SerpApi.Org
South Korea's Samsung Heavy Industries and Chicago-based nuclear engineering firm Sargent & Lundy signed a memorandum of understanding on July 21 in Washington, D.C., to jointly develop a universal floating small modular reactor platform — one designed from the outset to accept virtually any reactor type a customer chooses, not just a single vendor's design. The partnership moves Samsung Heavy's floating SMR program from the concept-design phase it completed last year into a full commercialization push, and it comes at a moment when offshore nuclear power has shifted from speculative to earnest: as of May 2026, thirteen announced projects have committed more than 9.8 gigawatts of nuclear capacity specifically to AI data center infrastructure.
The detail that makes this deal structurally significant is the one that most partnership announcements bury: neither company is betting on any single small modular reactor design to win the commercial race. Instead, they're building the platform that any winner could use.
What "Reactor-Agnostic" Actually Means — and Why It Matters Now
The engineering insight behind Samsung's floating SMR design is deceptively simple: split the platform into two separate compartments — one for the nuclear island (the reactor vessel, primary coolant loop, steam generators, and containment structures) and one for the power generation island (turbines, generators, electrical systems, and grid-connection equipment). A standardized interface connects the two compartments. Swap out the nuclear island for a different reactor model, and the power island doesn't have to move.
The American Bureau of Shipping confirmed this approach works in principle. The classification society granted Samsung Heavy an Approval in Principle for the concept in December 2025 — technically, the approval was granted on October 24, 2025, and publicly announced in mid-December. The initial concept was designed around two KAERI SMART100 reactors (each producing 110 megawatts of electricity), but ABS noted explicitly that the compartmental layout was "ultimately allowing the use of a variety of different SMR types."
The reason this architecture matters now, specifically, has to do with the economics of the SMR market. As of 2026, 127 distinct SMR designs exist globally, but only China and Russia have operational commercial units. No SMR design has yet produced the 40 to 70 serial units that most analysts believe are required before factory-scale manufacturing economics kick in. NuScale's flagship US project — the Carbon Free Power Project — was cancelled in November 2023 after its unsubsidized estimated generation cost rose to $119 per megawatt-hour, more than double an earlier projection. A 2023 Monte Carlo study published in the journal Energy found no SMR design among 19 surveyed was economically competitive without subsidy.
In that environment, a platform locked to a single reactor design is a commercial gamble on a specific vendor's regulatory and market journey. A reactor-agnostic platform is a hedge against exactly that uncertainty.
Samsung and Sargent & Lundy: What Each Brings
The partnership joins capabilities that neither company could easily replicate alone.
Samsung Heavy Industries, founded in August 1974 and headquartered in Seongnam, South Korea, is one of the world's three largest shipbuilders alongside Hanwha Ocean and HD Hyundai. The company's offshore engineering portfolio includes drillships, floating liquefied natural gas vessels, and floating production storage and offloading platforms — some of the most complex integrated industrial systems ever built and deployed at sea. Applying that offshore EPC (engineering, procurement, and construction) competence to a floating nuclear platform is a logical technology transfer, though a demanding one: nuclear-grade containment requirements and radiation environments create constraints that conventional offshore engineering doesn't address.
Sargent & Lundy, founded in 1891 and currently celebrating its 135th anniversary, has been involved continuously in nuclear power plant work since 1954. The firm is one of a handful of engineering companies in the world with hands-on small modular reactor design experience: it designed the nuclear island for NuScale Power's VOYGR SMR — the only SMR design fully licensed by the US Nuclear Regulatory Commission for commercial deployment in the United States. Industry publication ENR named Sargent & Lundy a leading firm in the nuclear plant sector in its 2026 Top Design Firms Sourcebook.
Under the MOU, Samsung contributes offshore EPC capability; Sargent & Lundy provides nuclear plant design and licensing know-how. The agreement covers joint development of a standard FSMR platform for multiple reactor types, with cooperation across the full commercialization pathway: development, licensing support, construction planning, and maritime deployment strategies, with an explicit focus on the US market.
"This collaboration represents a significant step forward for America's future clean energy," said Shiven Sulka, Sargent & Lundy's Senior Vice President and Chief Nuclear Officer. "By combining S&L's nuclear engineering leadership with Samsung Heavy Industries' proven offshore platform expertise, we will accelerate the development of an FSMR tailored for the U.S. and global energy market."
Kim Kyung-hee, Samsung Heavy Industries' Executive Vice President and Head of the Future Business Division, added that the Sargent & Lundy partnership would "take Samsung Heavy Industries' FSMR technology to the next level."
The Regulatory Mountain: Two Frameworks, Neither Designed for the Other
The technical elegance of the compartmental design masks a formidable regulatory challenge that the partnership must navigate — and that the draft of this article initially understated.
A floating nuclear power plant moored off a US coastline would need to satisfy two entirely separate regulatory systems designed in isolation from each other. The first is maritime classification law, administered in this case by ABS. An Approval in Principle confirms that the concept design complies with the intent of applicable ABS rules and industry codes — but it is explicitly an early-stage feasibility assessment, not a construction certificate.
The second and far more demanding system is national nuclear regulatory licensing. In the United States, that means the Nuclear Regulatory Commission. NuScale's VOYGR — the only US-licensed SMR as of 2025 — required more than a decade of NRC review. The SMART100 reactor holds standard design approval from South Korea's Nuclear Safety and Security Commission, granted in September 2024, but that approval does not transfer to US jurisdiction. A US deployment of any reactor type on the Samsung platform would require a separate, full NRC review.
There is no international framework that harmonizes maritime classification and nuclear regulatory licensing. The two systems were developed independently, by different institutions with different safety philosophies, and they overlap imperfectly. Bridging them — and doing so fast enough to be commercially relevant in the AI data center build-out window — is the central challenge the SHI–S&L partnership has accepted.
This is precisely where Sargent & Lundy's NRC relationship adds depth that Samsung cannot provide alone.
What AI Demand Has to Do With Floating Reactors
The market logic for this technology starts with a number: the International Energy Agency has projected that global data center electricity consumption could exceed 1,000 terawatt-hours annually by 2026 — roughly equivalent to Japan's entire national power consumption. Goldman Sachs Research projected in May 2026 that US data center power demand alone would climb from 31 gigawatts in 2025 to 66 gigawatts by 2027, with capacity additions accelerating dramatically. Globally, Goldman Sachs projects a 165% increase in data center power demand by 2030 compared to 2023 levels.
That demand is creating a land-permitting bottleneck that is among the most acute constraints facing the hyperscalers building AI infrastructure. Grid interconnection queues in the United States have grown to over 2,600 gigawatts of pending projects, with average wait times of five years. Securing a site for a new land-based power plant — nuclear or otherwise — requires navigating utility interconnection processes, local planning and zoning, state permitting, and community engagement that can stretch a decade.
A floating nuclear platform offshore circumvents much of that friction. It doesn't compete for scarce onshore sites. It can be positioned near coastal data center facilities or island populations that cannot easily connect to mainland grids. Samsung has positioned the FSMR explicitly for "coastal and island regions where securing onshore sites is difficult or grid connections are limited."
Whether that positioning is compelling enough for a paying customer depends on the regulatory timelines described above. But the Trump administration's May 2025 executive orders — which set a target of expanding US nuclear capacity from approximately 100 gigawatts today to 400 gigawatts by 2050 and directed an 18-month cap on NRC licensing timelines — signal federal intent to compress exactly those timelines.
Only One Operating Floating Nuclear Plant Exists — and Russia Built It
Context is useful here. The world's only commercial operational floating nuclear power plant is Russia's Akademik Lomonosov — a non-self-propelled power barge moored in the Arctic port of Pevek in the Chukotka region, operated by Rosatom. The vessel carries two KLT-40S pressurized-water reactors derived from nuclear icebreaker propulsion technology, producing up to 70 megawatts of electrical output. Connected to the regional grid in December 2019 and commercially operational by May 2020, the Akademik Lomonosov had produced its first billion kilowatt-hours by early 2025 and completed its first refueling cycle. It has increased its supply of power to the remote Chukotka region from about 20% to 60% of local demand.
It is, by any measure, a proof of concept. But it is also purpose-built for a state utility in a jurisdiction with no commercial equivalent, and it was designed around icebreaker reactor technology rather than a commercially exportable SMR. It is not a replicable template for what Samsung and Sargent & Lundy are building.
What the Samsung–S&L partnership is attempting is categorically different: a privately financed, internationally marketed, standardized platform governed by Western classification standards and designed for multiple reactor types, built by a commercial shipbuilder for commercial customers in commercial markets. Russia's approach proved the concept was physically viable. The new partnership is trying to prove it is commercially viable.
South Korea's Broader Bet
Samsung Heavy is not alone among South Korean shipbuilders in pursuing future offshore energy and platform markets. HD Hyundai is developing a nuclear supply chain partnership with TerraPower and accelerating development of nuclear-powered container ships and floating data centers. Hanwha Ocean is advancing unmanned surface vessels and zero-carbon shipping concepts. The South Korean government committed in May 2026 to invest 525 billion won (approximately $355 million, at the July 23, 2026 exchange rate of approximately 1,480 KRW per USD; exchange rate conversions are approximate) over five years to secure core technologies for seven categories of future vessel platforms.
A shipbuilding industry official quoted in Korean media put it plainly: if LNG carriers were the representative high-value vessel that drove the growth of South Korea's shipbuilding industry over the past three decades, the next competition will be in the future marine platform market combining energy, data, and defense.
Samsung Heavy itself is now competing simultaneously in two adjacent floating platform markets. In June 2026, the company signed agreements with Greek shipowner Capital Clean Energy Carriers and classification society Lloyd's Register at the Posidonia shipping exhibition in Athens — covering a non-nuclear, 50-megawatt floating data center platform using seawater cooling and onboard conventional power generation.
The two platforms — one nuclear, one conventional — are distinct products for distinct markets. But they represent the same strategic bet: that demand for offshore power and computing infrastructure is real enough to justify building the platforms before all the customers have materialized.
How Does a Floating Nuclear Power Plant Actually Generate Electricity?
The SMART100 reactor at the core of Samsung's initial concept is an integrated pressurized-water reactor. "Integrated" means all primary coolant system components — the reactor vessel, steam generators, pressurizer, and coolant pumps — are housed inside a single sealed reactor vessel, eliminating the large-bore primary coolant piping that represents a significant accident risk in conventional pressurized-water reactor designs.
The reactor heats water under pressure (preventing it from boiling) in a primary loop. That heated water transfers its heat through steam generators to a secondary water loop, which does boil, producing steam. The steam drives turbines connected to electrical generators. The turbines and generators are in the separate "power island" compartment of the floating platform.
The SMART100 is designed with fully passive safety systems — the safety mechanisms operate through natural forces (gravity, natural circulation, compressed gas pressure) rather than requiring active pumps, valves operated by electric power, or manual human intervention. This passive safety philosophy is essential for a floating deployment where the operating environment may be remote, crew sizes small, and emergency response limited.
After generating electricity, the steam is condensed back to water and recirculated. The platform uses seawater as its ultimate heat sink — a significant advantage over land-based plants, which must be sited near rivers or reservoirs for cooling water. The floating deployment provides essentially unlimited cooling water access.
Can a Different Reactor Just "Slot In"?
The compartmental architecture is elegant in concept; the engineering challenge is in the interface specifications. Every SMR design has a different thermal output, a different primary coolant chemistry, different structural dimensions, and different safety system requirements. A reactor-agnostic platform doesn't mean any reactor plugs in without modification — it means the platform's non-reactor components (the hull structure, power island, marine systems, and safety support systems) are designed to accommodate a range of reactor parameters rather than being precisely optimized for one.
The ABS Approval in Principle covers the platform's hull and non-reactor systems. The nuclear island itself — whatever reactor type a customer eventually selects — must go through its own separate nuclear regulatory approval before it can be installed. The modular interface between the two compartments defines what range of reactors can be accommodated, and that specification is exactly what the Samsung–Sargent & Lundy engineering collaboration will be working out during the development phase ahead.
Frequently Asked Questions
What is a floating small modular reactor, and how does it differ from a regular nuclear plant?
A floating small modular reactor (FSMR) is a nuclear power plant mounted on an offshore floating structure — a barge or ship hull — rather than anchored to land. Like a land-based nuclear plant, it uses fission heat to generate steam, which drives turbines and produces electricity. The differences are primarily in deployment: an FSMR can be positioned in coastal waters, moored offshore near the facilities it serves, and — in principle — repositioned between sites. It avoids the land permitting, grid interconnection queues, and community opposition that complicate onshore nuclear plant siting. The SMART100 design at the core of Samsung Heavy's initial concept also uses passive safety systems rather than active pumps and valves, which reduces staffing and operational complexity.
How is this different from the reactor-agnostic design, and why does that matter for buyers?
Traditional nuclear plants are designed around a specific reactor type, with every structural, cooling, and safety system precisely optimized for that reactor's dimensions and thermal output. Samsung Heavy's compartmental approach separates the reactor from the power generation systems. The nuclear island sits in its own compartment with standardized interface connections to the power island (turbines, generators, grid systems). In principle, a buyer can select a different SMR type by changing only the nuclear compartment, without rebuilding the rest of the platform. This matters because no SMR design has yet achieved the commercial-scale serial production that would demonstrate long-run economics — the market is still deciding which designs survive. A reactor-agnostic platform lets a customer defer that bet.
What nuclear regulatory approvals does Samsung Heavy's platform still need before it can operate in the United States?
A significant stack. The American Bureau of Shipping's Approval in Principle covers only the maritime hull and non-reactor engineering elements of the concept design — it is a feasibility assessment, not a construction or operational certificate. The nuclear reactor itself requires separate licensing from national nuclear regulators. In the US, that means a Nuclear Regulatory Commission review of the specific reactor design installed on the platform. The SMART100 holds standard design approval in South Korea only; a US deployment would require NRC review from scratch. NuScale's VOYGR, the only SMR the NRC has fully licensed to date, required over a decade of review. The Trump administration's executive orders directed the NRC to cap new review timelines at 18 months, but that mandate has not yet been tested. The partnership's decision to pursue the US market through Sargent & Lundy — which has deep NRC experience — reflects a clear-eyed reading of where the regulatory bottleneck lies.
Is there any operational proof that a floating nuclear plant can work safely over years of service?
Yes — one. Russia's Akademik Lomonosov, a barge-mounted plant carrying two KLT-40S pressurized-water reactors derived from nuclear icebreaker technology, has been commercially operational in the Arctic port of Pevek since May 2020. By early 2025, it had produced more than one billion kilowatt-hours of electricity and completed its first refueling cycle, while supplying an increasing share of the Chukotka region's power demand. The Akademik Lomonosov was built and operated by Rosatom, a Russian state nuclear company, for a state utility — not a commercial, export-oriented product. Its operational record demonstrates that floating nuclear power generation is physically viable. Whether that translates to a privately financed, standardized, commercially marketed platform is what the Samsung–Sargent & Lundy partnership is now attempting to prove.
