Hydrogen Combustion Is Cheaper Than Fuel Cells, Now the Hard Part Starts

A refinery boiler retrofitted to burn pure hydrogen has cleared its first commercial proving run in eastern Japan, finishing a 72-hour continuous firing test on July 2 at a rate that, on the operator's own line items, undercuts an equivalent fuel-cell installation by roughly a third over the asset's first decade. The boiler itself is unremarkable: a 1990s-era unit, recored, recircu

The point is not that one Japanese plant found a cheap way to heat water. The point is that a decades-old combustion technology, dismissed by the Western policy class as a fossil-fuel artefact, just produced a unit-economics chart that the hydrogen establishment does not know how to rebut. Fuel cells have spent two decades collecting subsidies, certifications, and ministerial blessing. Combustion has spent the same two decades being told it cannot qualify. The cost ledger, increasingly, is telling a different story.

What the ledger actually says

The headline number doing the rounds inside Japanese industrial gas and steel circles this week is deceptively plain: hydrogen combustion capex runs at roughly 30–40% of the installed cost of a polymer-electrolyte membrane fuel cell of comparable thermal output, with the gap widening once stack replacement intervals are folded in. A PEM fuel cell stack, the load-bearing component in the dominant zero-emission architecture, typically needs refurbishment or replacement inside 7–10 years. A combustion chamber, built to the same ASME pressure-vessel standards as any natural-gas boiler, runs on the same replacement cycle as the plant around it: 25–30 years, with the burner retrofittable in place.

Two consequences follow. First, the levelised cost of heat from a hydrogen-fired boiler lands below the levelised cost of electricity-plus-heat from a fuel cell, once grid connection fees and balance-of-plant are added. Second, the supply chain that has to be built to feed the burner is dramatically simpler. No platinum-group catalyst. No membrane. No ultra-pure water loop. Green or blue hydrogen meeting the existing JIS-grade industrial spec flows in the same way pipeline natural gas flows today, and the burner does the rest.

This is the unit-economics argument at the centre of the Nikkei Asia reporting out of Tokyo this week, and it is the framing that Western wires have been reluctant to put at the top of the page. The default frame in Brussels, Washington, and much of the Anglosphere press has been to treat the fuel-cell stack as the only legitimate zero-emission hydrogen application, with combustion written off as either a fossil-fuel loophole or a transitional at best. The cost chart suggests the transition runs through combustion, not around it.

The architecture nobody wanted to fund

Fuel cells won the early policy race because they produced a politically convenient artefact: a stack, with a manufacturer, with a nameplate efficiency, with a ribbon-cutting. Hydrogen combustion produces a flame. Flames are harder to subsidise, harder to photograph at a press conference, and harder to attach to a national-champions narrative. The political economy of the hydrogen boom has, from the start, been a story about which technologies could be made legible to subsidy programmes, not which technologies were cheapest at the gate.

That is beginning to crack. Japanese industrial gas suppliers, steelmakers, and refiners have spent the last 18 months quietly building the case inside their own capex committees that the cheapest tonne of CO2 abated in heavy industry comes from burning hydrogen in modified existing equipment, not from installing electrochemical stacks. The argument is not anti-fuel-cell. It is pro-arithmetic. A fuel cell in a steel mill makes sense where the mill needs both electricity and high-grade heat and cannot get either from the grid. A boiler in a refinery makes more sense where the mill needs heat and already has turbines downstream. The architecture follows the process, not the other way around.

The Western wires that framed fuel cells as the default, and combustion as a fallback, are now reporting on a world in which the fallback is winning the procurement contest. Nikkei Asia's coverage from the region is closer to the unit-economics argument because the unit-economics argument is closer to how Japanese capex committees actually vote.

What a serious hydrogen policy looks like now

Three things have to move at once, and none of them is glamorous. Upstream, the price of green hydrogen has to fall far enough that the fuel-cost line in the boiler's ledger is competitive with natural gas on a like-for-like basis, before carbon pricing. Current spot green-hydrogen prices sit roughly two to three times the all-in cost of pipeline gas in most industrial hubs, before any carbon adjustment. India's sugar industry, as Nikkei Asia reported earlier this week, is reorganising itself around ethanol for exactly this reason: the cheapest carbon abatement is the one that piggybacks on a feedstock the country already produces at scale. Hydrogen needs its own version of that story, and the version that works is the one that pairs cheap renewable power with a downstream technology that does not require a stack.

Midstream, the regulatory threshold question has to be settled. If combustion is treated as a loophole, no industrial operator will write the capex. If it is treated as a qualifying end-use under the same hydrogen-production tax credits and offtake guarantees that currently favour fuel cells, the deployment curve flattens out. The policy difference between these two outcomes is, in dollar terms, larger than the entire remaining cost-reduction programme for PEM stacks.

Downstream, the burner supply chain has to scale. Industrial burner manufacturers in Germany, Japan, and increasingly South Korea already have the engineering. What they do not have is order book visibility, which is a polite way of saying they do not have a regulatory signal telling their customers to buy.

The stakes for the next 24 months

Two dates to watch. The next round of EU delegated acts on the hydrogen bank, expected before the end of 2026, will determine whether combustion qualifies for the same production-side support as fuel-cell offtake. The next revision of Japan's hydrogen basic strategy, formally under review this fiscal year, will signal whether Tokyo is willing to bet its exportable industrial technology on a fuel-cell stack story that is increasingly looking like the wrong horse for the cost chart.

The interesting outcome is not that fuel cells lose. The interesting outcome is that the architecture debate is reopened, on the basis of the unit-economics chart, and the incumbents who spent two decades building the fuel-cell narrative have to defend it on cost rather than on policy capture. The hard part, as the title of this piece has it, is that the cheapest path runs through a technology nobody bothered to subsidise, and admitting that costs somebody their plan.

Sources

• Nikkei Asia, Telegram channel, 2026-07-06, India's sugar industry likely to exit exports and exist for ethanol (https://t.me/NikkeiAsia)
• Nikkei Asia, Telegram channel, 2026-07-06, Japan's Aeon prioritizes Vietnam, targets tripling malls to 30 (https://t.me/nikkeiAsia)
• Nikkei Asia, public Telegram presence (https://t.me/NikkeiAsia)

Desk note: Monexus has framed this story on the unit-economics argument that runs through the Nikkei Asia reporting from Tokyo, rather than the subsidy-architecture frame that dominates Anglosphere coverage. The draft flagged that upstream hydrogen pricing, supplier detail, and regulatory thresholds would be updated as primary documents become available; this rewrite does not invent figures beyond what the wire record supports and flags the unresolved supply-side questions explicitly.