Romania’s Hydrogen Train Deal Reveals a Governance Failure, Not a Technology Win

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

Romania’s award of a contract for 12 hydrogen trains to Siemens Mobility looks, at first glance, like a late but determined embrace of cleaner regional rail. Read more closely, it looks like something else. It looks like a governance failure made visible. The contract was awarded only after repeated failed tenders, after the original financing window under Romania’s National Recovery and Resilience Plan had effectively been lost, and after the authority had to acknowledge that it was seeking a new funding source. That is not what successful technology selection looks like. It is what institutional path dependence looks like when a public body keeps pushing a project forward after the market, the schedule, and the financing have all signaled trouble.

The basic facts are straightforward enough. Romania’s Railway Reform Authority, ARF, just awarded Siemens a contract covering 12 hydrogen fuel cell trains with long-term maintenance, with reported contract values ranging from about RON 1.6 billion to RON 2.5 billion (€314 million to €490 million) depending on maintenance duration. Public summaries of the intended service pattern focus on Bucharest-radiating non-electrified routes such as Bucharest to Henri Coandă Airport, Bucharest to Târgoviște, Bucharest to Pitești, and Bucharest to Curtea de Argeș via Pitești, with some coverage also mentioning westward extension toward Craiova. The procurement is being presented as a way to decarbonize regional rail without the full cost of electrifying those lines. But the country no longer has the original PNRR-backed funding lined up for the deal, and that matters because financing discipline is part of technology discipline. A country that chooses a technology cleanly does not usually lose the money, rerun the tender multiple times, and then award the contract while still hunting for replacement financing.

The year-by-year history makes the point better than any abstract critique. In 2021, the idea of hydrogen trains entered Romania’s policy conversation as a promising modern option for cleaner rail. In 2022, ARF launched market consultation for hydrogen electric multiple units, which is a normal early step in a procurement pipeline. In April 2023, ARF signed the PNRR financing contract for the acquisition of 12 hydrogen units. But in the same general period, the first formal tender failed because no bids were submitted. Romania relaunched the tender and received only one bid, from the Alstom and Linde Gaz Romania association, and that second procedure was cancelled in November 2023 because the sole offer was non-compliant. ARF relaunched again in December 2023. The third attempt collapsed in May 2024 after no bids were submitted. By August 2024, Prime Minister Marcel Ciolacu was publicly saying Romania would abandon the purchase of hydrogen trains through the PNRR and redirect the money to highways, with some reporting quoting him saying, “Because there are no hydrogen trains.” Yet in December 2024 the authority launched a fourth tender. In April 2025 Siemens submitted the first eligible bid. In April 2026 Siemens won the tender, after the original financing window had been missed and ARF had to say it was seeking a new source of money. That is not one procurement bump. It is a multi-year institutional struggle with repeated warning signs.

Those repeated failures are not background noise. They are the market speaking. Public procurement is a market test. If a government puts a well-designed, financeable, operationally credible project into the field, suppliers usually show up. They may compete hard, or they may not, but they show up. Romania ran this hydrogen train test four times. Twice it got no bids. Once it got a non-compliant bid. Only on the fourth round did it get a single viable bidder. That kind of market behavior usually points to a bundle of problems rather than one isolated issue. The technology may look risky. The risk allocation may look unattractive. The timeline may look unrealistic. The associated fueling and infrastructure plan may look underdefined. The commercial upside may look too small for the complexity involved. Whatever the mix, the signal is the same. The supplier base did not see this as an inviting mainstream procurement. Romania did not discover a healthy hydrogen rail market. It persisted until one supplier was willing to take the deal.

One of the clearest signs that Romania is buying into a shrinking niche rather than a rising market is the retreat of the companies most visibly associated with hydrogen rail. Cummins, whose fuel cells powered Alstom’s flagship Coradia iLint fleets, sold its rail-dedicated hydrogen fuel cell activities to Alstom in April 2026, and Alstom did not present the deal as a growth push into a booming segment. It described the acquisition in operational terms, focused on supporting the installed fleet and concluding contracted programs. That is the language of consolidation and managed follow-through, not expansion. Cummins itself was once promoted as a key enabling supplier for hydrogen rail, including Lower Saxony’s 14-train fleet, so its exit from rail-dedicated fuel cell activities matters. It suggests that after years of trials, subsidies, and demonstration value, hydrogen trains still did not create the scale, margins, or confidence needed to sustain a broader supplier ecosystem. Romania’s award comes at the point where one major supplier has stepped back and the main train manufacturer is folding the remaining capabilities inward to support existing obligations, which is a poor backdrop for claiming that hydrogen rail has entered a durable commercial phase.

That is why the governance frame matters more than the technology frame. Romania is not revealing confidence in hydrogen. It is revealing what happens when public institutions become committed to a framing and continue after the evidence has weakened. Once a project acquires strategic language, policy sponsorship, budget line items, and bureaucratic labor, it becomes hard to step back and reopen first-order questions. The institution has already spent years organizing around one answer. At that point, abandoning the answer feels like failure in a way that redesigning the problem from scratch does not. Bureaucracies, like corporations, are vulnerable to sunk cost behavior. The harder and longer the procurement becomes, the more likely decision-makers are to interpret persistence as prudence instead of asking whether persistence itself has become the error. Romania’s hydrogen train award has that pattern all over it.

The deeper systems problem is that this was never just a train purchase. It was an infrastructure decision disguised as a rolling stock decision. Hydrogen is politically appealing because it looks like a way to decarbonize rail without making a hard infrastructure choice. Overhead wiring requires corridor planning, grid connection work, approvals, construction windows, capital discipline, and coordination with the infrastructure manager. Battery-electric rail requires selective wiring, charging strategy, timetable adjustment, and system planning. Hydrogen lets a ministry tell itself that it can avoid those choices by buying a more complex train. But that does not make the infrastructure problem go away. It just moves it sideways into fuel production, storage, compression, refueling logistics, maintenance complexity, and long-term supply risk. Hydrogen is not an escape from infrastructure. It is a different and often more fragile infrastructure bet. The appeal is political because the assets are less visible along the line. The costs and risks are still there.

This is where the German evidence matters. Baden-Württemberg and Lower Saxony did not approach rail decarbonization as a branding exercise. They studied and, in Lower Saxony’s case, operated the options. The VDE study on alternatives to diesel multiple units concluded that fuel cell trains were the least economical option in all modeled cases. It pointed to hydrogen cost above €5 per kg, shorter life for fuel-cell-related components, and the added cost of hydrogen fueling infrastructure. Baden-Württemberg’s subsequent work through the NVBW SteFaNS study recommended battery-hybrid trains for several subnetworks because of clear cost advantages. Press reporting on the state’s findings summarized the result even more sharply, with hydrogen trains described as up to 80% more expensive over their lifetime than electric alternatives. Germany did not conclude that hydrogen had no use anywhere. It concluded that wires made sense where justified, batteries plus selective wiring made sense across many regional networks, and hydrogen was a narrow residual niche carrying a cost premium.

Lower Saxony is even more instructive because it moved beyond modeling into operation. Its hydrogen deployment around Bremervörde, using Alstom iLint trains, was the showcase project for daily passenger hydrogen rail service. More than €85 million went into the trains, federal support was added, and the whole point was to prove that hydrogen trains could replace diesel on a real network. But the lesson Lower Saxony appears to have drawn was not to scale hydrogen outward. In March 2024 the state announced support for 70 battery-electric trains, including 40 for LNVG, as part of the next stage of diesel replacement. Reporting on the state’s decision-making indicates that batteries and electrification won on cost grounds for the remaining networks. Meanwhile the flagship hydrogen fleet itself ran into reliability and supply issues. Industry reporting described problems obtaining hydrogen reliably, service substitutions with diesel and buses, and later shortages of fuel-cell replacement modules that cut train availability sharply. At present only 4 of 14 trains are operating. Lower Saxony was the strongest European test of hydrogen rail. It turned into an argument for batteries.

Romania’s choice, then, is not simply a national experiment. It is a late move into a segment that its strongest European exemplars are already moving away from. That matters because some technology choices deserve patience while the field matures. Others deserve skepticism when the field matures in the opposite direction. Hydrogen rail now looks more like the second case. The European pattern is not one of accelerating adoption based on field success. It is one of narrowing use cases, rising scrutiny, and greater preference for batteries plus wires. Romania is not stepping confidently into a growing mainstream. It is buying into a residual niche at the moment the niche is being reclassified.

The economics of the Romanian contracts for electric vs hydrogen trains reinforce that point. The Alstom electric train contract for 37 electric multiple units plus 15 years of maintenance was reported at about RON 2.422 billion excluding VAT, or roughly €486 million. The Siemens hydrogen contract is reported around €325 million for 12 trains plus a similar term of maintenance. On a crude per-trainset basis, the electric trains come out at around €13.1 million each, while the hydrogen trains are around €27.1 million each. That is more than double per trainset. It is not a perfect apples-to-apples comparison because the electric trains run on electrified lines and the hydrogen trains are meant for non-electrified routes. But the comparison still tells readers something important. Hydrogen rolling stock is expensive before any one asks where the hydrogen will come from.

That caveat about line electrification is important, because the real comparison is not hydrogen versus conventional electric trains on already wired corridors. The real comparison is hydrogen versus batteries plus selective wiring and charging strategy on the targeted non-electrified corridors. The Romanian route list does not obviously fall outside that decision frame. Bucharest to Henri Coandă Airport is about 19 km. Bucharest to Târgoviște is around 75 to 80 km. Bucharest to Pitești is about 108 km. Bucharest to Curtea de Argeș via Pitești is on the order of 140 km. Those are not all trivial distances, but they are very much within the range band that European studies have examined when deciding between batteries, selective electrification, and hydrogen. Siemens and Alstom already offer 120 km range electric train sets. This is exactly the terrain where Germany’s evidence base became skeptical of hydrogen and supportive of battery-hybrid solutions with targeted overhead line investment. Romania did not pick hydrogen on some uniquely long, isolated desert railway. It picked hydrogen on routes that look like the sort of routes that make the battery-plus-wires question unavoidable.

The operational question that Romania has not answered clearly is where the hydrogen comes from. Early procurement reporting indicated that the project included associated hydrogen supply and fueling facilities, and the 2023 Alstom bid paired Alstom with Linde Gaz Romania, which strongly suggested that fuel and fueling infrastructure were part of the package. But public summaries of the 2026 Siemens contract focus on the trains and maintenance, not on a clearly identified hydrogen supply plan. I was not able to verify authoritative public documentation stating whether the intended hydrogen would be green, blue, gray, or otherwise, or whether the current contract includes fuel supply or leaves it to a separate arrangement. That silence matters. Hydrogen rail is often discussed as if the fuel molecule will simply appear on schedule, at the right purity, in the right pressure regime, at the right price, every day for decades. That is not how infrastructure works. If fuel sourcing is not plain in the public record, then the project is carrying unresolved operational risk.

Romania’s broader hydrogen landscape does not make the answer simpler. The country has some green hydrogen projects and ambitions, but not a large mature pipeline by European standards. OMV Petrom is developing a 20 MW green hydrogen project at Petrobrazi and has described related work tied to lower-carbon fuels. Rompetrol has described a partner-seeking green hydrogen concept tied to the Petromidia refinery. Romania’s strategy documents and summaries talk about targets for electrolyzer capacity by 2030 and support under the NRRP. But that is not the same thing as having a deep bench of operating green hydrogen supply available to support a rail system with daily service obligations. At the same time, much of Romania’s existing industrial hydrogen still comes from conventional steam methane reforming. That means the easy rhetorical move, “the trains will run on green hydrogen,” cannot be taken for granted. It needs an actual project, price, and delivery chain behind it.

Across Europe, hydrogen sourcing has repeatedly turned out to be the hidden fragility in hydrogen transport projects, not just for trains but for buses and ferries as well. In Lower Saxony’s flagship hydrogen rail network, Linde’s technical problems with hydrogen supply contributed to train cancellations and substitutions with diesel units and buses in 2024, while later fleet availability problems compounded the operational strain. In Polish bus fleets, Poznań temporarily withdrew its hydrogen buses after onboard diagnostics flagged faults that reports linked to concerns over hydrogen quality. In ferries, many of the better-known European hydrogen vessel projects are still at the demonstration or first-of-a-kind stage, which is another way of saying the fuel chain has not matured into a routine, bankable operating input in the way marine diesel, shore power, or even LNG have. The pattern is not that Europe lacks hydrogen announcements. It is that getting the right hydrogen, at the right purity and pressure, in the right place, every day, at an acceptable cost, remains hard enough that the fuel supply itself keeps becoming a service risk.

How much hydrogen are we talking about? Not an industrial cluster amount, but not a trivial demonstration amount either. Siemens says a two-car Mireo Plus H can reach up to 1,200 km range, and reporting around the platform indicates about 180 kg of hydrogen per fill. That implies a rough planning consumption of around 0.15 kg of hydrogen per km. Using current broad service levels on the named Bucharest-centered corridors as a proxy, the likely fleet demand lands around 0.8 to 1.0 tons of hydrogen per day. The airport shuttle alone could need around 180 to 205 kg per day. Bucharest to Pitești could need around 345 to 389 kg per day. Bucharest to Târgoviște could need around 192 to 216 kg per day. Curtea de Argeș at sparse service could add around 114 to 129 kg per day. Add those together and the system lands in the neighborhood of a ton per day. The number is approximate because ARF has not published a full operating diagram tied to the hydrogen fleet, but it is large enough to turn “where will the hydrogen come from” from a rhetorical question into an operational one.

If Romania wanted to cover that demand with electrolysis, the size of the electrolyzer is not enormous. At 50 to 55 kWh per kg of hydrogen, 800 to 1,000 kg per day translates into roughly 1.7 to 2.3 MW of continuous electrolyzer load. A real project would likely want more like 2.5 to 3 MW installed once compression, storage, downtime, and operating margin are considered. That is small enough to show that the rail project is not some massive hydrogen economy anchor. But it is also large enough to require a real asset, a real connection, a real siting process, and real economics. It is not nothing. And that creates a question that goes to the heart of the governance critique. If the required hydrogen production is only in the low single-digit megawatt range, why buy a highly specialized hydrogen fleet at all instead of buying battery trains and solving the corridor problem more directly with selective wiring and charging. A small electrolyzer requirement does not rescue the hydrogen case. It makes the alternative look even more ordinary and accessible.

One operational point in hydrogen’s favor is that the currently named routes are mostly Bucharest-centered, so a single Bucharest-area fueling facility appears plausible for the initial network. Given the train’s advertised range, multiple daily round trips could be completed before returning to a common refueling location. But that is a limited comfort. A single-facility arrangement still requires dependable fuel supply, compression, storage, maintenance, backup plans, and operational resilience. If the network later extends westward toward Craiova, the neatness of a one-depot story declines. I was not able to find authoritative public confirmation of a final depot and refueling plan, so even this relatively favorable assumption sits in the category of plausible rather than confirmed. That again is part of the story. The further one looks into the project, the more one finds broad claims and fewer concrete answers.

The climate case becomes even less comfortable if green hydrogen does not materialize and the trains run on gray hydrogen instead. Using the Siemens consumption estimate of roughly 0.15 kg H2 per km, and using the IEA’s estimate that gray hydrogen from unabated natural gas carries around 10 to 12 kg CO2e per kg of hydrogen, the climate performance is no longer remotely “zero emission.” Add hydrogen leakage to the calculation and the picture worsens. Recent literature places hydrogen’s indirect warming impact around GWP20 of roughly 37 and GWP100 of roughly 11, and leakage reviews indicate that real supply-chain leakage rates can be a few % rather than trivial fractions. Under reasonable assumptions, gray hydrogen rail lands around 1.5 to 2.0 kg CO2e per train-km. A comparable two-car diesel regional train, using Great Britain’s diesel passenger traction averages as a proxy, lands around 2.7 kg CO2 per train-km from combustion alone and around 3.2 kg CO2e per train-km with a modest upstream allowance. That means gray hydrogen is probably somewhat better than diesel on a fuel-cycle basis, but only by a middling margin, and the advantage shrinks when leakage is counted, especially over a 20-year horizon. Gray hydrogen rail is not decarbonization. It is partial improvement dressed in zero-emission language.

At fleet scale, the climate arithmetic sharpens the point. For a Romanian hydrogen fleet burning around 0.8 to 1.0 tons of hydrogen per day, gray hydrogen would imply something like 8 to 14 tons of CO2e per day depending on hydrogen carbon intensity, leakage, and whether one uses GWP20 or GWP100. A comparable diesel service would be higher, but not so much higher that one could honestly keep calling the hydrogen trains “clean” without qualification. Once the fuel pathway stops being an abstraction and starts being a real procurement and logistics question, the rhetorical gap between the image of hydrogen rail and the likely reality becomes harder to ignore. That does not mean the project will necessarily run on gray hydrogen. It means the governance failure includes putting a public decarbonization story on the table without making the fuel pathway plain enough to test.

Another uncomfortable piece of this story is that Romania’s contract award appears to have survived the loss of the original funding in part because the tender documents had already anticipated financing uncertainty. Reporting on the December 2024 relaunch said the authority reserved the right to cancel if the financing contract was not signed or if state budget funds could not be approved. In other words, the procurement architecture itself recognized that money might not be there. That does not make the award unlawful or irrational on its own. It does underline that the country has been working for some time inside a framework where financing instability was not a surprise. Good governance would normally treat that as a reason to reopen the core decision. Instead, Romania appears to have treated it as a reason to keep the procedural machinery moving while hoping the financing could be reconstructed around the project later.

That brings the story back to the central diagnosis. Romania is not buying a technology. It is revealing a governance failure. The failure is not that the country wants cleaner rail. That ambition is sound. The failure is not even that Romania explored hydrogen. Many countries did. The failure is that after years of failed tenders, loss of the original financing window, weak supplier appetite, uncertain fuel sourcing, and a growing European evidence base pointing toward batteries plus wires, the institutional response was to keep pushing the hydrogen project forward rather than step back and reframe the question. Romania did not arrive at hydrogen because the evidence became compelling. It arrived at hydrogen because the bureaucracy stayed committed after the evidence weakened.

With the contingent nature of the contract, the high expense, the clear success of electric trains and the lack of any apparent green hydrogen, I only give the odds of the contract actually being exercised and trains being delivered a 30% chance. And if they are delivered and put into operation, it’s only 5% likely that they would still be running after ten years. Globally, hydrogen transit fleets are much more notable for being abandoned than perpetuated.

That is why this case matters beyond 12 trains. It is a study in how governments buy the wrong complexity for the wrong reasons. They define the problem too narrowly as a vehicle procurement. They separate rolling stock from infrastructure strategy. They overvalue symbolic innovation. They underweight market signals when tenders fail. They treat financing slippage as a procedural headache rather than a warning that the project concept itself needs review. They let a modern-sounding answer displace a disciplined system comparison. Once that sequence is underway, the final award can look like progress even when it is really the residue of an earlier decision that was never properly reexamined. Romania’s hydrogen train deal is not important because it will transform European rail. It is important because it shows how public institutions keep turning decarbonization into a branding exercise when the harder, more valuable work is infrastructure planning, option comparison, and governance that is willing to change course when the evidence changes.