The Gateway Theory

Stand at any gas station in America and watch. People lean on their cars, scroll their phones and chat while a pump feeds 15 gallons of a liquid into a tank a few feet from their knees, a liquid whose fumes, once they reach the engine, explode a few thousand times a minute. Nobody flinches. Nobody reads the warning stickers. We grew up with it.

Now say the word "hydrogen" in the same parking lot and watch faces change.

My argument is simple. The thing standing between most drivers and a hydrogen car isn't engineering and it isn't chemistry. It's familiarity, and the electric car is quietly supplying it. The research on how people get used to new technology backs that up. What it can't do is fix hydrogen's price, and I'll get to that too, because an argument that hides its weak spots isn't worth reading.

The trouble with the word "gas"

Think about the company that word keeps. Poison gas in the trenches of the First World War. The gas chambers. The gas leak on the evening news, the house that exploded, the propane tank, the smell you're told to run from. Then add hydrogen's own baggage: the Hindenburg burning over New Jersey in 1937, and the bomb that carries its name.

None of that is a scientific finding. It's my observation, and I'd bet it's yours too. What's striking is that gasoline isn't even a gas. It's a liquid you can see, splash and smell, and that visibility is part of why we tolerate it. Hydrogen is the opposite: clear, odorless, weightless and invisible even when it burns. We have been here before. In 1937 a natural-gas leak nobody could smell destroyed a school in New London, Texas, and killed nearly 300 people, most of them children. Texas answered by becoming the first state to require that a smell be added to natural gas. The gas didn't get any safer that year. It got noticeable, and noticeable was enough for people to live with it.

Electricity was the scary one

If you want to know whether people can learn to trust an invisible, deadly energy source, you don't have to guess. We already did it once, with electricity, and for a long time it went badly.

Thurber played his relative's fear for laughs, but read it again. She knew the stuff could kill you and couldn't see where it was, so she assumed it was everywhere. That is almost word for word how people talk about hydrogen now.

And here's the part that matters: electricity won without anyone learning how it works. Ask a room of adults to explain what actually moves through a wire and most can't. Physics teachers have spent decades documenting how stubbornly students believe that current gets "used up" as it goes around a circuit, and one study found even working electricians could do their jobs without a full understanding of it. Psychologists have a name for this, the illusion of explanatory depth: people feel they understand zippers, toilets and light switches right up until they're asked to explain one. We don't trust electricity because we understand it. We trust it because it's been in the wall our whole lives.

Why fear fades

Psychologists have been measuring this process for more than half a century, and the findings line up neatly with the history.

Exposure breeds liking. In 1968 Robert Zajonc showed that simply seeing something repeatedly makes people like it more, even when they can't say why. A review of 208 experiments in 1989 confirmed the effect and found it strongest for unfamiliar things, building over the first 10 to 20 exposures. (It has limits: a later review found that past a point, liking levels off and can dip. Nobody needs a hydrogen jingle.)

We fear the new and the invisible most. Paul Slovic's work on risk perception, published in Science in 1987, found that 2 things drive public fear more than actual danger: dread, meaning a risk that feels catastrophic and out of your control, and the unknown, meaning a risk that's new, hard to see and slow to show its harm. An odorless, invisible, flammable gas is close to a perfect score on both. The same research explains the way out: as a hazard stops being new, it stops scoring as unknown.

New things spread when they feel like old things. Everett Rogers, whose Diffusion of Innovations is the standard text on how new ideas catch on, found that a handful of qualities explain most of the speed. One of the biggest is compatibility, meaning how well the new thing fits what people already know and do. Others include whether you can try it first and whether you can see other people using it.

The microwave is a good example. Amana's countertop Radarange arrived in 1967 at $495 (about $4,900 today), into a country worried about radiation leaking out of the door. In 1976 only 4% of homes had one. By 2001 more than 90% did.

The self-service elevator is better still. Elevators that ran without an operator existed by 1900, and people refused to ride them alone for about half a century. It took a 1945 strike by New York's elevator operators, which stranded some 1.5 million office workers, to make building owners push hard. What finally won riders over wasn't an engineering breakthrough. It was reassurance: a big red stop button, an emergency telephone, a calm recorded voice, and advertising that showed grandmothers and small children pressing the buttons. The machine didn't change. The feeling did.

Even gasoline had to go through it. Hiram Percy Maxim, who built some of America's first cars, remembered a colleague standing next to their gasoline tricycle as nervously as if he were standing next to a ton of dynamite. Within 2 generations, nobody at a gas pump gave it a second thought.

A hydrogen car is an electric car

This is where the electric car comes in, and it's the heart of my argument.

Most people picture a hydrogen car as something exotic, a rolling tank of explosive gas with an engine that burns it. The ones on sale don't burn anything. A fuel-cell car feeds hydrogen from its tank into a fuel cell, where it combines with oxygen from the air to make electricity. That electricity drives an electric motor, with a small battery in between to store energy from braking. The only thing out of the tailpipe is water.

Look at what that means. The silence, the instant shove off the line, the regenerative braking, the dashboard showing kilowatts instead of rpm: all of it is the EV experience. The difference is where the electricity comes from. An EV carries it in a battery that you fill from a plug. A fuel-cell car makes it on board from a tank that you fill at a pump, in about 5 minutes.

So every EV on the road is doing hydrogen a favor. It is making electric drive normal, and it is doing it at scale: about 1 new car in 17 in America is now a battery EV, nearly 1 in 4 in Europe, and in China, plug-in cars were nearly 2 in 3 new sales in August. Every Uber ride in a Tesla, every neighbor's Ioniq, every quiet crossover pulling away from a light is an exposure. Zajonc would call it repetition. Rogers would call it observability. I'd call it the warm-up act.

The EV teaches people the driving. Hydrogen only has to teach them the pump.

The industry is already building the bridge. Honda's CR-V e:FCEV is a plug-in car with a fuel cell: it runs 29 miles on its battery from a wall socket, then switches to hydrogen for the rest of a 270-mile range. It is, quite literally, an EV with a hydrogen chaser. BMW plans to start series production of its first hydrogen car in 2028, with a fuel cell developed alongside Toyota.

The evidence it works

This isn't only theory. When researchers at UC Davis put 182 people through hydrogen ride-and-drive events, more than 80% came away with a positive view, and the researchers concluded that even brief exposure improved how people rated hydrogen's performance and safety. In a separate fleet trial, drivers' rating of hydrogen cars as being as safe as gasoline rose the longer they drove them, from 3.74 to 4.18 out of 5. And in 2026, a study of car buyers in Beijing tested almost exactly the idea in this article. People who already owned a battery EV expected more from hydrogen cars than people who didn't.

The fear, it turns out, is also thinner than the headlines suggest. When London ran hydrogen buses, fewer than 1% of the people surveyed mentioned the Hindenburg, and more than 70% supported the buses. In a national survey in Australia, 9% of people linked hydrogen with bombs or weapons; far more thought of chemistry or water. The bigger barrier across these studies is plain unfamiliarity. In a European survey of more than 7,000 people, fewer than half were even aware of hydrogen as an energy technology.

Where my theory gets complicated

Here's where the argument needs its fair hearing.

EV owners may fall in love with the plug. The same psychology that makes new things feel normal also makes whatever you have now feel right. Economists call it status quo bias. For an EV owner who charges in the garage every night, home charging is the new status quo, and a trip to a pump is a step backward. When researchers asked 39 Tesla owners about hydrogen, their biggest objections were the lack of stations and the fact that most hydrogen is made from fossil fuels. Of the 39, 10 still raised safety.

California's hydrogen buyers didn't come from EVs. A UC Davis study of 906 fuel-cell households found a third had owned a hybrid before, while only 15% had owned a battery EV. More of them than of battery-EV buyers rented or lived in apartments, without a place to charge. That's a twist worth thinking about: hydrogen's most natural customer may be the driver the EV left out, the person with no garage and no plug.

Forcing it backfires. Britain planned to switch the village of Whitby to hydrogen heating. In 2023 the plan was shelved for lack of local support, with residents citing safety, cost and being made to switch. Slovic predicted exactly that: a risk you choose feels smaller than a risk someone chooses for you. Hydrogen will be adopted the way the microwave was, one willing buyer at a time, or not at all.

What familiarity can't fix

Now the uncomfortable part. Even if every driver in America made peace with hydrogen tomorrow, most still wouldn't buy a hydrogen car, because of the bill.

California's hydrogen went from $14.95 a kilogram in the spring of 2022 (about $17 today) to $35.47 by the spring of 2025. The state's energy commission put it plainly: that is like paying about $14 for a gallon of gasoline. At that price a Mirai costs about 3 times as much per mile as a gasoline car, and about 10 times as much as an EV.

Part of the reason is physics. Start with the same renewable electricity, and a battery car gets about 77% of it to the wheels. Turn that electricity into hydrogen, compress it, truck it, and turn it back into electricity in the car, and about 33% arrives, according to the European research group Transport & Environment. Compressing hydrogen to the 700 bar a car's tank needs takes roughly a tenth of the energy it holds.

The rest is infrastructure. California once aimed for 200 hydrogen stations by 2025. The September 2026 station list shows 42 open for retail and 13 unavailable. Shell closed all 7 of its California stations in 2024, and after a hydrogen tanker exploded in Colton in February 2026, dozens of the state's stations went offline. Toyota sold 2,737 Mirais in America in 2023 and 210 in 2025. Owners have sued over the empty pumps and over resale values, and new Mirais have been sold with discounts of more than $30,000 plus a $15,000 fuel card.

Some automakers have stepped away. Stellantis dropped its hydrogen vans in July 2025, saying the market was a niche with no prospect of paying for itself any time soon. GM wound down its fuel-cell business that October, and the fuel-cell factory it shares with Honda in Michigan stops production by the end of this year. Nikola, the hydrogen-truck start-up, went bankrupt in February 2025.

But hydrogen isn't dead. It's just not here.

Global fuel-cell sales rose about 24% in 2025, to 16,011. Almost all of that was China, mostly buses and trucks, and South Korea, where Hyundai sold its new Nexo. Hyundai is building a factory for 30,000 fuel-cell systems a year, due in 2027. Toyota is readying a third-generation fuel cell with twice the durability and about 20% more range. Daimler Truck starts a run of 100 hydrogen semis at the end of this year. The money is still flowing. It's flowing into the places where the arithmetic already works.

The most abundant element in the universe

Hydrogen's fans love to point out that it's the most common element in the universe, and they're right: roughly 74% of all ordinary matter, by mass. It seems to follow that we could never run out, and that a business built on it could keep its doors open forever.

The catch is Earth. Hydrogen is so light it escapes our atmosphere and so eager to bond with other elements that almost all of ours is locked up in water and hydrocarbons. To get it out, you spend energy, either splitting water with electricity or stripping hydrogen off natural gas. Today it's mostly the second: the International Energy Agency counts low-emission hydrogen at under 1% of supply.

That makes hydrogen an energy carrier, like electricity, rather than an energy source, like oil. It never runs out in the same way water never runs out, but every kilogram costs whatever energy went into making it. Oddly, that strengthens the gateway theory. The EV is teaching a generation of drivers that energy is something you make and store, not something you dig up. Hydrogen is one more way of storing it.

There is one tantalizing exception. In 1987 a well drilled for water near the village of Bourakébougou, in Mali, turned out to be full of almost pure hydrogen. Since 2012 it has powered the village, and its pressure has held steady. French researchers announced a large deposit in Lorraine in 2023, though geologists argue about its size. And in December 2024 the U.S. Geological Survey estimated that the planet may hold enough natural hydrogen underground that recovering even a small fraction could cover today's demand for about 200 years. Nobody is producing it commercially yet. But if natural hydrogen can be drilled cheaply, hydrogen stops being a carrier and becomes a fuel you pump out of the ground, like oil, and every number in the chart above changes.

Follow the money

A common theory is that hydrogen is being held back because an endless fuel is less profitable than a scarce one. I went looking for evidence of that, and the money points the other way.

Hydrogen is already a business of about 100 million metric tons (110 million U.S. tons) a year, and most of it belongs to the oil and gas industry: it's made from natural gas and used to refine fuel and make fertilizer. Shell and Total were founding members of the industry's Hydrogen Council in 2017, and BP, ExxonMobil, Saudi Aramco and Chevron joined later. Toyota, Hyundai, Honda, GM and BMW have spent billions on fuel cells. When companies pulled back, they gave the same reasons every time: cost, missing stations, too few customers.

Scarcity isn't what sets the price at the pump, either. California's hydrogen costs $35 a kilogram because making it, compressing it and trucking tiny volumes of it to a few dozen stations is expensive. That's the chicken and the egg: no stations means no buyers, and no buyers means no stations.

The instinct behind the theory is still right in one way. An abundant fuel can make plenty of people rich, just not by charging for the fuel. The profit sits in the electrolyzers, compressors, pumps, tanks, fuel cells and cars.

The clock

So is hydrogen the best alternative that hasn't happened yet? For the car in your driveway, not soon. The best-known ranking of where hydrogen makes sense, the energy analyst Michael Liebreich's "hydrogen ladder," puts passenger cars on the uncompetitive rung, and our own guide to every powertrain, Natural Selection, reached the same conclusion: hydrogen's best chance is the long-haul truck lane.

But the gateway theory survives, and I think it's the most important thing hydrogen has going for it. For hydrogen cars to come back, 3 things have to happen. The fuel has to get cheap, through falling electrolyzer costs or by drilling natural hydrogen. The stations have to stay open, probably starting where trucks refuel. And it needs a trusted face, the way stop buttons won over elevator riders. None of those is psychological. But when they arrive, the psychology will already be done, because millions of drivers will have spent a decade learning that electric cars are quiet, quick and completely ordinary.

Electricity was 38 years old before a third of American homes had it. The hydrogen car is 18.

If a kilogram of hydrogen cost the same as a gallon of gas tomorrow, would you pull up to the pump?