Hydrogen at Sea Just Got Even More Expensive: What DNV’s Safety Findings Mean

Support CleanTechnica’s work through a Substack subscription or on Stripe.

The final DNV study for the European Maritime Safety Agency (EMSA) matters because it changes the hydrogen shipping debate from an argument about theoretical decarbonization potential into an argument about engineering burden and cost. DNV is not saying hydrogen-fuelled ships are impossible. It is saying that hydrogen requires a design-based safety regime with stronger barriers than LNG, not to mention stronger barriers than alcohol fuels or batteries. That sounds like a regulatory nuance, but it is really a capital allocation signal. Every extra layer of containment, detection, isolation, ventilation, hazardous-area treatment, validation, and operating discipline has to be paid for, maintained, inspected, and proven to class and flag satisfaction. In shipping, safety architecture is never a side issue. It is embedded in steel, controls, procedures, and downtime risk. DNV’s report makes that impossible to ignore.

It is worth being precise about what the study actually says. The EMSA guidance is explicitly non-mandatory and advisory. Its purpose is to support practical design solutions for ships using hydrogen as fuel, not to prohibit them. At the same time, it differs from the draft IMO interim guidelines in a way that matters commercially. EMSA’s hydrogen guidance recommends that all hydrogen leak sources be protected by secondary enclosures, including on open deck, specifically to prevent leaks from spreading uncontrollably. That is not a minor change. It means the emerging conservative view is that buoyancy and open-air dispersion are not enough to carry the safety case on their own. If that becomes the accepted design direction, hydrogen’s cost floor moves upward across the maritime sector.

The engineering reason is simple. Hydrogen is not just another gas. It leaks easily, ignites with very little energy, and can form dangerous clouds quickly. DNV’s public summary says that design-based safety is required because incidents can escalate faster than detection and response. In other words, hydrogen’s problem at sea is not only the probability of a leak. It is the combination of leak behavior, ignition sensitivity, and response time. A fuel that can move from leak to dangerous concentration before a human operator or a contact sensor can intervene pushes designers toward passive and automatic protections. Those protections are expensive, and the cost is not confined to one subsystem. It spills into layout, piping, instrumentation, commissioning, crew procedures, bunkering interfaces, maintenance plans, and survey regimes.

That is why the cost conversation has to be broken into layers. The first layer is direct equipment cost. Secondary enclosures, double-walled or otherwise protected pipe runs, tank connection spaces, vent masts, automatic shutdown valves, redundant gas detection, fire suppression, pressure management, and control systems all add capex. The second layer is engineering cost. Hydrogen systems require more detailed quantitative risk assessment, more CFD work around leak dispersion and explosion consequences, more class engagement, more documentation, and often more iterations in design approval. The third layer is operating cost. Specialized inspections, calibration, crew training, emergency drills, purging procedures, spare parts, shutdown testing, and bunkering choreography all consume labour and availability. The fourth layer is infrastructure cost. A battery ferry can often rely on electricity and charging equipment. A methanol ferry can bunker a liquid fuel through a comparatively familiar logistics chain. A hydrogen ferry needs an entire hydrogen value chain to be in place and working correctly.

That cost hierarchy matters when hydrogen is compared with actual alternatives instead of an abstract decarbonization wish list. Methanol and ethanol are not risk free, but they are governed by interim IMO guidelines specifically written for methyl and ethyl alcohol as fuels. DNV’s 2025 methanol report describes methanol-fuelled engines and technical systems as having reached a high readiness level, with compatibility with existing port infrastructure reducing complexity and cost for shipowners. Batteries have their own hazard set, especially thermal runaway and off-gas management, although both are diminishing rapidly with runaway prevention engineering and new battery chemistries, but EMSA’s battery safety work treats those as concentrated shipboard risks with a defined architecture of battery management systems, fire boundaries, ventilation, and suppression. Hydrogen, by contrast, is a distributed safety problem. It spreads across storage, transfer, conditioning, supply, and bunkering, and the new DNV work is telling the market that the conservative answer is more containment and more automation, not less.

That has implications far beyond ferries. Ferries are among hydrogen’s most favourable use cases because they run fixed routes, have predictable schedules, and can rely on dedicated shore infrastructure. If hydrogen struggles to be competitive there, the picture does not improve when moving to looser operating patterns, more ports, and more fragmented ownership structures. The Faraday Institution’s 2025 maritime battery overview makes the split clear. Full electrification is increasingly suited to short routes, while hybrids are emerging for longer distances. That is not a statement that batteries solve every ship. It is a statement that for the short-sea segment, where hydrogen has often been presented as an obvious fit, batteries and battery hybrids keep advancing while hydrogen keeps accumulating extra system cost and engineering conditions.

The Vestfjorden, often referred to through the Lofoten ferries, is where the abstractions turn into steel and money. The Norwegian Public Roads Administration’s contract with Torghatten Nord is worth roughly NOK 4.98 billion (€450 million) over 15 years. The two new ferries are about 117 to 120 metres long, carry 599 passengers, 120 cars, and 12 trucks each, and are expected to operate at 16 to 17 knots on a route of about 90 to 100 km that takes roughly 3.5 to 4 hours. Public project material says the ferries are to operate on a minimum of 85% hydrogen and a maximum of 15% biofuel, and that they will consume about 5 to 6 tons of hydrogen per day. PowerCell’s fuel-cell supply contract alone was valued at €19.2 million, with about 13 MW of installed power across the two vessels. The dedicated Bodø hydrogen project, backed by GreenH and Luxcara, is planned with a 20 MW electrolyser and up to 3,100 tons of annual green hydrogen production.

Those numbers are enough to sketch the scale of the economics, and they align with the energy cost framing in my recent Lofoten analysis. At 5.5 tons of hydrogen per day, the ferries would consume about 2,000 tons per year, around 65% of the first phase output of the Bodø plant. Hydrogen contains about 33.3 kWh/kg on a lower heating value basis, so 5.5 tons per day means about 183 MWh of chemical energy each day. If the fuel cells deliver around 55% conversion efficiency, that becomes about 101 MWh per day of useful shipboard electricity. Because hydrogen is meant to supply 85% of annual energy, the implied total annual propulsion and hotel load is around 43 GWh. The electricity input into the hydrogen plant is also easy to bound. A 20 MW electrolyser running through the year implies about 175 GWh of electricity for up to 3,100 tons of hydrogen, or roughly 56.5 kWh/kg. Using Statistics Norway’s average Q2 2025 electricity price of 40.1 øre/kWh for energy-intensive industry, the electricity portion alone works out to about NOK 22.7/kg, which translates to roughly NOK 45 million (€4.0 million) per year for the ferry service before capital recovery, plant O&M, compression, margin, and bunkering losses. That is before anyone talks about safety.

The public design philosophy for the Lofoten ferries makes the DNV report especially relevant. As reported by The Motorship in 2023, the Norwegian Ship Design Company’s concept placed all hydrogen installations in a concentrated area on the uppermost deck and fed no hydrogen below deck, explicitly so that any leaked gas would disperse upward and away from the vessel. The project’s later public technical deck also showed vent masts, cofferdams or isolating spaces, ten 20-foot storage containers holding about 5,000 kg of hydrogen at 350 bar, independent emergency shutdown, gas and fire detection, redundancy strategies, water-based fire protection, and direct ship-to-plant bunkering. Lloyd’s Register granted Approval in Principle in 2022, later confirmed the vessels were set for LR class, and LR’s own hydrogen rules were written to support a risk-based design process and flag administration acceptance under alternative design arrangements. This is not a casual project. It already had a heavy safety case.

But it is exactly that dispersion logic that now looks less comfortable. The new DNV guidance says the more conservative route is to assume that all hydrogen leak sources should be protected by secondary enclosures. That does not prove the Lofoten ferries are unsafe. Lloyd’s Register and the Norwegian Maritime Authority are the relevant approval bodies, not DNV. It does mean, however, that one of the project’s most visible public design assumptions is now under pressure from the latest sector-wide safety work. If a design case leans materially on unobstructed upward dispersion as the primary barrier for open-deck hydrogen systems, and the emerging guidance is moving toward physical secondary containment, then either the safety case has to show that existing barriers already achieve an equivalent level of safety, or the design has to move closer to the more conservative position. That kind of tension is where change orders come from.

Recent 2025 studies of hydrogen releases in semi-confined and partially open environments reinforce the same point emerging from the maritime safety work, which is that openness does not reliably eliminate risk. Experimental work such as Runefors et al. on hydrogen releases in semi-confined vehicle compartments shows that even in relatively open and lightly obstructed geometries, ignition can still produce measurable overpressure and rapid flame acceleration, with outcomes strongly dependent on leak size and local conditions. Complementary work by Yoon et al. on hydrogen-air explosions in semi-confined structures demonstrates that venting can reduce external overpressure to low kilopascal levels, but also that more confined or poorly vented cases can escalate to damaging pressure and fragment hazards. The consistent finding across these studies is not that all semi-confined hydrogen events become catastrophic, but that dispersion alone is not a sufficiently reliable primary safety barrier. Ignitable clouds can form quickly and, as highlighted in the EMSA and DNV maritime safety work, often faster than detection and shutdown systems can respond, which is why current guidance is shifting toward physical containment, controlled venting, and automated isolation rather than relying on hydrogen to dissipate safely into open air.

The likely rework, if it happens, is not a complete redesign of the ships. It is more likely to be concentrated in the hydrogen process plant on the upper deck. MAN Cryo was contracted to provide detail design for the bunkering systems, hydrogen piping, and vent masts, and all hydrogen equipment is being installed in Norway during outfitting. That makes changes possible, but not cheap. The probable scope would include enclosure treatment for open-deck leak sources, revised vent routing, additional instrumentation, more isolation logic, possible inerting provisions, updated CFD and QRA, and a new round of class and flag review. My estimate for a targeted rework case is NOK 50 million (€4.5 million) to NOK 150 million (€13.5 million) for the pair, with a schedule impact of roughly four to nine months. A low-end paper and instrumentation case could be much smaller. A full redesign could be much worse. But the middle case is the one that matters, because it is credible and it would still land on top of a project that was originally aimed at 2025 service and is now framed around 2026.

Even without rework, hydrogen’s safety burden makes it harder to defend against alternatives. In my assessment of the Lofoten route, I argued that Norway already solved ferry decarbonization with batteries and then chose to treat Vestfjorden as a hydrogen industrial-policy flagship rather than a simple transport problem. Norway had roughly 70 battery-electric ferries in operation by 2025, with many more hybrids, while the Vestfjorden project required an integrated chain of hydrogen production, compression, storage, bunkering, and fuel-cell operation. The article’s core point remains sound in light of the DNV report. Hydrogen on this route is not competing with diesel alone. It is competing with a battery ecosystem Norway already knows how to buy, install, operate, and repeat.

The battery comparison becomes even stronger once the hydrogen consumption is translated into delivered energy. The 5 to 6 tons of daily hydrogen consumption imply roughly 43 GWh of annual vessel energy demand for the service. That is a demanding ferry application, but not a mystical one. It suggests a battery-heavy or battery-hybrid architecture in the tens of MWh per vessel, combined with high-power shore charging and possibly shore-side battery buffers. Norway already has the operating base for that kind of electrification, and BC Ferries has now specified up to 70 MWh of battery energy storage and shore charging above 60 MW on its new hybrid major vessels. Full battery-electric on Vestfjorden might still be challenging, but the existence of large marine batteries, containerized shore-charged marine batteries, and extreme charging power is no longer hypothetical. Hydrogen has to justify why it deserves to win over battery-heavy designs that are following the cost curve of manufacturing industries instead of custom fuel chains.

Methanol and ethanol create a different comparison, and one that is not flattering for hydrogen either. IMO already has interim safety guidelines for methyl and ethyl alcohol as marine fuels, and DNV’s 2025 methanol report says methanol systems have reached high readiness with infrastructure compatibility that can reduce complexity and cost. Ethanol has much less commercial maritime deployment than methanol, but the technical pathway is related and far more conventional than high-pressure hydrogen with fuel cells. Both alcohols remain liquid-fuel systems. They need proper tank protection, ventilation, leak detection, hazardous-area management, and crew procedures, but they do not force shipowners into the same degree of distributed containment and response-time anxiety that hydrogen now does under the DNV framing. That does not automatically make renewable methanol or ethanol cheap. It does make their safety overhead easier to understand and finance.

This is the commercial damage done by the DNV report. It does not kill hydrogen through prohibition. It kills it by hardening the engineering assumptions around it. Maritime hydrogen was already struggling with poor round-trip efficiency, high fuel cost, bespoke infrastructure, and uncertain operational durability in high-duty applications. The new safety guidance adds another structural handicap. If the conservative answer is secondary enclosures around all leak sources, including on open deck, then hydrogen ships become more expensive to design, more expensive to build, more expensive to validate, and more expensive to operate. That pushes hydrogen farther away from batteries on short-sea routes and farther away from alcohols where liquid-fuel logistics are acceptable.

The Lofoten ferries are the best case for seeing the pattern. This is a fixed route in a country with low-carbon power, strong maritime engineering, a state supportive of maritime hydrogen drive trains, public procurement leverage, a dedicated hydrogen production project, direct bunkering from the plant, and one of the most mature ferry electrification ecosystems in the world. If hydrogen still requires a custom safety philosophy, custom infrastructure, a major fuel-cell package, a dedicated hydrogen plant, possible late-stage design tension around dispersion assumptions, and continuing questions about operational economics, then the lesson is not that more hydrogen projects are needed to prove the concept. The lesson is that even under favourable conditions, hydrogen remains the most difficult and costly answer to a problem that batteries and low-carbon liquid biofuels are already addressing more directly.

That is why the DNV report feels like one more nail in the coffin of hydrogen for maritime shipping. Not the first nail, and not the only nail. Energy efficiency put nails in early. Infrastructure complexity added more. Weak economics in road transport and industrial hydrogen have been adding nails for years. What DNV has now contributed is something quieter and more damaging. It has shown that when the maritime sector treats hydrogen’s hazards seriously and follows them through to conservative design practice, the resulting system becomes harder to justify against the alternatives. On ferries, where batteries have already won the decade, that matters a great deal. Across the rest of short-sea shipping, it matters even more.