the principal route covering the Murmansk to Petropavlovsk-Kamchatsky run via the Northern Sea Route. The vessel was refuelled in 2003 and again in 2013, and the current operational plan extends to 2034. Sevmorput is the only operational reference for civilian nuclear-merchant cargo operation and is the basis for several technical lessons that inform the 2024-2025 SMR-marine programme designs.
The four historic civilian newbuilds delivered cumulative operational experience of approximately 50 ship-years and 2.5 million nautical miles, with no major reactor incidents in commercial service (the Mutsu sea-trial leak occurred during commissioning rather than commercial operation). The technical envelope was demonstrated successfully, and the constraints that ended civilian nuclear-merchant programmes were principally economic, regulatory and political rather than technical.
Nuclear marine reactor types: PWR, KLT-40S, RITM-200, SMR
Marine nuclear propulsion reactors are predominantly pressurised-water reactors (PWR), with several variants tailored to marine duty cycles and shipboard space envelope.
Pressurised-water reactor (PWR) is the dominant marine reactor type for both military and civilian applications. The reactor uses light water as both the moderator and the primary coolant, pressurised to approximately 150 bar to prevent boiling in the primary loop. The fuel is low-enriched uranium dioxide (UO2) for civilian applications (4 to 20 percent U-235) or highly enriched uranium (HEU, above 20 percent and up to 93 percent for some military applications). The PWR design is mature, with more than 60 years of operational experience and a well-understood safety case, and is the basis for the United States Navy submarine and carrier fleet, the Russian civilian icebreaker fleet, and the historic NS Savannah, Otto Hahn and Mutsu programmes. The thermal efficiency at the steam-turbine boundary is approximately 30 to 33 percent, comparable to a conventional steam plant.
KLT-40S is the Russian commercial pressurised-water marine reactor used in the Atomflot icebreaker fleet from the Taymyr class onwards. The reactor delivers approximately 150 MWt thermal output and 35 MWe electrical at the turbine boundary, with a fuel cycle of approximately 4 to 5 years between refuelling on a 18.6 percent enriched UO2 fuel. The KLT-40S design is based on the earlier KLT-40 used in Sevmorput and the Lenin reactor, with refinements for shipboard reliability and cold-environment operation. The KLT-40S is also the reactor design of the floating nuclear power plant Akademik Lomonosov, commissioned in 2020 and providing 70 MWe of electricity to the Pevek port in the Russian Far East.
RITM-200 is the new Russian integrated pressurised-water marine reactor used in the Project 22220 Arktika-class icebreakers. The reactor is an integrated design with the steam generators, pressuriser and primary pumps housed within a single pressure vessel, which reduces the reactor compartment volume by approximately 50 percent versus the KLT-40S. The RITM-200 delivers 175 MWt thermal output and 35 MWe electrical at the turbine boundary, with a fuel cycle of 7 to 10 years between refuelling on 20 percent enriched UO2 fuel. Each Project 22220 vessel carries two RITM-200 reactors for a combined 350 MWt thermal output and 60 MWe of installed propulsion power. The RITM-200 is the technical basis for the floating nuclear power plant successor to Akademik Lomonosov and is the leading reference for the integrated-PWR approach to commercial marine SMR.
Small modular reactors (SMR) is a generic category covering reactors typically below 300 MWe electrical output, with modular factory-fabrication and shipboard or floating-platform deployment as the principal commercial cases. Marine SMR designs draw on three reactor families: integrated PWR (RITM-200, NuScale, Holtec, BWX Technologies), molten-salt reactor (Core Power UK partnered with TerraPower), and lead-cooled fast reactor (Newcleo Italy partnered with Fincantieri). The thermal efficiency at the turbine boundary is approximately 30 to 35 percent for integrated PWR, 40 to 45 percent for molten-salt reactor with a high-temperature steam cycle, and 40 to 45 percent for lead-cooled fast reactor with a supercritical CO2 or steam cycle. The SMR-marine designs at the 2026 design horizon are predominantly in the 30 to 80 MWe range, sized for an Aframax tanker or a 14,000 TEU container ship at single-reactor configuration.
The reactor-type choice for a marine application interacts with the fuel-cycle and waste-cycle profile. PWR reactors with LEU fuel produce spent fuel that requires geological disposal but does not require reprocessing, with the spent-fuel handling infrastructure available at most nuclear-fuel-cycle states. Molten-salt reactors offer the potential for online refuelling and reduced spent-fuel volume, but the fuel-cycle infrastructure is at pilot scale only. Lead-cooled fast reactors offer the potential for spent-fuel reprocessing into mixed-oxide (MOX) fuel and a closed fuel cycle, but the reactor-design experience base is much smaller than PWR. The 2024-2025 commercial marine SMR programmes are therefore weighted toward the integrated-PWR family, with the molten-salt and lead-cooled fast reactor approaches as longer-horizon options.
Nuclear fuel cycle: U-235 enrichment, refuelling intervals
The nuclear fuel cycle for marine propulsion comprises five stages: uranium mining, uranium conversion, uranium enrichment, fuel fabrication, and reactor loading. Each stage contributes to the lifecycle WtW intensity, with the enrichment stage typically dominating the lifecycle electricity demand and the consequent emissions on a fossil-grid enrichment site.
Uranium mining extracts natural uranium ore from underground or open-pit mines, or in-situ recovery (ISR) operations that leach uranium from porous host rock through injection wells. The principal uranium-producing states in 2026 are Kazakhstan (approximately 40 percent of global supply), Canada, Australia, Namibia, Niger, Uzbekistan and Russia. The lifecycle emissions of uranium mining and milling are approximately 2 to 4 gCO2eq per MJ of reactor energy on a typical ore-grade basis (0.1 to 0.2 percent U3O8), with the figure rising to 6 to 10 gCO2eq per MJ for low-grade African and Australian ores below 0.05 percent.
Uranium conversion converts the milled uranium oxide (yellowcake, U3O8) to uranium hexafluoride (UF6), the gaseous form required for enrichment. The conversion is a chemical process using fluorine and is approximately 99 percent efficient. The lifecycle emissions of conversion are approximately 0.5 to 1 gCO2eq per MJ of reactor energy on a typical conversion-plant electricity intensity.
Uranium enrichment increases the U-235 fraction in the uranium from the natural 0.7 percent to the marine-fuel target enrichment. Civilian marine reactors use LEU at 4 to 20 percent enrichment: NS Savannah at 4.4 percent, Otto Hahn at 3.5 percent, Mutsu at 3.2 percent, Sevmorput KLT-40 at 90 percent (a Russian exception), KLT-40S at 18.6 percent, and RITM-200 at 20 percent. Naval reactors typically use HEU at 20 to 93 percent enrichment for compactness, with the United States Navy at the high end of the range (97 percent in the Virginia-class submarine fleet) and the French and Chinese navies at the lower end (5 to 30 percent in some designs). The 20 percent threshold is the IAEA-defined boundary between LEU and HEU, with weapons-relevant control implications above 20 percent.
Fuel fabrication processes the enriched UF6 back to UO2 powder, presses the powder into ceramic pellets, sinters the pellets at approximately 1,750 degrees Celsius, loads the pellets into zircaloy fuel rods, and assembles the rods into fuel bundles for reactor loading. The lifecycle emissions of fuel fabrication are approximately 0.3 to 0.7 gCO2eq per MJ of reactor energy.
Reactor loading and operation is the use stage of the fuel cycle. The reactor is loaded with the fresh fuel bundles, brought to criticality, and operated at full power for the design fuel-cycle duration. The fuel cycle for marine reactors is materially longer than for shore-based commercial nuclear power plants because of the higher enrichment and the design optimisation for refuelling-free operation: 7 to 12 years for marine SMR designs at 20 percent LEU enrichment, against 18 to 24 months for typical 4 percent LEU shore-based PWRs. The operational stage produces zero TtW CO2 because the energy source is fission rather than chemical combustion.
The refuelling interval is the principal operational difference between nuclear marine propulsion and conventional bunkering. A conventional VLSFO-fuelled deep-sea vessel takes on bunkers every 7 to 10 days at intervals across the operational route, with the bunker logistics chain spanning more than 100 ports globally. A nuclear-powered marine vessel takes on fresh fuel every 5 to 12 years at the producing state’s reactor-loading facility, with the fuel logistics chain limited to a small set of approved facilities (currently approximately 5 globally for civilian marine fuel: Russia, France, the United Kingdom, the United States and China). The nuclear-fuel logistics simplification is one of the operational arguments for marine SMR, but the geographical constraint on fuel supply is one of the regulatory constraints.
Atomflot fleet: Yamal, 50 Let Pobedy, Arktika, Project 22220
FSUE Atomflot is the Russian state operator of the civilian nuclear-icebreaker fleet, headquartered at Murmansk and reporting to the State Atomic Energy Corporation Rosatom. The fleet at end of 2025 comprises six operational nuclear-powered vessels (five icebreakers and one cargo carrier) and four Project 22220 newbuilds in various stages of construction or sea trials.
Yamal is a 1992-commissioned Arktika-class icebreaker (Project 10521 design) of 23,500 tonnes displacement, with two KLT-40 reactors of 171 MWt thermal each and a combined 75,000 shaft horsepower. The vessel is one of the principal workhorses of the Atomflot fleet and conducts approximately 30 to 40 escort operations per year on the Northern Sea Route. Yamal is recognisable for the painted shark-mouth bow that has been retained since the 1990s. The vessel has been refuelled twice (1998 and 2010) and is scheduled for a third refuelling in 2027.
50 Let Pobedy (50 Years of Victory) is the largest Arktika-class icebreaker built, completed in 2007 to a stretched Project 10580 design at 25,800 tonnes displacement, with the same twin KLT-40 reactor configuration as Yamal. The vessel performs Arctic tourism cruises in addition to escort duty, with approximately 8 to 12 tourist cruises per year to the geographic North Pole during the summer season.
Vaygach and Taymyr are 1990-1989-commissioned Taymyr-class shallow-draft icebreakers of approximately 20,000 tonnes displacement, with a single KLT-40M reactor each at 171 MWt thermal and 50,000 shaft horsepower. The shallow-draft design allows operation in the river estuaries and shallow coastal waters of the Russian Arctic, principally the Yenisei and Ob rivers.
Arktika is the lead vessel of the Project 22220 class, commissioned in 2020 at 33,500 tonnes displacement with two RITM-200 reactors of 175 MWt thermal each. The vessel is the largest and most powerful nuclear icebreaker ever built, with a combined 60 MWe of installed propulsion power and the capability to break ice up to 2.8 metres thick at continuous speed. The Project 22220 design uses dual-draft operation: full draft for ocean cruise to maximise propulsion efficiency, and shallow draft for river-estuary operation by ballasting the vessel up. The first three Project 22220 vessels (Arktika, Sibir, Ural) are in operational service as of 2026, with Yakutia at sea trials and Chukotka at the construction stage.
Sevmorput is the LASH cargo carrier described in the historic-fleet section, classified administratively under Atomflot and operating commercial cargo voyages on the Northern Sea Route under Rosatomflot direction.
The Atomflot fleet is the largest civilian nuclear-marine operation globally and is the only continuously operating civilian nuclear-marine fleet at scale. The cumulative operational experience of the Atomflot fleet exceeds 250 ship-years and 30 million nautical miles, with no major reactor incidents in commercial service. The fleet is the principal technical reference for the integrated-PWR marine design family and is the basis for the Russian export of the floating nuclear power plant concept (Akademik Lomonosov class) and the proposed RITM-200 marine licence to third-country operators.
Northern Sea Route nuclear icebreaker traffic
The Northern Sea Route (NSR) connects the Atlantic and Pacific oceans through the Russian Arctic seas (Barents, Kara, Laptev, East Siberian, Chukchi), with a length of approximately 5,600 kilometres against the equivalent Suez Canal route of approximately 11,000 kilometres for a Hamburg-Yokohama voyage. The NSR is governed by the Russian Northern Sea Route Administration (NSRA), which issues mandatory transit permits and provides icebreaker escort services through Atomflot.
The annual cargo throughput on the NSR has grown from approximately 4 million tonnes in 2014 to approximately 36 million tonnes in 2023, with the Russian government targeting 80 million tonnes by 2030 and 150 million tonnes by 2035. The cargo mix is dominated by Russian energy exports (LNG from the Yamal LNG and Arctic LNG-2 plants, crude oil from Novy Port and Prirazlomnaya, and metallurgical concentrate from Norilsk) and bulk cargo on the western leg (Murmansk-Sabetta-Dudinka). The eastern transit traffic to Asia has grown from approximately 5 transits in 2014 to approximately 60 transits in 2023, with the principal trade lanes being LNG to China and Japan and crude to India.
The icebreaker escort regime requires NSR transits in winter and shoulder seasons (October to June) to be escorted by an Atomflot nuclear icebreaker, with the escort fee structured by vessel class, cargo type and ice-class certification. Independent transit by ice-class 1A vessels is permitted in the summer season (July to September) on segments of the NSR with thinner ice. The icebreaker escort regime is the principal source of revenue for Atomflot beyond the Russian state subsidy, and the Project 22220 fleet expansion is sized to support the targeted cargo growth.
The interaction between the NSR icebreaker fleet and the Polar Code is operationally significant. The Polar Code requires polar-class certification for vessels operating in polar waters, with the Polar Ship Certificate documenting the ice-class certification, the equipment standard, and the operational limits. The Atomflot nuclear icebreakers operate as Polar Class 2 (PC2) under the IACS Polar Class system, which is the highest ice-going class assigned to a non-icebreaker. The Polar Code interaction with the IMO civilian nuclear regime under A.491(XII) is a niche regulatory question that the under-revision IMO instrument is expected to clarify.
MEPC.391(82) treatment of nuclear (currently absent)
The IMO MEPC.391(82) Lifecycle GHG Intensity of Marine Fuels Guidelines (2023 LCA Guidelines) provide default emission factors per fuel and per production pathway for the principal marine fuels covered by the IMO net-zero framework. The current text of MEPC.391(82) Annex 1 provides defaults for VLSFO and MGO, LNG (Otto and Diesel cycles), LPG, methanol grades, ammonia grades, hydrogen grades, bio-LNG, HVO, FAME and e-fuels. Nuclear marine propulsion is not listed as a fuel pathway in the current text.
The absence of nuclear from MEPC.391(82) reflects three structural points. First, the LCA Guidelines were drafted in the 2022-2023 window when the active marine-fuel pathway discussion was concentrated on chemical fuels and biofuels, with nuclear treated as a separate civilian-aviation-style category outside the standard marine bunkering framework. Second, the methodology of MEPC.391(82) is built around the bunker-manifold delivery concept, where a defined mass of fuel with a defined LCV is delivered to the vessel and consumed in a propulsion plant. The nuclear-fuel-cycle delivery of a fuel core every 5 to 12 years does not fit naturally into the bunker-manifold accounting structure. Third, the IMO civilian nuclear regulatory regime under Resolution A.491(XII) is separate from the MARPOL Annex VI framework that hosts the LCA Guidelines, and the cross-instrument coordination has not yet been completed.
The current absence of a nuclear pathway in MEPC.391(82) is under discussion at the IMO MEPC and the Sub-Committee on Pollution Prevention and Response (PPR), with the principal questions covering the scope (whether to include civilian nuclear merchant ships under the MARPOL Annex VI lifecycle framework or to maintain the separate A.491(XII) instrument), the methodology (how to translate the fuel-core lifecycle into a per-MJ bunker-manifold-equivalent intensity), and the verification (how to integrate the IAEA safeguards regime with the MARPOL verification regime). The IMO MEPC has scheduled the question for substantive discussion at MEPC 84 (mid-2026), with a possible draft amendment to MEPC.391(82) to add a nuclear pathway by 2027 to 2028.
The absence of a nuclear pathway has practical compliance consequences for any near-term nuclear-powered merchant newbuild. A nuclear-powered vessel that operates internationally cannot be assessed against a defined GFI attained value under the IMO net-zero framework because there is no Annex 1 default for the propulsion energy. The flag-state and class-society approval framework has been used to bridge the gap on the safety side under A.491(XII), but the GHG-compliance side remains open. The 2024-2025 commercial SMR-marine programmes are therefore proceeding on a Class-society approval-in-principle basis with the expectation of MEPC.391(82) amendment in the late 2020s to support the operational compliance regime.
Typical WtW for nuclear: lifecycle accounting
The lifecycle WtW intensity of nuclear marine propulsion is calculated on the same lifecycle accounting basis as other marine fuels, with the energy delivered to the propeller shaft (or to the ship-electrical bus in the integrated-electric configuration) as the denominator and the lifecycle emissions across the fuel cycle as the numerator. The lifecycle stages include uranium mining and milling, uranium conversion, uranium enrichment, fuel fabrication, reactor operation, spent-fuel handling, geological disposal, and reactor decommissioning.
The IPCC AR6 Working Group III lifecycle assessment of nuclear electricity provides the reference figures for the upstream and downstream stages. The AR6 median lifecycle emissions for nuclear electricity are approximately 12 gCO2eq/kWh, with a 5th-to-95th percentile range of approximately 5 to 110 gCO2eq/kWh that reflects variation across uranium-ore grade, enrichment technology, and end-of-life decommissioning assumptions. Translated to gCO2eq per MJ of reactor energy on a 33 percent thermal-efficiency basis, the median figure is approximately 4 gCO2eq/MJ of reactor heat, or approximately 12 gCO2eq/MJ of shaft energy at the propeller boundary.
The breakdown across the fuel-cycle stages, on a typical centrifuge-enrichment LEU pathway, is approximately:
- Uranium mining and milling: 2 to 4 gCO2eq/MJ shaft energy
- Uranium conversion: 0.5 to 1 gCO2eq/MJ
- Uranium enrichment (centrifuge): 1 to 3 gCO2eq/MJ
- Fuel fabrication: 0.3 to 0.7 gCO2eq/MJ
- Reactor operation: approximately 0 gCO2eq/MJ at the reactor boundary
- Spent-fuel handling and storage: 0.5 to 1.5 gCO2eq/MJ
- Geological disposal (long-term): 0.5 to 1 gCO2eq/MJ amortised over the fuel cycle
- Reactor decommissioning: 0.5 to 2 gCO2eq/MJ amortised over the operational life
The aggregate WtW intensity for centrifuge-enriched LEU marine fuel is approximately 5 to 15 gCO2eq/MJ at the shaft boundary, comparable to the best-case green-hydrogen pathway and well below conventional fossil fuels.
The diffusion-enrichment legacy pathway is materially worse. Gaseous-diffusion enrichment plants (the United States Paducah and Portsmouth plants, decommissioned in the early 2010s, and the French Eurodif plant, decommissioned in 2012) consumed approximately 2,400 kWh of electricity per Separative Work Unit (SWU), against approximately 50 to 65 kWh/SWU for modern centrifuge plants. The lifecycle emissions of diffusion-enriched LEU produced on a fossil-grid enrichment site (the Paducah plant was supplied principally by coal-fired electricity at approximately 700 gCO2eq/kWh) reach approximately 25 to 50 gCO2eq/MJ of reactor energy. The diffusion-enrichment pathway is no longer in commercial operation as of 2026, but the legacy accounting matters for the assessment of older nuclear vessels (Sevmorput, the Russian icebreaker fleet, NS Savannah, Otto Hahn, Mutsu) that drew at least part of their fuel inventory from diffusion-enriched stocks.
Diffusion vs centrifuge enrichment
The two principal uranium-enrichment technologies are gaseous diffusion (legacy, decommissioned by the mid-2010s) and gas centrifuge (current commercial standard).
Gaseous diffusion uses a cascade of porous-membrane stages to enrich the uranium-235 fraction by exploiting the slightly faster diffusion rate of the lighter U-235 hexafluoride molecule versus the heavier U-238 hexafluoride molecule. The enrichment factor per stage is small (approximately 1.0043), so a typical commercial enrichment cascade requires several thousand stages connected in series, with substantial inter-stage compression and cooling. The electricity demand is approximately 2,400 kWh per SWU at the cascade boundary, with additional demand for the support utilities (cooling, compression, gas handling). The diffusion-enrichment plants in commercial service were the US Department of Energy plants at Oak Ridge, Paducah and Portsmouth (United States, decommissioned 2010-2013), the French Eurodif plant at Tricastin (decommissioned 2012), the Russian Verkh-Neyvinsky plant (decommissioned 1990s), and the Chinese Lanzhou plant (decommissioned 1980s).
Gas centrifuge uses high-speed rotation to separate U-235 hexafluoride from U-238 hexafluoride by exploiting the small mass difference. The centrifuge spins at approximately 50,000 to 100,000 rpm in a vacuum chamber, with the heavier U-238 migrating to the outer wall and the lighter U-235 concentrating along the rotation axis. The enrichment factor per stage is much larger than diffusion (approximately 1.3 to 1.6 per centrifuge), so a commercial cascade requires only a few hundred centrifuges connected in series and parallel groups. The electricity demand is approximately 50 to 65 kWh per SWU at the cascade boundary, approximately 40 times less than diffusion. The current commercial centrifuge-enrichment plants are URENCO (UK, Netherlands, Germany, USA), Orano (France, replacement for Eurodif), Rosatom Tenex (Russia), CNNC (China), and JNFL (Japan).
The transition from diffusion to centrifuge is essentially complete on the global commercial scale as of 2026. The difference in electricity consumption per SWU translates directly into the lifecycle WtW intensity: a centrifuge-enriched LEU fuel has approximately 1 to 3 gCO2eq/MJ of enrichment-stage emissions, while a diffusion-enriched LEU fuel has approximately 15 to 35 gCO2eq/MJ of enrichment-stage emissions (depending on the grid intensity at the enrichment site).
The marine-fuel implication is that any nuclear-marine programme launched in 2026 onwards uses centrifuge-enriched LEU and benefits from the reduced enrichment-stage intensity, while the legacy fleet that drew on diffusion-enriched fuel inventory carries a higher historical lifecycle intensity. The Atomflot icebreaker fleet has been on centrifuge-enriched fuel since the 2000s, after the closure of the Soviet diffusion plants. The Sevmorput LASH carrier and the NS Savannah, Otto Hahn and Mutsu cargo ships drew on diffusion-enriched fuel during their operational lives.
FuelEU Annex II current and proposed treatment
The FuelEU Maritime Regulation (EU) 2023/1805 implements the lifecycle GHG-intensity methodology for the EU and EEA voyage scope, with Annex II providing default WtW emission factors per fuel and pathway. The current text of Annex II mirrors the MEPC.391(82) Annex 1 structure and lists defaults for the principal chemical-fuel pathways (VLSFO, MGO, LNG, LPG, methanol, ammonia, hydrogen, bio-LNG, HVO, FAME, e-fuels). Nuclear marine propulsion is not listed as a fuel pathway in the current text.
The absence of nuclear from FuelEU Annex II reflects the parallel structure with MEPC.391(82): the EU LCA framework was drafted in 2022-2023 with reference to the IMO LCA framework, and the IMO framework did not include nuclear at the time. The EU Industrial Decarbonisation Strategy (Clean Industrial Deal, 2025) and the European Council conclusions on nuclear energy (2024) have added nuclear to the EU’s industrial decarbonisation toolkit, with the implication that the next FuelEU revision (scheduled for 2030) is likely to add a nuclear-marine pathway to Annex II.
The proposed treatment under discussion at the European Commission and Member State level is to add nuclear as a separate Annex II category with a defined default emission factor based on the IPCC AR6 lifecycle figure of approximately 5 to 15 gCO2eq/MJ for centrifuge-enriched LEU at modern PWR or SMR reactor design. The default would apply to civilian nuclear-powered merchant ships operating under EU port calls and would be subject to the same verifier audit and chain-of-custody documentation as the chemical-fuel pathways. The discussion at the EU level is interlocking with the IMO discussion at MEPC.391(82) revision, with a target of harmonised inclusion in both frameworks by 2028 to 2030.
The RFNBO multiplier under FuelEU Article 5(7) does not currently apply to nuclear because nuclear-powered electricity is excluded from the RED III renewable-energy definition. The proposed FuelEU revision under discussion does not propose to extend the RFNBO multiplier to nuclear, but to introduce a separate low-carbon fuel category that includes nuclear and other zero-carbon non-renewable pathways. The compliance arithmetic for a nuclear-powered vessel under the proposed framework would treat the nuclear energy at the AR6 lifecycle intensity, with no multiplier and no RED III certification requirement.
IMO Code of Safety A.491(XII) 1981 and revision
The IMO regulatory framework for civilian nuclear-powered merchant ships is provided by Resolution A.491(XII), Code of Safety for Nuclear Merchant Ships, adopted at the IMO Assembly in November 1981. The Code is the standing instrument for the safety of civilian nuclear-merchant ships and covers the reactor design, the radiological safety case, the safety assessment report, the operational requirements, the emergency preparedness, and the port-state interaction. The Code was developed in the context of the NS Savannah, Otto Hahn and Mutsu programmes and reflects the technical envelope of the late 1970s pressurised-water reactor designs.
The Code is widely regarded as outdated as of 2026 for several reasons. First, the technical envelope has shifted: modern integrated PWR (RITM-200, NuScale), molten-salt reactor and lead-cooled fast reactor designs were not anticipated in the 1981 text, and the safety-case structure for these reactor families differs from the loop-PWR baseline of the Code. Second, the regulatory architecture has shifted: the 1981 Code predates the modern goal-based standards approach used in the IGF Code, the IGC Code and the LSA Code, and the prescriptive provisions of A.491(XII) do not align with the modern functional-requirement framework. Third, the radiological emergency preparedness landscape has evolved: the post-Fukushima emergency-response framework under the IAEA Convention on Early Notification of a Nuclear Accident and the Convention on Assistance in the Case of a Nuclear Accident is not reflected in the 1981 Code text. Fourth, the operational envelope has shifted: the modern SMR-marine designs envision unmanned reactor compartments with remote operation, which is not anticipated in the 1981 manning provisions.
The IMO has initiated a revision of A.491(XII) through the Sub-Committee on Carriage of Cargoes and Containers (CCC) and the Sub-Committee on Ship Design and Construction (SDC). The work was added to the IMO biennial agenda at MSC 108 (May 2024) with a target completion at MSC 112 (mid-2028). The principal scope of the revision covers (a) modernisation of the safety-case framework to a goal-based standard structure, (b) inclusion of the new reactor families (integrated PWR, MSR, LFR), (c) alignment with the IAEA safeguards and emergency-preparedness regime, (d) integration with the IGF Code and the IGC Code, (e) provision for unmanned reactor operation, and (f) coordination with the MARPOL Annex VI lifecycle framework on the GHG-compliance side.
The revised A.491 instrument is expected to be a self-standing IMO Code rather than a chapter of SOLAS or MARPOL, mirroring the structure of the IGF Code and the IGC Code. The instrument is expected to apply to civilian nuclear-powered merchant ships of any flag operating in international waters, with port-state acceptance of nuclear-powered ships remaining at the discretion of the port state under UNCLOS Article 22 and Article 23 (the innocent-passage and transit rules, which do not bind a port state to accept a nuclear-powered ship at port).
IGF Code silence on nuclear
The IMO IGF Code (International Code of Safety for Ships using Gases or other Low-flashpoint Fuels) covers the safety case for ships using gases or low-flashpoint fuels other than oil-based fuels. The Code has Part A (general requirements), Part A-1 (specific requirements for natural gas as fuel) and Part B (the framework for additional fuels through Interim Guidelines, with separate circulars for methanol, ammonia and LPG, and the under-development hydrogen amendments).
The IGF Code is silent on nuclear fuel as a covered category. The Code’s defined scope is gases and low-flashpoint chemical fuels, which excludes by definition the fission-fuel category that nuclear marine propulsion uses. The implication is that there is no current applicable IMO regulatory framework for new commercial nuclear-marine vessels under the IGF Code; the framework is provided by the parallel A.491(XII) instrument as discussed in the prior section.
The under-revision A.491 instrument is expected to be coordinated with the IGF Code on cross-cutting issues such as the engine-room safety case, the hazardous-area classification, the fire-fighting protocol and the crew training and certification. The two instruments will operate in parallel with cross-references, similar to the current parallel operation of the IGF Code (chemical fuels) and the IGC Code (liquefied gas carriers as cargo). The cross-instrument coordination is being managed at the IMO Sub-Committee on Carriage of Cargoes and Containers, with the SDC and the SSE Sub-Committees providing input on specific technical areas.
The class-society approval-in-principle framework currently bridges the gap on the safety side. The principal class societies have published nuclear-marine and SMR-marine rule sets that draw on A.491(XII), the IAEA safety standards, the national civilian-nuclear regulatory framework of the producing state, and the relevant IGF Code provisions for the non-nuclear systems. The approval-in-principle approach allows newbuild designs to progress through preliminary design and detailed design ahead of the formal IMO instrument entry into force, with the expectation that the final approval will be issued under the revised A.491 once that instrument enters force in the late 2020s.
2024-2025 SMR commercial programs (Core Power, Newcleo, HD Hyundai, FN Group)
The 2024-2025 commercial SMR-marine programme pipeline is the most active period of civilian nuclear-merchant development since the 1970s. Four programmes anchor the pipeline.
Core Power is a UK-based nuclear-marine technology developer founded in 2018 with a focus on molten-salt reactor (MSR) marine propulsion and floating nuclear power. Core Power is the lead partner in the m-MSR (marine molten-salt reactor) programme together with TerraPower (Bill Gates-backed advanced reactor developer), Southern Company (US utility) and Orano (French nuclear fuel cycle company). The m-MSR design uses chloride-salt fuel at approximately 600 to 700 degrees Celsius, with a thermal efficiency of approximately 45 percent at the supercritical-CO2 turbine boundary. The marine application is a 30 to 100 MWe nuclear-electric Aframax tanker concept developed in partnership with Ulstein (Norwegian shipbuilder) and announced in 2023, with a target first delivery in the 2032-2035 window subject to regulatory approval. Core Power has secured Lloyd’s Register approval-in-principle for the m-MSR concept and is engaged with the UK Maritime and Coastguard Agency on the flag-state regulatory pathway.
Newcleo is an Italian-UK developer of lead-cooled fast reactor (LFR) technology, founded in 2021 by Stefano Buono. The reactor design is a 30 MWe LFR with mixed-oxide (MOX) fuel using reprocessed plutonium and depleted uranium, sized for shipboard or floating-platform deployment. Newcleo announced a strategic partnership with Fincantieri (Italian shipbuilder) in 2023 covering the development of marine and floating-power applications of the LFR design. The principal commercial pathway for Newcleo is a 200 MWe LFR for shore-based commercial nuclear power (target first-of-a-kind operation in 2031 in France), with the marine and floating applications as derivatives of the shore-based design. Newcleo has secured Bureau Veritas approval-in-principle for the marine-application reactor design.
HD Hyundai Heavy Industries (HHI) is the Korean shipbuilder that received Lloyd’s Register approval-in-principle in 2024 for a 14,000 TEU SMR-powered container ship concept. The reactor is a 30 to 70 MWe integrated PWR design developed by HHI in partnership with USNC (Ultra Safe Nuclear Corporation, US-based SMR developer) and TerraPower for an alternative reactor option. The container-ship concept uses dual-reactor configuration for redundancy and continuous operation during refuelling, and is sized for transpacific liner trade between East Asia and the US West Coast. The HHI programme is the most commercially advanced of the SMR-container-ship programmes and is targeted for first delivery in the 2030 to 2032 window.
FN Group is a Greek shipowner consortium that announced the Mr. Yi nuclear-powered Aframax tanker concept in 2024, with American Bureau of Shipping (ABS) class engagement and a 30 MWe SMR reactor (specific reactor type subject to selection between three candidates: NuScale, BWX, and HHI integrated PWR). The Mr. Yi concept is the first publicly announced nuclear-powered Aframax tanker by a Greek shipowner, signalling Greek shipping’s interest in the SMR-marine pathway alongside the more established LNG and methanol pathways. The FN Group programme is at concept-design stage and is targeted for first delivery in the 2033 to 2035 window.
The four programmes are commercially distinct but technically interlocking. The integrated-PWR family (HD Hyundai, FN Group with NuScale or BWX, and the Russian RITM-200) is the dominant near-term reactor choice, with the molten-salt and lead-cooled fast reactor families (Core Power, Newcleo) at longer-horizon development. The class-society approval-in-principle framework provides a common technical basis across the programmes, with Lloyd’s Register, Bureau Veritas, ABS, DNV, ClassNK and RINA all developing nuclear-marine and SMR-marine rule sets in parallel.
Class society approval-in-principle (LR, BV, ABS)
The class society approval-in-principle (AiP) framework is the principal technical-approval pathway for nuclear-marine newbuild designs in the absence of a formal IMO instrument. AiP is a structured process under which the class society reviews the design concept against its rules and the relevant IMO and national regulatory framework, and issues a written statement that the design can in principle be approved subject to the resolution of identified open items.
Lloyd’s Register (LR) has been the most active class society on SMR-marine AiP, with public AiP issuances for HD Hyundai Heavy Industries (2024, 14,000 TEU container ship with SMR), Core Power (2024, m-MSR Aframax with Ulstein), and several confidential AiP for owners on integrated-PWR designs. LR’s nuclear-marine rule set draws on its long history with naval reactor support (UK Royal Navy submarine fleet) and its civilian nuclear shore-based experience.
Bureau Veritas (BV) has issued AiP for Newcleo’s LFR marine application (2024) and is engaged with several Mediterranean and Northern European owners on SMR-marine concepts. BV’s rule set is organised around the goal-based standard structure used in the IGF Code and is positioned for alignment with the under-revision A.491 instrument.
American Bureau of Shipping (ABS) has the longest US naval-reactor classification history through US Navy nuclear-vessel work and has issued AiP for the FN Group Mr. Yi concept (2024) and for several US operators on offshore floating-power applications. ABS’s rule set is organised around the US NRC (Nuclear Regulatory Commission) framework on the nuclear side and the ABS marine framework on the ship side.
DNV has published a position paper on SMR-marine and has rule development under way, with engagement on Norwegian and Northern European concepts. DNV’s framework draws on its experience with the Norwegian floating power experience (including offshore floating wind and floating LNG) and its civilian nuclear shore-based experience.
ClassNK has rule development under way for the Japanese Nuclear Marine Initiative (a public-private programme to develop a Japanese SMR-marine concept following the Mutsu legacy), and RINA has rule development for Italian and Mediterranean concepts including the Newcleo programme partner work.
The AiP framework is a bridge to formal instrument approval rather than a substitute for it. The AiP statement does not authorise commercial operation of the vessel; that authorisation is provided by the flag-state administration based on the IMO instrument framework (currently A.491(XII), expected revised A.491 from the late 2020s) and the national civilian-nuclear regulatory framework of the flag state. The AiP-to-final-approval pathway typically takes 3 to 7 years for the full design and construction cycle, with the flag-state engagement starting from the AiP stage.
Regulatory hurdles: flag states, port-state acceptance, UNCLOS Article 23
The regulatory hurdles for civilian nuclear-marine commercial operation are concentrated in three areas: flag-state regulation, port-state acceptance, and UNCLOS interaction.
Flag-state regulation is the primary regulatory framework for any commercial nuclear-marine operation. The flag state issues the certificate of registry, the safety certificates and the safety-management certificates, and is responsible for the overall regulatory regime for vessels under its flag. For nuclear-powered merchant ships, the flag state must additionally have a domestic civilian-nuclear regulatory framework that covers reactor licensing, operational safety, radiological emergency preparedness, and decommissioning. The set of flag states with a domestic civilian-nuclear regulatory framework that supports merchant-ship operation is small as of 2026: Russia (extensive framework supporting the Atomflot fleet), the United States (limited framework, principally for naval reactors), the United Kingdom (limited framework, principally for naval reactors and the Sellafield reprocessing site), France, and a few others. The major commercial flag states (Liberia, Marshall Islands, Panama, Bahamas, Singapore) do not have a domestic civilian-nuclear regulatory framework as of 2026, which means that a commercial nuclear-powered merchant ship cannot register under those flags without a substantial regulatory development effort.
Port-state acceptance is the principal operational constraint on commercial nuclear-marine operation. The port state has full sovereign discretion to refuse port entry to any vessel under its national law and the IMO conventions. For nuclear-powered ships, port-state acceptance has historically been limited and conditional, with the Otto Hahn experience (refused entry to several West African and South American ports) the cautionary reference. As of 2026, the set of ports globally that accept nuclear-powered merchant calls is estimated at approximately 15, including the Russian Murmansk, Sabetta, Pevek and Vladivostok ports, the US Newport News (for naval visits, with limited civilian capability), the UK Faslane and Devonport (naval), and a handful of others. The expansion of port-state acceptance to support the 2024-2025 SMR-marine programmes is one of the regulatory hurdles being addressed at the bilateral level between programme owners and the destination port states.
UNCLOS Article 23 (“Foreign nuclear-powered ships and ships carrying nuclear or other inherently dangerous or noxious substances”) provides that “foreign nuclear-powered ships and ships carrying nuclear or other inherently dangerous or noxious substances shall, when exercising the right of innocent passage through the territorial sea, carry documents and observe special precautionary measures established for such ships by international agreements.” The Article confirms the right of innocent passage for nuclear-powered ships through the territorial sea (12 nautical miles) of any coastal state, conditional on documentation and special precautionary measures. The Article does not address port-state acceptance, which is governed by separate provisions (UNCLOS Article 25 and Article 211) that preserve the coastal state’s discretion over port entry.
UNCLOS Article 22 (“Sea lanes and traffic separation schemes in the territorial sea”) allows the coastal state to require nuclear-powered ships to confine passage to designated sea lanes, with the practical effect that the coastal state can route nuclear-powered transit traffic through a defined corridor that does not approach densely populated coastlines.
The combined effect of UNCLOS Articles 22 and 23 is that a nuclear-powered merchant ship can transit any maritime zone in international navigation but can be required to use specific sea lanes and to carry specific documentation, while the same ship has no automatic right of port entry in any port other than its flag-state ports. The practical implication for commercial operation is that the nuclear-powered ship must have a defined route and a defined set of accepting ports for the operational profile to be commercially viable, which currently constrains the SMR-marine programmes to liner trades on routes between accepting flag-state and port-state pairs.
Cost economics: CAPEX-heavy, OPEX-light, lifetime fuel-free
The cost economics of nuclear marine propulsion are structurally different from conventional and alternative-fuel propulsion. The capital expenditure (CAPEX) per installed kilowatt is several times higher than a conventional propulsion plant, and the operational expenditure (OPEX) is dominated by reactor operation, regulatory compliance, and end-of-life decommissioning rather than fuel cost.
CAPEX: A 30 MWe SMR-powered Aframax tanker is estimated at approximately USD 200 to 500 million capital cost as of 2026 design horizon, against USD 80 to 100 million for a conventional VLSFO/methanol Aframax newbuild and USD 150 to 200 million for an LNG-fuelled or ammonia-fuelled equivalent. The nuclear premium is concentrated in three areas: the reactor and primary cooling system (approximately USD 100 to 250 million), the radiological safety and containment systems (approximately USD 50 to 100 million), and the integrated electric propulsion plant with redundancy (approximately USD 30 to 60 million). The incremental CAPEX per shaft horsepower for the SMR-marine option is approximately 3 to 5 times the conventional baseline.
OPEX: The fuel cost for a nuclear-powered vessel is concentrated in the periodic core-replacement event (every 5 to 12 years), with an estimated cost of USD 50 to 100 million per core for a 30 MWe SMR. Amortised over the inter-refuelling interval, the fuel cost is approximately USD 5 to 10 million per year, against USD 20 to 40 million per year for a conventional VLSFO Aframax at 2025 fuel prices. The OPEX advantage is substantial but conditional on the fuel-cycle cost remaining stable across the operational life, and on the reactor compartment achieving the design fuel-cycle interval without unexpected refuelling events.
Crewing and operations: A nuclear-powered merchant ship requires a dual nuclear-and-marine crew, with approximately 4 to 6 reactor operators and supervisors in addition to the conventional deck and engine-room manning. The crewing cost is approximately USD 1 to 2 million per year above the conventional baseline. The 2024-2025 SMR-marine programmes are exploring reduced manning through autonomous reactor monitoring and remote operation, but the initial commercial operations are expected to use full manning for regulatory acceptance.
Insurance: The marine-insurance cost for a nuclear-powered merchant ship is estimated at approximately 2 to 5 times the conventional baseline, reflecting the additional perils of radiological release, port-state liability, and decommissioning. The IAEA Convention on Civil Liability for Nuclear Damage (Vienna Convention) and the OECD Paris Convention provide the international liability framework, with the tonnage-based liability cap at approximately EUR 1.2 billion per incident under the 2004 amendment to the Paris Convention.
Decommissioning: The end-of-life decommissioning cost for a nuclear-powered merchant ship is estimated at approximately USD 100 to 200 million per vessel, principally for the reactor pressure vessel removal, the spent-fuel handling, and the radiological clean-up of the engine-room compartment. The decommissioning cost is typically funded through a sinking-fund mechanism over the operational life, with approximately USD 5 to 10 million per year set aside.
The lifecycle cost-of-ownership comparison between nuclear-marine and conventional propulsion is sensitive to three swing variables: the carbon-pricing trajectory (a high carbon price favours nuclear because of the zero TtW CO2), the alternative-fuel cost trajectory (a high green-hydrogen or ammonia price favours nuclear), and the regulatory-compliance cost trajectory (high CCS, RFNBO certification, or port-state-acceptance cost favours nuclear). The 2024-2025 SMR-marine programmes have been developed on the assumption of all three swing variables moving in favour of nuclear, with the commercial case sized at approximately 15 to 25 percent IRR over a 30-year operational life under those assumptions.
Nuclear waste disposal infrastructure
The nuclear waste disposal infrastructure for civilian merchant nuclear-marine programmes is the long-tail liability that conditions the lifecycle accounting. The principal waste streams are spent fuel (high-level waste), intermediate-level waste from reactor operation and decommissioning, and low-level waste from operational maintenance.
Spent fuel is the principal high-level waste stream. A 30 MWe SMR with a 7-year fuel cycle produces approximately 5 to 10 tonnes of spent fuel per refuelling event. The spent fuel is initially cooled in a spent-fuel pool at a designated land-based facility for 5 to 10 years, then transferred to dry-cask storage for an interim period of 30 to 60 years, and finally placed in a geological repository for permanent disposal. The geological repository infrastructure is operational only in Finland (Onkalo, accepting first emplacement in 2025) and at advanced design stage in Sweden, France and Switzerland; most nuclear-fuel-cycle states currently maintain interim dry-cask storage with deferred final-disposal decision.
The marine-spent-fuel handling pathway requires the producing state to operate a designated land-based facility that can receive the fuel from the marine reactor, conduct the cooling and conditioning, and transfer the fuel to the national waste-management infrastructure. The current set of states with the operational facility for civilian marine spent-fuel is small: Russia (Mayak and Andreyeva facilities for the Atomflot fleet), the United States (the Idaho National Laboratory for naval-reactor spent fuel, with civilian capability under development), the United Kingdom (Sellafield), and France (La Hague).
Intermediate-level waste is the activated structural material from reactor operation and the decommissioned reactor pressure vessel and primary cooling system. The volume per reactor-life is approximately 50 to 200 cubic metres per reactor, with the disposal pathway through near-surface engineered repositories (the French Aube facility, the UK Drigg facility, the Finnish Loviisa facility).
Low-level waste is the operational maintenance waste (gloves, filters, contaminated tooling) and is managed through the producing state’s standard low-level waste infrastructure.
The nuclear-waste disposal infrastructure is a state-operated rather than commercially-operated function, which differs from the chemical-fuel waste-management framework where the user (the shipowner or the bunker supplier) is generally responsible for the waste disposal. For nuclear-marine commercial operation, the producing state of the fuel is the principal waste-cycle counterparty, and the commercial agreement between the shipowner and the fuel supplier must include the spent-fuel return and waste-disposal provisions. The IAEA INFCIRC/254 nuclear-supplier guidelines provide the international framework for this commercial relationship, with bilateral state-to-state agreements at the implementation level.
The waste-disposal cost contribution to the lifecycle WtW intensity is approximately 1 to 3 gCO2eq/MJ of reactor energy on a typical centrifuge-LEU pathway, dominated by the energy demand of the spent-fuel cooling, the dry-cask construction, the geological-repository excavation, and the long-term monitoring infrastructure. The cost contribution is approximately 5 to 10 percent of the total lifecycle CAPEX of the marine programme, structured as a sinking-fund contribution over the operational life.
Comparison with hydrogen (cross-link)
Nuclear marine propulsion and hydrogen marine propulsion share the structural feature of zero tank-to-wake CO2 because neither energy source contains carbon atoms. Both pathways are therefore “zero TtW carbon” options under MEPC.391(82) and FuelEU Annex II, conditional on the relevant lifecycle accounting being applied correctly.
The two pathways differ structurally on several dimensions.
Lifecycle WtW intensity: Centrifuge-enriched LEU nuclear at approximately 5 to 15 gCO2eq/MJ is comparable to direct-coupled green hydrogen at approximately 1 to 5 gCO2eq/MJ at the plant gate. After liquefaction, green liquid hydrogen at a Norwegian-grid liquefier is approximately 12 to 18 gCO2eq/MJ, comparable to the centrifuge-LEU range. Grey hydrogen at approximately 104 gCO2eq/MJ is materially worse than nuclear, and brown hydrogen from coal gasification is worse still. The two pathways are roughly comparable on lifecycle WtW intensity at their best-case versions.
Volumetric energy density: Nuclear fission fuel at approximately 80 million MJ/kg is six orders of magnitude higher than hydrogen at 120 MJ/kg, and the volumetric burden of nuclear fuel storage is essentially zero against the substantial bunker-tank size penalty for liquid or compressed hydrogen.
Refuelling logistics: Nuclear refuelling every 5 to 12 years versus hydrogen bunkering every 7 to 10 days. Nuclear simplifies the logistics chain at the operational level but constrains the refuelling site to a small set of approved facilities; hydrogen requires a substantial bunker-port infrastructure but allows operation at any port with bunker supply.
Capital cost: Nuclear at approximately USD 200 to 500 million per Aframax versus hydrogen at approximately USD 120 to 180 million for a hydrogen-fuel-cell or hydrogen-ICE Aframax (subject to fuel-cell stack cost). Nuclear is materially CAPEX-heavier.
Operating cost: Nuclear OPEX at approximately USD 5 to 10 million per year amortised fuel cost versus hydrogen at approximately USD 15 to 25 million per year on green hydrogen at projected 2030 prices (USD 3 to 5 per kg green LH2). Nuclear is OPEX-lighter.
Regulatory framework: Nuclear under A.491(XII) and the under-revision A.491; hydrogen under the under-development IGF Code amendments and MSC.1/Circ.1671 interim guidelines. The hydrogen framework is more advanced as of 2026, with first-of-a-kind newbuilds in service (MS Topeka, MF Hydra) and the IGF Code amendments expected to enter force in 2028. The nuclear framework requires the A.491 revision to complete (target MSC 112 in mid-2028) and the MEPC.391(82) and FuelEU Annex II inclusion to be added (target 2027 to 2030).
Port-state acceptance: Nuclear-powered ships are accepted at approximately 15 ports globally as of 2026, principally Russian; hydrogen-fuelled ships have rapidly expanding port acceptance through the under-development IGF Code framework, with several Northern European, Norwegian, Singaporean and Korean ports developing hydrogen bunker capability.
The two pathways are likely to be complementary rather than competing for the 2030-2050 decarbonisation transition. Nuclear is positioned for the deep-sea liner trades on stable routes between accepting flag-state and port-state pairs, while hydrogen (and ammonia, e-methanol and e-LNG) are positioned for the broader merchant fleet with flexible bunker-port operation.
Formula, assumptions, and limits
Formula
The well-to-wake intensity of nuclear marine propulsion in a lifecycle accounting framework consistent with MEPC.391(82) and FuelEU Annex II takes the form:
where EF_mining+enrichment+fab is the upstream lifecycle intensity from uranium mining, conversion, enrichment and fuel fabrication; EF_operation,zero is the operational-stage intensity at the reactor boundary (approximately zero gCO2eq/MJ for fission); and EF_spent-fuel+decommission is the downstream lifecycle intensity from spent-fuel handling, geological disposal, and reactor decommissioning, amortised over the fuel-cycle and operational life.
For centrifuge-enriched LEU fuel at modern PWR or SMR reactor design:
For diffusion-enriched LEU fuel (legacy pathway, no longer in commercial production as of 2026):
The enrichment-stage contribution is the dominant variable across the two pathways:
where SWU per kg fuel depends on the enrichment level (approximately 5 SWU/kg at 4.4 percent, 30 SWU/kg at 20 percent), kWh per SWU depends on the technology (approximately 50 to 65 kWh/SWU centrifuge, 2,400 kWh/SWU diffusion), and EF_grid,enrichment site is the carbon intensity of the electricity supply at the enrichment plant.
Derivation
The WtW formula for nuclear marine propulsion is derived from the standard fuel-lifecycle methodology with the operational-stage CO2 set to zero. The fission reaction U-235 + n yields fission products + 2.4 n + approximately 200 MeV of energy, with no carbon atom in the fuel and no carbon atom in the fission products. The reactor coolant (water in PWR, salt in MSR, lead in LFR) does not participate in the energy-release reaction, so there is no operational-stage CO2 release.
The upstream stages (mining, conversion, enrichment, fabrication) are characterised by their electricity and process-energy demand and the corresponding emissions on the grid intensity at the producing site. The principal electricity-demand stage is enrichment, which dominates the upstream lifecycle when conducted by gaseous diffusion and is a smaller contributor when conducted by gas centrifuge. The mining and milling stages contribute through diesel fuel for mining equipment and grid electricity for milling, with the figures sensitive to the ore grade and the mine type (open-pit, underground, ISR). The conversion and fabrication stages are smaller contributors.
The downstream stages (spent-fuel handling, dry-cask storage, geological disposal, decommissioning) are characterised by their long-duration energy and material demand, amortised over the fuel-cycle and the operational life. The dry-cask storage requires construction energy and material (concrete, steel) with associated embodied emissions; the geological repository requires excavation, backfill and long-term monitoring infrastructure; the decommissioning requires reactor pressure vessel removal, contaminated material handling and site clean-up. The downstream contribution is approximately 1 to 4 gCO2eq/MJ in typical accounting, with the larger figure applying to states with longer interim-storage periods before final disposal.
The IPCC AR6 lifecycle figure of approximately 12 gCO2eq/kWh (median, 5 to 110 gCO2eq/kWh range) for nuclear electricity is the basis for the marine-application figure. The conversion from electricity to shaft energy is at the marine reactor and turbine boundary (approximately 30 to 35 percent thermal efficiency), with the WtW intensity expressed in gCO2eq/MJ of shaft energy or, equivalently, gCO2eq/MJ of bunker-equivalent energy at the LCV reference.
Assumptions
The framework assumes:
- The fuel core is sourced from a recognised civilian nuclear-fuel-cycle supplier with documented chain of custody from mining to fabrication, and the producing-state safety regime is consistent with IAEA safeguards under the Comprehensive Safeguards Agreement and the Additional Protocol where applicable.
- The enrichment is by gas centrifuge technology, with a documented kWh per SWU and grid-intensity attribution at the enrichment site.
- The reactor operation is at the design fuel-cycle interval without unplanned refuelling events, and the spent-fuel inventory is returned to the producing-state for handling and disposal.
- The geological repository is operational at the time of final emplacement, or a deferred-disposal credit is taken with the long-term monitoring cost amortised through the sinking fund.
- The reactor compartment achieves the design operational life (typically 30 to 60 years) without major maintenance events that affect the lifecycle accounting.
- The decommissioning cost and emissions are funded through a sinking-fund mechanism over the operational life and are included in the lifecycle accounting.
- The flag-state regulatory framework supports civilian nuclear-merchant operation under the IMO instrument (currently A.491(XII), expected revised A.491) and the relevant national civilian-nuclear regulation.
- The port-state acceptance is documented for the operational route, with bilateral agreements covering the radiological emergency-response framework.
Worked example
A nuclear-powered Aframax tanker uses a 30 MWe integrated PWR (75 MWt thermal at 40 percent thermal efficiency) with 20 percent enriched UO2 fuel from a centrifuge-enrichment supplier. The fuel core mass is approximately 2.5 tonnes of U-235 equivalent, sized for a 7-year fuel cycle at full power. The lifecycle WtW intensity components are:
- Uranium mining and milling: 3 gCO2eq/MJ
- Uranium conversion: 0.7 gCO2eq/MJ
- Uranium enrichment (centrifuge at 60 kWh/SWU on 50 gCO2eq/kWh grid): 1.8 gCO2eq/MJ
- Fuel fabrication: 0.5 gCO2eq/MJ
- Reactor operation: 0 gCO2eq/MJ
- Spent-fuel handling and geological disposal: 1.5 gCO2eq/MJ
- Reactor decommissioning: 1.0 gCO2eq/MJ
Aggregate WtW intensity: approximately 8.5 gCO2eq/MJ on a shaft-energy basis.
The vessel’s compliance value under the proposed FuelEU treatment with a target of approximately 88.0 gCO2eq/MJ in 2026 would be 88.0 minus 8.5 = 79.5 gCO2eq/MJ of surplus per MJ of shaft energy. The compliance arithmetic is detailed at /calculators/fuel-wtw-nuclear and benchmarked against hydrogen, ammonia, and the GFI attained calculation.
A diffusion-enriched legacy counterfactual at approximately 35 gCO2eq/MJ (with the enrichment-stage contribution rising to approximately 25 gCO2eq/MJ on a 700 gCO2eq/kWh fossil-grid Paducah-equivalent enrichment site) would still deliver compliance surplus of approximately 53 gCO2eq/MJ versus the 2026 target. The legacy pathway is no longer in commercial production, so the example is academic, but it illustrates the sensitivity to the enrichment-stage technology choice.
Edge cases and limits
- HEU naval reactors: military naval reactors using HEU at 20 to 93 percent enrichment are excluded from the civilian merchant framework under MARPOL Annex VI and FuelEU. The military reactors operate under each navy’s national regulatory regime and are not subject to the IMO commercial framework.
- Russian KLT-40 90 percent enrichment exception: the Sevmorput LASH carrier and the Lenin icebreaker (decommissioned) used 90 percent HEU fuel as a Russian civilian exception, with the modern Russian fleet (KLT-40S, RITM-200) at LEU enrichment levels (18.6 to 20 percent). The 90 percent fuel was produced by Russian diffusion enrichment in the Soviet era, and the lifecycle WtW for that fuel is materially higher than the modern centrifuge-LEU baseline.
- Reactor decommissioning timing: the decommissioning cost is amortised over the operational life (typically 30 to 60 years), but the actual decommissioning may occur 30 to 100 years after end of operation if the reactor compartment is placed in interim storage. The discount rate and the time-value of the decommissioning cost are sensitive to this assumption.
- Geological repository availability: the assumption that a geological repository is operational at the time of final spent-fuel emplacement is sensitive to the Finnish Onkalo, Swedish SKB, French Cigeo and other repository programmes maintaining their operational schedules. A deferred-disposal scenario adds approximately 1 to 3 gCO2eq/MJ to the lifecycle intensity for the extended interim storage period.
- Spent-fuel reprocessing for MOX: a closed fuel cycle with spent-fuel reprocessing (the French La Hague pathway, the proposed UK Sellafield extension, the Russian Mayak pathway) reduces the spent-fuel volume by approximately 75 percent and produces MOX fuel for re-use in suitable reactors. The lifecycle WtW intensity for a closed fuel cycle is approximately 60 to 80 percent of the once-through cycle, principally through the reduced uranium-mining demand. The closed cycle requires a reprocessing infrastructure that is operational only in France, Russia, the UK (limited) and Japan (limited).
- Methane leakage in fossil-grid enrichment electricity: the enrichment-stage contribution is sensitive to the methane leakage rate in the natural-gas supply for any enrichment plants supplied by gas-fired electricity. The figure is similar to the methane-leakage sensitivity for blue hydrogen.
- Reactor compartment removal at end of life: the marine reactor compartment is typically removed as a single unit at end of life and shipped to the producing-state’s land-based facility for further dismantling. The transport energy and emissions for the removal voyage are included in the decommissioning lifecycle, typically at 0.2 to 0.5 gCO2eq/MJ.
- First-of-a-kind premium: the first commercial SMR-marine newbuilds are expected to carry a first-of-a-kind cost premium of approximately 20 to 40 percent above the nth-of-a-kind cost. The lifecycle WtW intensity is not affected by the first-of-a-kind premium, but the financial economics are.
- Combined heat and power: a marine SMR can be operated as a combined-heat-and-power (CHP) plant for a port-served heat application during port stays, raising the effective utilisation of the reactor and reducing the amortised lifecycle intensity per unit of useful energy. The CHP option is being explored in the Russian Pevek floating power plant model and is a candidate for future SMR-marine designs at port-rich operational profiles.
- Hydrogen-leakage analogy: the indirect-GHG impact of nuclear-fuel-cycle releases (for example, krypton-85 from spent-fuel reprocessing) is not currently incorporated in the FuelEU or MEPC.391(82) lifecycle methodology, but is a topic for future revision similar to the under-discussion hydrogen-leakage GWP integration.
Regulatory basis
- IMO Resolution A.491(XII), 1981, Code of Safety for Nuclear Merchant Ships (the standing IMO instrument under revision).
- IMO MEPC.391(82), 2023 Guidelines on Lifecycle GHG Intensity of Marine Fuels (LCA Guidelines), Annex 1 (currently no nuclear pathway, under amendment).
- Regulation (EU) 2023/1805 (FuelEU Maritime), Annex II default WtW emission factors (currently no nuclear pathway, under proposed revision for 2030 amendment).
- IMO IGF Code Part A (currently silent on nuclear; nuclear-marine framework provided by parallel A.491(XII) instrument).
- United Nations Convention on the Law of the Sea (UNCLOS) Article 22 and Article 23 (transit and innocent-passage rules for nuclear-powered ships).
- IAEA Convention on Civil Liability for Nuclear Damage (Vienna Convention) and OECD Paris Convention (commercial liability framework for nuclear-powered ship operations).
- IAEA Convention on Early Notification of a Nuclear Accident and the Convention on Assistance in the Case of a Nuclear Accident (radiological emergency-response framework).
- IAEA INFCIRC/254 nuclear-supplier guidelines (international framework for civilian-nuclear fuel-cycle commercial relationships).
- IPCC AR6 Working Group III, lifecycle GHG intensity of nuclear electricity generation including the median 12 gCO2eq/kWh figure and the 5 to 110 gCO2eq/kWh range.
- IAEA Comprehensive Safeguards Agreement (CSA) and Additional Protocol (AP) for civilian-nuclear fuel-cycle verification.
- National civilian-nuclear regulatory frameworks of the producing states (Russia, USA, UK, France, China, plus emerging frameworks in Korea, Japan, Italy and others).
- Class society rule sets for nuclear-marine and SMR-marine application (Lloyd’s Register, Bureau Veritas, ABS, DNV, ClassNK, RINA), drawing on the under-revision A.491 instrument and the relevant national civilian-nuclear framework.
Common errors
- Treating nuclear marine propulsion as a zero-emission fuel because the operational stage produces zero TtW CO2. The lifecycle WtW intensity of approximately 5 to 15 gCO2eq/MJ for centrifuge-enriched LEU is small but non-zero, dominated by the upstream uranium mining, enrichment and fabrication stages.
- Using the diffusion-enrichment lifecycle figure for a modern centrifuge-enrichment fuel core. The two pathways differ by a factor of approximately 2 to 5 in WtW intensity, with diffusion no longer in commercial production as of 2026.
- Omitting the downstream waste-cycle contribution from the WtW accounting. The spent-fuel handling, geological disposal and reactor decommissioning add approximately 1 to 4 gCO2eq/MJ to the lifecycle intensity and must be amortised over the operational life through a sinking-fund mechanism.
- Conflating military naval reactor performance with civilian merchant reactor performance. The military reactors use HEU and operate under different regulatory and lifecycle accounting frameworks; the civilian merchant reactors use LEU and operate under MARPOL Annex VI and the relevant IMO instrument.
- Assuming the IMO MEPC.391(82) and FuelEU Annex II include a nuclear-marine pathway. The current text of both instruments excludes nuclear, with the inclusion under discussion for the next amendment cycle (2027 to 2030 expected).
- Ignoring the port-state acceptance constraint when assessing operational routes for a nuclear-powered merchant ship. The set of accepting ports is small (approximately 15 globally as of 2026) and conditions the commercial route choice.
- Treating the AiP (approval-in-principle) statement as equivalent to final regulatory approval. The AiP is a class-society technical statement and does not authorise commercial operation; final approval is provided by the flag-state administration based on the IMO instrument and the national civilian-nuclear framework.
- Assuming the refuelling interval translates directly to operational availability. The reactor compartment requires periodic maintenance shutdowns (typically every 18 to 24 months) for instrumentation calibration, primary-coolant chemistry checks, and steam-generator inspection, in addition to the 5 to 12 year refuelling interval.
- Applying the CAPEX figure for a first-of-a-kind newbuild to a series production. The first SMR-marine newbuilds carry a first-of-a-kind premium of 20 to 40 percent above the nth-of-a-kind cost, and the comparison with conventional propulsion economics should be conducted on the appropriate cost basis.
- Excluding the marine insurance premium for nuclear-powered ships from the cost-of-ownership accounting. The insurance cost is approximately 2 to 5 times the conventional baseline, reflecting the radiological liability under the Vienna and Paris conventions.
- Using LCV-based comparison between nuclear fission energy (no chemical LCV) and chemical fuels at chemical LCV. The energy-equivalence is at the shaft-energy boundary, not at the bunker-manifold LCV, and the comparison should normalise to the propulsion energy delivered.
See also
- /wiki/per-fuel-wtw-hydrogen
- /wiki/per-fuel-wtw-ammonia-grades
- /wiki/per-fuel-wtw-methanol-grades
- /wiki/per-fuel-wtw-bio-lng
- /wiki/per-fuel-wtw-vlsfo-mgo
- /wiki/marine-gfs-methodology
- /wiki/marpol-annex-vi
- /wiki/imo-net-zero-framework
- /wiki/fueleu-intensity-formula-breakdown
- /wiki/antarctic-special-area-and-polar-code
- /calculators/fuel-wtw-nuclear
- /calculators/fuel-wtw-blend
- /calculators/fuel-wtw-hydrogen
- /calculators/fuel-wtw-ammonia
- /calculators/fueleu-ghg-intensity
- /calculators/gfi-attained
Related calculators
- e-Diesel / FT e-Fuel - Well-to-Wake
- VLSFO - Well-to-Wake
- LPG - Well-to-Wake
- LNG - Well-to-Wake by engine pathway
- HFO - Well-to-Wake
- MGO / MDO - Well-to-Wake
- Methanol - Well-to-Wake by pathway
- HVO / Renewable Diesel - Well-to-Wake