261102 – The Hydrogen Car Update

by | Nov 2, 2026 | Physics | 0 comments

Hydrogen Cars in 2026: Where the Dream Stands

· @Thomas Lee Abshier, ND

The hydrogen passenger car has lost the race to the battery car, and I do not expect it to come back. Hydrogen itself has not lost. Its future is in heavy machines, long-haul and remote work, and above all in chemistry, and a nuclear-powered economy would make that future larger, not smaller.

A post has been circulating on social media announcing that JCB, the British construction-equipment maker, has received approval to sell its hydrogen combustion engine. The news is real but not new. JCB’s first approvals came from the Dutch vehicle authority and ten other European licensing bodies in January 2025, and full EU type-approval for non-road machinery followed in May 2025. JCB spent roughly £100 million and three years with 150 engineers to build it. The post is right that the engine emits no carbon dioxide at the tailpipe. It leaves out that a hydrogen flame in air still makes nitrogen oxides, and that nearly all hydrogen sold today is made from natural gas.

The post sent me back fifty years. In the early 1970s at UCLA, in Boelter Hall, engineering students were building a car that ran on hydrogen. I have wondered since then why that idea never arrived. If we had abundant nuclear energy, could we split water, store the hydrogen, and drive on it, in fuel cells or in engines, or in hybrids of both? This article is my attempt to answer that question with the physics and the market as they stand in November 2026.

The car I remember: UCLA’s hydrogen Gremlin, 1972

The car I remember was a yellow 1972 AMC Gremlin, donated by American Motors and converted by UCLA engineering students and faculty for the 1972 Urban Vehicle Design Competition. Sixty-four North American universities entered that contest, held at the General Motors proving ground in August 1972. The UCLA car won the prize for lowest emissions; its exhaust was essentially water vapor. The project leader was a student named Frank Lynch, with Joe Finegold, Ned Baker, and others, and the team reported the work to the Society of Automotive Engineers in 1973 (SAE paper 730507, The UCLA Hydrogen Car: Design, Construction, and Performance).

The Gremlin was an internal combustion car, not a fuel cell car. It burned hydrogen in a modified piston engine, which is the same basic approach JCB took fifty years later. Lynch said decades afterward that the car proved hydrogen could run a car in a practical sense, but that there was no direct line from it to today’s prototypes. That is a fair summary of the whole half-century: hydrogen has been shown to work many times, and has never been shown to pay.

Two ways to drive on hydrogen

There are two ways to get motion out of hydrogen, and they behave very differently.

The fuel cell car (Toyota Mirai, Hyundai Nexo, Honda CR-V e:FCEV) is an electric car. A fuel cell combines hydrogen with oxygen from the air across a membrane and produces electricity, heat, and water. The electricity drives a motor, with a small battery as a buffer. The only exhaust is water. The cell is roughly 50 to 60 percent efficient at turning the hydrogen’s energy into electricity.

The hydrogen combustion engine (the UCLA Gremlin, BMW’s hydrogen 7-Series of the 2000s, Toyota’s hydrogen racing Corollas, and JCB’s engine) burns hydrogen in cylinders much like gasoline or diesel. It is cheap to build because it reuses a century of engine know-how. But it is less efficient, roughly 35 to 40 percent at best, and because the flame is hot and the air is 78 percent nitrogen, it makes nitrogen oxides. Engineers manage this by running the engine very lean, which cuts power, or by adding exhaust aftertreatment as on a modern diesel. A trace of carbon dioxide also comes from burned lubricating oil.

One point of terms is worth getting straight. A fuel cell consumes hydrogen; an electrolyzer makes it, by running electricity through water. The two are the same reaction in opposite directions, and some high-temperature solid-oxide cells can run either way. But in a vehicle, the fuel cell uses hydrogen, never the source.

Both kinds of car share hydrogen’s basic storage problem. Hydrogen holds about three times the energy of gasoline per kilogram, but it is the lightest gas. Even squeezed to 700 times atmospheric pressure, a liter of it holds only about one-seventh the energy of a liter of gasoline. That is why fuel cell cars carry heavy carbon-fiber tanks, why the gas must be compressed or chilled to −253 °C at great energy cost, and why hydrogen leaks through seals and weakens some steels.

Where hydrogen cars stand in 2026

The world bought about 16,000 hydrogen fuel cell cars in 2025, against more than 20 million battery-electric cars. Fuel cell sales peaked at about 20,700 in 2022, fell to 12,900 in 2024, and recovered in 2025 almost entirely because Hyundai launched a new Nexo in South Korea. In the first half of 2026, global sales were about 4,600; Korea grew while Toyota, Honda, and China fell.

The fueling network is the deeper problem, and it is shrinking. A car that can refuel at only a few dozen stations is a car few people will buy, and stations without nearby customers are unlikely to survive.

  • California, the only real U.S. market, has had about 50 retail stations. Shell closed all seven of its car stations in early 2024. On February 24, 2026, an explosion at a hydrogen trailer yard in Colton killed a technician, and the safety stand-down that followed froze trucked supply; by late March, 32 of 52 stations were closed, and Mirai and Nexo owners waited in long lines at the few that remained. Pump prices run about $10 to $16 per kilogram.
  • Japan, which once planned for 800,000 fuel cell cars by 2030, sold just 431 in 2025, down 83 percent from 2021. Its station count fell to 149, and the government is cutting the fuel cell car subsidy while raising the battery car subsidy.
  • Europe: Austria’s OMV left the hydrogen station business entirely in 2025, and Germany’s H2 Mobility closed roughly 30 percent of its network.

At about 70 miles per kilogram, a Mirai driver paying $13 a kilogram spends roughly 19 cents a mile on fuel. A 30-mpg gasoline car at $3.50 a gallon spends about 12 cents, and a battery car charged at home at 17 cents per kilowatt-hour spends about 5 cents. (These are my round-number estimates; local prices vary.)

Company Where it stands on hydrogen vehicles
Hyundai Committed: new Nexo (Korea 2025, Europe 2026), fuel cell plant in Ulsan, Xcient fuel cell trucks
Toyota Shifting from private cars to commercial use: Tokyo fuel cell taxis (600 targeted by 2031), trucks; still testing hydrogen combustion engines
BMW Plans series production of a hydrogen iX5 in 2028, with Toyota fuel cell technology
Honda Sells the CR-V e:FCEV in small numbers; postponed its next fuel cell module plant
General Motors Halted next-generation fuel cell development; ended joint production with Honda
Stellantis Ended its fuel cell program in July 2025, calling hydrogen a niche with no mid-term economic sustainability
Renault Ended its Hyvia hydrogen van venture with Plug Power in 2025

U.S. policy has also turned. The 2025 budget law moved the deadline to begin construction for the clean-hydrogen production credit (section 45V, up to $3 per kilogram) from 2033 to the end of 2027, and the Department of Energy canceled $2.2 billion for the two West Coast hydrogen hubs in October 2025.

The energy chain: why the battery wins the passenger car

The market’s verdict follows from physics, not fashion. Every energy conversion loses some, and hydrogen requires more conversions than a battery. Start with the same kilowatt-hour of electricity and follow it to the wheels:

Path from electricity to wheels Main steps (approximate efficiency) Share reaching the wheels
Battery car transmission and charging (~0.85) × battery, inverter, motor (~0.90) about 75–80%
Fuel cell car electrolysis (~0.70) × compression and delivery (~0.85) × fuel cell (~0.55) × motor (~0.90) about 25–35%
Hydrogen combustion car electrolysis (~0.70) × compression and delivery (~0.85) × engine (~0.38) × drivetrain (~0.90) about 20%

Multiply the steps and the result is plain: a fuel cell car needs roughly two and a half to three times as much electricity per mile as a battery car, and a hydrogen combustion car needs roughly four times as much. Independent studies, including work at the University of Queensland and an analysis Volkswagen published, land in the same range. Whatever electricity costs, the hydrogen car pays for it three or four times over.

Hydrogen’s real advantages are fast refueling (about five minutes) and low weight per unit of stored energy at scale. For a commuter car, those advantages matter less. Batteries have fallen roughly 90 percent in cost since 2010, most driving is under 40 miles a day, and a car can charge overnight in its own garage. A battery car brings its own fuel station home; a hydrogen car needs a whole new industry of production plants, tanker trucks and high-pressure pumps before it can leave the dealer’s lot.

Nuclear power and hydrogen: the honest arithmetic

Nuclear power is a good partner for hydrogen, but the best place for nuclear hydrogen is the factory, not the family car. The reasoning runs as follows.

Nuclear makes hydrogen cheaper in one important way. An electrolyzer is an expensive machine, and its cost per kilogram depends on how many hours it runs each year. Solar and wind feed an electrolyzer perhaps a quarter to a third of the time; a reactor runs over 90 percent of the time. Steady power spreads the machine’s cost over three times as much hydrogen.

Nuclear heat can do part of the splitting. A conventional reactor turns only about a third of its heat into electricity, so ordinary electrolysis yields hydrogen holding roughly a quarter of the reactor’s heat. High-temperature steam electrolysis, which Idaho National Laboratory has developed since 2003, supplies part of the splitting energy as heat. Coupled to a high-temperature gas-cooled reactor, INL’s design estimates that a 600-megawatt-thermal reactor could make about 2.5 kilograms of hydrogen per second, which works out to about half the reactor’s heat captured in the hydrogen. That roughly doubles the yield.

It has begun, at demonstration scale. Constellation started the first U.S. nuclear-powered hydrogen production at Nine Mile Point in New York in 2023 with a 1-megawatt electrolyzer. Davis-Besse in Ohio and Prairie Island in Minnesota, the latter with a high-temperature solid-oxide unit, were funded alongside it.

But abundance does not repeal the energy chain. Suppose a reactor makes a kilowatt-hour that costs almost nothing. Sent down the wire to a battery car, it moves the car about three miles. Turned into hydrogen and sent to a fuel cell car, it moves the car about one mile. Even with high-temperature electrolysis, the battery still delivers more than twice the miles per unit of reactor heat. Cheap energy narrows the cost gap, but it cannot close a gap set by thermodynamics, and the hydrogen path still needs its own pipelines, trucks and stations.

Nuclear hydrogen makes clear sense where hydrogen is used as a chemical rather than as a fuel. The world already makes about 100 million tonnes of hydrogen a year, almost all from natural gas or coal, for fertilizer, oil refining and methanol. Replacing that with nuclear hydrogen, and adding steelmaking and synthetic liquid fuels, would cut more pollution per reactor than putting hydrogen in cars, and it would happen at a few large sites, with no need for a fueling station on every corner.

Hydrogen hybrids for long distances

The idea of a hybrid, a battery for daily driving with hydrogen for long trips, is sound engineering, and the industry has built it. Every fuel cell car is already a hybrid with a small buffer battery. Honda’s CR-V e:FCEV goes further: it is a plug-in, charging at home for roughly 29 miles of daily driving and using hydrogen beyond that. Toyota has been road-testing a van with both a hydrogen combustion engine and a battery.

The hybrid’s weakness is the long trip itself. The hydrogen half is needed only on the highway, far from home, which is exactly where stations are thinnest and most expensive to keep open. A hydrogen highway would be many lightly used stations spread across empty country, each costing from half a million to several million dollars. Meanwhile, the electric grid already runs along every interstate, so fast chargers can be added for a fraction of that cost, and today’s long-range battery cars cover 300 miles or more between stops. For the rare driver who regularly goes beyond that, a small gasoline or biofuel range extender uses existing fuel stations.

So the hybrid solves a real problem, but by the most expensive route available. It would make sense only if hydrogen stations became common for some other reason, such as heavy trucking, and cars could ride along on a network that trucks were paying for.

Where hydrogen does belong: the heavy, the remote, and the chemical

Hydrogen earns its place where a battery would be too heavy, too slow to recharge, or too far from a plug, and where one fueling point serves many machines. That is why JCB’s engine is a sensible product even though a hydrogen family car is not.

  • Construction and farm machines. An excavator or telehandler works all day on a site with no grid connection. A battery large enough would be enormous and could not be recharged during a lunch break. A hydrogen engine refuels in minutes from a tank trailer parked on site, and JCB built it from ordinary engine parts its dealers already know how to service. JCB has also certified the engine to EU Stage V emissions limits, which addresses the nitrogen oxide concern.
  • Captive fleets. Buses, taxis, forklifts and delivery trucks that return to one depot each night can share one well-used station. Toyota’s Tokyo taxi project and Hyundai’s fuel cell trucks rest on this logic.
  • Long-haul trucks, ships and some trains. Here the contest with batteries is still open; hydrogen and its derivative fuels (ammonia and synthetic diesel) have a case where weight and refueling time decide the matter.
  • Industry. Fertilizer, refining, steel, and synthetic fuels already need hydrogen as a raw material. This is the largest and surest market, and the one nuclear hydrogen fits best.

One caution applies to all of these. A hydrogen engine is “zero carbon” only if its hydrogen is. Most hydrogen today is made from natural gas, releasing about ten kilograms of carbon dioxide for every kilogram of hydrogen. The JCB headline is a promise about the engine, not about the fuel.

Why the hydrogen car is not entirely foolish

The case for the hydrogen car still has real points, and it is worth stating them at their strongest.

  • Batteries have their own limits. They depend on lithium, nickel, cobalt and graphite mined and refined largely in China, and on a grid that would have to grow substantially to charge every car. Hydrogen can be made from water anywhere there is power.
  • Truly abundant energy changes the weighting. The efficiency argument assumes energy is scarce. If advanced reactors ever make energy as cheap as fission’s early advocates hoped, a threefold loss matters far less than convenience, and five-minute refueling becomes a real selling point.
  • Hydrogen stores energy across seasons. Batteries hold energy for hours; hydrogen can be stored for months, so a nation might build a hydrogen system for grid storage and industry anyway. Once it exists, cars could share it.
  • Serious companies have not quit. Toyota, Hyundai, and BMW continue to invest, and Toyota and Hyundai both unveiled improved fuel cell systems in 2025. They have engineering reasons, and they may see a future that sales charts do not yet show.

These points are why I would not call hydrogen cars impossible. They are why I would call them a bet that depends on two things not yet in hand: very cheap energy, and a hydrogen network paid for by someone other than car owners.

Conclusion

I stand with the battery for the passenger car, and with hydrogen for heavy machines, captive fleets, and industrial chemistry. The UCLA Gremlin proved in 1972 that a car could run on hydrogen. Fifty-four years later, the question is no longer whether it can, but whether it should, when the same electricity would carry a battery car three times as far without a new fuel network.

Abundant nuclear power would be a great good for this country, and hydrogen would be one of its best uses: made steadily at the reactor, with reactor heat doing part of the work, and sent to fertilizer plants, refineries, steel mills and synthetic-fuel plants. Some of that hydrogen may one day fuel trucks and machines like JCB’s. I expect very little of it will end up in the family car.

Sources and their limits

The sales, station and policy figures come mostly from trade press and industry analysts (SNE Research, Hydrogen Insight, Automotive World), company press releases, and news reports. These lean toward the mainstream and industry record; hydrogen boosters and battery advocates each publish selectively, and CleanTechnica in particular argues against hydrogen cars. The efficiency table is my own arithmetic using typical published component efficiencies; individual vehicles and plants vary by several percentage points. Background figures on global hydrogen production, carbon released by natural-gas reforming, battery cost decline, and storage density are standard reference values not tied to a single source below.

Submit a Comment