Steel production accounts for around seven per cent of humanity’s greenhouse-gas emissions. There are two reasons for this startling fact. First, steel is made using metallurgic methods that our Iron Age forebears would find familiar; second, it is part of seemingly everything, including buildings, bridges, fridges, planes, trains, and automobiles. According to some estimates, global demand for steel will nearly double by 2050. Green steel, therefore, is urgently needed if we’re to confront climate change.

To understand steel, you need to think at the level of high-school chemistry—even the chemistry you learned on the first day will suffice. Basically, steel is iron, with a little carbon added in to increase strength: tiny carbon atoms nestle between the larger iron ones, making the steel denser and more ductile. In a sense, iron isn’t so hard to find—it makes up five per cent of the earth’s crust, by weight—but metals in rock are mixed with other elements. You must get them out, in pure form, before you can build that sword or Eiffel Tower. In this respect, iron presents a particular challenge: iron atoms bind tightly with oxygen atoms, like complementary pieces in a jigsaw puzzle. Two irons and three oxygens make ferric oxide, or Fe₂O₃—a complete picture that’s hard to pull apart. Ferric oxide forms easily—so easily that, in the presence of water, naked iron will stick to oxygen in the air, creating rust.

For most of human history, therefore, the problem of iron extraction was unsolvable. Five thousand years ago, the ancient Egyptians made beads out of iron—but they got their metal from meteorites, in which it had already been split from oxygen by some unknown extraterrestrial process. Another thousand years would elapse before making usable iron became possible, through a process called reduction. Sometime around 2000 B.C.E., it was discovered, possibly by accident, that iron-heavy rock, or ore, became malleable when it was heated over charcoal fires. Today, we can explain why this happens: at high enough temperatures, iron atoms loosen their grip on oxygen atoms. The oxygen binds to the carbon in the charcoal, forming CO₂, which flies off into the air. What’s left behind is purified, or “reduced,” iron. The process of reduction allowed the Iron Age to begin.

It’s hard to say exactly when steel was first made. From time to time, it would be created when carbon diffused from the charcoal into the iron, strengthening it. But steel production was hard to control until a few hundred years ago, when the blast furnace was invented. Using bellows, steelworkers increased the temperatures of their coal fires to nearly three thousand degrees—hot enough to melt iron in large quantities. Today, blast furnaces are still the main method used to reduce steel. Current models are about a hundred feet tall, and can produce ten thousand tons of iron in a day. Instead of charcoal, they use coke, a processed form of coal. Coke and ore go in the top of the furnace, and molten iron comes out the bottom, infused with carbon; this iron can be easily processed into steel. The steel industry produces around two billion tons of it each year, in a $2.5-trillion market, while emitting more than three billion tons of CO₂ annually, most of it from blast furnaces.

Fortunately, we’ve since learned that there’s more than one way to purify iron. Instead of using carbon to remove the oxygen from ore, creating CO₂, we can use hydrogen, creating H₂O—that is, water. Many companies are working on this approach; this summer, a Swedish venture used it to make steel at a pilot plant. If the technique were widely employed, it could cut the steel industry’s emissions by ninety per cent, and our global emissions by nearly six per cent. That’s a big step toward saving the world.

The Swedish project, called hybrit—Hydrogen Breakthrough Ironmaking Technology—built its pilot plant in Luleå, in the northern part of Sweden. “hybrit” is written in a sleek, sans-serif font on the sides of the facility; a tall assemblage of gray boxes, it calls to mind a space-shuttle hangar or contemporary-art museum—the clean future, not the gritty past. The project is a collaboration between Vattenfall, Sweden’s state-owned electrical utility; L.K.A.B., its state-owned iron-ore miner; and S.S.A.B., a private steelmaking corporation. When the plant opened, last August, the Swedish Prime Minister gave a speech, and described it as representing “a historic opportunity.” I asked hybrit for a video tour, and its representatives declined, citing a need to protect proprietary technology. But, although what happens inside is something of a secret, what came out this summer was plain to see: “green” iron that, for the first time, was turned into steel and delivered to a customer.

Typically, steel results from several stages of production. Most commonly, iron ore is crushed and pelletized. Meanwhile, coal is processed into coke. Ore, coke, and limestone go into the blast furnace, creating glowing liquid iron, along with a by-product called slag and huge quantities of CO₂. The purified iron is then heated a second time, without coke, in what’s known as a “basic oxygen” furnace. During this stage, oxygen is blown over the surface of the molten iron, to encourage the production of CO and CO₂. This decreases the iron’s carbon content from about four per cent to less than one per cent. At this point, it becomes steel. “It’s a bit like a big cooking recipe,” Valentin Vogl, a grad student who’s writing his dissertation on the decarbonization of the steel industry, at Lund University, in Sweden, told me. “There’s people working in steel mills whose life is monitoring the blast furnace, and they understand the blast furnace on an intuitive level.” The finished product is cast into plates and squeezed into sheets, then rolled and shipped.

hybrit uses a different, greener strategy, which a few other ventures are also pursuing. In its system, iron-ore pellets go into the top of a so-called shaft furnace, which is roughly the same size as a blast furnace. Instead of coke, hydrogen gas goes in lower down. Inside, a process that’s known as “direct reduction” occurs. The furnace reaches about fifteen hundred degrees, which isn’t hot enough to melt the iron; as a result, the “direct-reduced iron” that comes out is still solid. It contains almost no carbon, so it goes into an electric-arc furnace—a vessel that passes bolts of electricity between internal electrodes. There, it’s melted along with a bit of coal, producing steel (and a tiny bit of CO₂). The old-school method emits oodles of carbon at every stage; the new process emits as little as possible. The hybrit pilot plant now produces about a ton of steel per hour. The next step is to build a commercial-scale demo plant, in Gällivare, also in the north, that will produce 1.3 million tons of steel a year by 2026.

Reducing steel with hydrogen has been done on a small scale in laboratories for years. Martin Pei, the chief technology officer of S.S.A.B., the steelmaking company, told me that there were no great scientific hurdles to scaling up the process. Instead, it’s mostly been a matter of optimizing the operating conditions: for instance, engineers needed to experiment with the machinery that heats the hydrogen before it’s pumped in. The real hurdle, Pei said, is the hydrogen supply. Pure hydrogen comes mostly from natural gas, typically methane—but getting hydrogen out of methane requires energy, and also creates carbon monoxide, which produces CO₂ when burned. There is a green source of hydrogen: water. It’s possible to split water into hydrogen and oxygen, by running current through it, in a process called electrolysis. But electrolysis, in turn, is green only if the electrons involved also come from renewable energy.

hybrit’s pilot plant is small, and they have no problem securing green hydrogen. But, its engineers say, creating enough green hydrogen through electrolysis to make a ton of steel requires about twenty-six hundred kilowatt-hours of electricity—enough to power an average American home for three months. hybrit also plans to use green electricity to power the preparation of the ore, the electric-arc furnace, and the steel rollers, for a total of thirty-five hundred kilowatt-hours per ton of steel. Multiply that by the nearly two billion tons of steel we currently make in a year, and you get almost seven thousand terawatt-hours of electricity. To fill that demand without producing CO₂, we’d need to nearly double the world’s annual supply of nuclear and renewable electricity. This would mean building roughly a hundred copies of humanity’s biggest existing nuclear facility, the Kashiwazaki-Kariwa Nuclear Power Plant, in Japan. So power is a problem. We’d also have to replace our existing iron and steel plants, and build massive electrolysis facilities. Even then, because mining and transportation will likely still emit CO₂, production would not be completely green.

Still, an overhaul has to start somewhere—in this case, Sweden. The country plans to be the first to achieve zero net emissions, by 2045. It hopes that other nations will follow; China, which supplies most of the world’s steel, also aims to reduce its emissions. Pei told me that he expects green steel to cost twenty to thirty per cent more than traditional steel, at least at first. But, as electrolysis processes and green-energy sources become more efficient, the cost could come down. Meanwhile, subsidies, taxes, tariffs, and other government interventions could make green steel competitive. If it becomes cheaper, it will take over.

hybrit’s green iron, after being made into steel, was rolled by S.S.A.B. That company’s first green-steel delivery went to Volvo, and this month S.S.A.B. also announced a partnership with Mercedes-Benz. Gökçe Mete, who leads the industry transition group at the Stockholm Environment Institute, told me that cars made using green steel will cost about three hundred euros extra. (Washing machines, another potential product, will cost about twenty euros more.) She thinks that many buyers will happily pay the premium. “Green steel is so prominent in Sweden,” Mete said. “You can hear young people, even hipsters, speaking about it, in cafés, having their poke bowls. Green steel is becoming a really hot topic in everyday life.” She credits the enthusiasm to a combination of media coverage, a widespread passion about the climate, and Sweden’s industrial economy: an estimated one in ten Swedes works in advanced manufacturing.

Svante Axelsson, the national coördinator for the government initiative Fossil Free Sweden, is charged with helping government and industry agree on how to transform the economy. “We have all parties with us, all unions, and also people in the streets, because they’re working in these companies,” he told me. “In one way, we have changed from a climate issue to, How can we create jobs in the future?” Axelsson said that “the state’s new role” was “to reduce risks if we want to act in an open economy.” Among other things, this involves trying to make public procurement, bank investment, permitting, education for workers, and regulation work in a harmonious way, around shared goals. “I’ve changed my language from ‘it’s two to tango’ to ‘square dance,’” he said. “Because we need so many actors to do the right steps in the right direction.”

Green steel may not be cool in America, but a similar and potentially more consequential program exists in the U.S. While the direct reduction of iron with pure hydrogen is new, direct reduction with natural gas is not. Midrex, a steelmaking firm based in North Carolina, pioneered the latter method, and built its first pilot plant in 1967. Today, the company has dozens of plants operating on that principle around the world, producing more direct-reduced iron than its competitors combined. Midrex turns natural gas into carbon monoxide and hydrogen, which together reduce iron in a shaft furnace; compared with a blast furnace using coke, this produces a third to a half less carbon dioxide.

Midrex, like hybrit, aims to go full hydrogen. In Hamburg, Germany, it is planning to build a commercial-scale demo plant for ArcelorMittal, the world’s second-largest steelmaker, by 2025; the plant will be able to use either hydrogen and carbon monoxide or pure hydrogen, and the German government will be covering half of its hundred-and-ten-million-euro cost. Switching between the two methods poses some engineering challenges. “It is not so obvious that you just change the hose and inject the hydrogen,” Lutz Bandusch, an ArcelorMittal executive who is the manager of the Hamburg site and six others in Europe, said. When you use natural gas to reduce iron, a useful shell of carbon forms on the surface of the iron pellets; this protects the pellets from rust and combustion. Without such a layer, the company will need to modify the way it melts, stores, and handles its iron. Fabrice Patisson, an engineer at the Nancy School of Mines, in France, has studied hydrogen-based direct-reduced iron in a lab, and built computer simulations of full-scale Midrex furnaces; he sees no deal breakers, just questions—about the optimal furnace shape, or the best place to add hydrogen—that need answering.

Pattison suspects that steelmakers will be harder to engineer. “The main difficulty, at least in Europe, is that they’ve relied on the blast furnace for a century, and they don’t like at all the idea of discarding it,” he told me. The resistance, he said, is “of course economic, because it means building new things. And also psychologic, for the practitioners of steelmaking. The blast furnaces are quite optimized.” Bandusch, the ArcelorMittal manager, concurred. “We have spent now two hundred years to optimize the blast furnace,” he said. “But we don’t have another two hundred years to convert the steel industry. We have to do all this in ten to twenty years. And a lot of colleagues I know are very scared.” Still, Midrex hopes to work all this out at its demo plant, and then convert the rest. There is a competitive spirit animating the effort. Referring to hybrit’s deal with Mercedes-Benz, Bandusch said, “You get the impression that they produce thousands of cars out of this steel. In fact, if you look at the volume, it’s almost nothing.” He argues that Midrex and ArcelorMittal’s scale and experience give them a huge competitive advantage when it comes to green steel.

All the same, industry-wide, the switch will be slow. “A lot of people want to make one giant leap all the way to hydrogen,” Stephen Montague, the C.E.O. of Midrex, told me. “And it’s just tough right now. The scale of green hydrogen and its cost is not quite there.” Still, Montague said, “The financial community has made it pretty clear that they don’t want to lend money to projects that aren’t more green.” Montague described the hybrit and Midrex plants as “ lighthouses,” which are “going to really show the way for everyone else.” He added, “Socially, we’ve got to be ready to pay something extra. The government can’t, frankly, afford to bear the burden of all this themselves. It’s going to have to also come down to society appreciating that green steel is a good thing.” I asked Montague about the more expensive Volvo cars that would be produced using hybrit’s clean steel. “I think there’s a lot of people who might be ready and willing to help pay for that,” he said. “I would.” Hearing an executive in heavy industry tell me that he’d pay extra for a green Volvo, I was surprisingly moved.

Hydrogen-based steelmaking isn’t the only way forward. Other groups are exploring more experimental green-steel methods. An Austrian project called SuSteel—sustainable steel—uses hydrogen plasma, which is much hotter than hydrogen gas, to reduce and melt the iron ore; while it melts, they add carbon, combining ironmaking and steelmaking into one step. Boston Metal, an M.I.T. spinout company, is based on “molten oxide electrolysis,” in which electricity is run through melted iron oxide, producing steel and oxygen. ArcelorMittal is also testing “electrowinning”—a process in which it passes current through a solution containing particles of iron oxide, so that the iron collects on one electrode while oxygen collects at the other.

We can also, of course, make other changes. We could use steel more efficiently in construction. We could recycle more of it. (Old steel can be mixed with new iron in arc furnaces.) Reducing steelmaking’s CO₂ contribution from seven per cent to one per cent or lower is possible, but it will require commitment and coördination from many participants—engineers, industrialists, bankers, legislators, and consumers. We’ll all have to make contributions, to this effort and many others. We’ll all have to learn to square-dance.