Bracing for the Hard Part
Green Steel at the Door
by gernot wagner
GERNOT WAGNER is a climate economist at Columbia Business School and faculty director of its Climate Knowledge Initiative.
Published October 30, 2026
Start with the quip the industry itself invites: there is no such thing as low-carbon steel. Carbon is not some contaminant that clever metallurgy is meant to purge from the material. It is what turns soft, rust-prone iron into steel in the first place. It is also equally important for figuring out the role of steel in the low-carbon future. Focusing on steel manufacturing itself as a source of carbon emissions is largely misplaced. The carbon we should all worry about is emitted during the production of primary iron, the input in steel.
Seventy percent of the sector’s carbon dioxide (CO2) emissions comes from just this process. The magic ingredient: coal. Chemists, please excuse the liberties taken here with chemical formulas that have existed since long before the periodic table, but the basic equation is:
Iron ore (mostly Fe2O3) + Coal* = Iron (Fe) + CO2 *(in the form of nearly pure carbon called coke)
And while simplifying otherwise complex processes often dramatizes the result, this simplification underplays it. In reality, shoveling coke into a blast furnace and using it to strip oxygen off iron ore produces over two tons of CO2 for every ton of iron metal.
The additional CO2 released when iron is further processed into steel is small by comparison. It is also the easiest part of the process to manage emissions. Iron, once turned into steel, can enter an almost perfect circular-economy loop of steel turned into scrap turned back into steel.
Yes, steel is almost fully recyclable, and indeed over 85 percent of it already is recycled (over 90 percent in Europe). Even in the United States, not usually seen as a paragon of fastidious climate policy, some 70 percent of steel is already produced by feeding scrap into electric-arc furnaces to produce steel. Thus, instead of the over two tons of CO2 per ton of steel produced via the coal-fed route from iron ore to iron to steel, scrap-fed electric-arc furnaces produce as little as a third of a ton of CO2 per ton of steel.
Imagine the circular economy in which electricity from wind turbines turns scrap steel back into steel to make more of those turbines. But that imagining also points to part of the problem: steel use keeps rising, driven by construction and infrastructure demand in fast-growing emerging economies and, increasingly, for use in wind turbines, transmission towers and much else the net-zero energy transition itself requires.
There is simply no version of the energy tran-sition that uses less steel than the fossil-fueled status quo it replaces – only one that, if the in-dustry gets this right, uses steel made without cooking the planet in the process.
Steel’s climate problem is also, awkwardly, part of the climate solution. A single onshore wind turbine contains upward of 150 tons of steel in its tower and nacelle alone. The International Energy Agency reckons that hitting global net-zero carbon emission pledges would itself add meaningfully to steel demand, even as it subtracts from demand for the fossil fuels that steel furnaces still burn. There is simply no version of the energy transition that uses less steel than the fossil-fueled status quo it replaces – only one that, if the industry gets this right, uses steel made without cooking the planet in the process.
Hard to Abate, Twice Over
Earlier this year in these pages, I made the case that cement – along with steel, one of only two manufactured materials produced and consumed at global scale – shares its chemistry, its market structure and its climate predicament with its metallic cousin. Both cement and steel are brutally price-competitive commodity businesses dominated by industrial processes that are centuries old. And both are, at long last, in the early stages of a carbon-sparing technological reinvention that could not come soon enough.
But the reinvention of steel production is not without significant market headwinds, once again mirroring the economics of cement production. Blast furnaces are colossal, multibillion-dollar machines designed to run continuously for 15 to 20 years between timeouts to reline them with heat-resistant materials. Nobody switches off a blast furnace mid-campaign in order to add a green equipment upgrade. Thus the economic moment to make a fundamental technology choice comes only every couple of decades when a furnace is due for a rebuild, anyway. Cement plants face something similar with their kilns, but the sheer scale of steel’s capital assets – and a global industry skewed toward furnaces recently built or rebuilt across fast-industrializing Asia – makes for an even longer wait between opportunities to change course.
That inertia shows up starkly in what is actually getting built right now. For every steel plant adapted to a low-carbon world that is currently in the planning or construction pipeline worldwide, more than two conventional coal-fired blast furnaces are being planned or built alongside it. And every one of those new coal-fired furnaces will very likely still be running in 2050.
Scrap Heap of History
Before turning to the industry’s moonshots, then, it is worth pausing to contemplate an unglamorous decarbonization lever that already works at scale today: recycling. Steel is that rare manufactured material that can be melted down and reused repeatedly without meaningful loss of quality. Moreover, recycling via an electric-arc furnace uses only a fraction of the energy – and none of the coal – required to make primary iron from ore. Combined with the modest inroads made by a competitive process, natural-gas-fired direct reduction, these lower-carbon pathways already account for nearly a third of global steel production.
The catch, of course, is the remaining two-thirds, still produced the old-fashioned way. Recycling is limited by the amount of scrap available, and scrap availability tracks a country’s stock of steel already in use: plentiful in the mature economies of North America and Europe that have decades of old cars, appliances and buildings to draw on, and comparatively scarce in China, India and the rest of Asia, where they are still in the business of accumulating steel machines and structures rather than recycling them.

That geographic mismatch matters because it is precisely the fast-growing economies of Asia that will determine whether carbon emissions from global steel production rise or fall over the coming decades. Scrap-fed electric-arc furnaces are close to a free lunch (as far as carbon emissions go) where scrap is abundant. It is no lunch at all, where it isn’t.
Light-Green Steel
There is another underappreciated lever, one cheap enough that it ought to have been pulled already. Mining metallurgical coal – the specific grade of coal modified to feed into blast furnaces – releases methane directly into the atmosphere. Methane is a far more potent greenhouse gas than CO2, trapping around 80 times more heat over a 20-year horizon. Thus, coalbed methane alone adds the equivalent of a billion tons of CO2 to blast-furnace steel’s already outsized climate footprint each year, around a quarter on top of the direct carbon emissions from the furnaces themselves.
The fix here does not require reinventing iron- or steelmaking. Metallurgical coal is mined by a relatively small number of companies concentrated in a handful of countries, Australia and China chief among them. China alone accounts for over half of metallurgical coal and steel production – which means the methane could plausibly be mitigated through a comparatively small number of targeted partnerships at a cost of under 1 percent of finished steel’s market price. Cleaning up these methane emissions linked to mining would not by itself produce green steel. But it is the metaphoric $20 bill lying on the sidewalk that skeptical economists insist cannot really be there, and yet it is.
Green Hydrogen for Green Steel
The more consequential fight, the one that will determine whether steel can ever be made genuinely clean (net-zero carbon), is over how to replace coal in the ironmaking phase. Two rival approaches have emerged, and they diverge less over the destination than over how much of the existing industrial furniture gets kept along the way.
The first, further along and more heavily capitalized, consists of swapping hydrogen for natural gas as the source of heat in the furnace. Instead of coke and a blast furnace, a continuously operating “vertical shaft” furnace is fed iron ore pellets and a stream of hydrogen. The hydrogen strips the oxygen from the ore to leave behind lumps of “sponge iron,” with water vapor (H2O) rather than CO2 as the byproduct.
That sponge iron then goes into an electric-arc furnace, the very same equipment already used to melt scrap steel, to be turned into finished steel.
Done with genuinely green hydrogen – that is, hydrogen produced by electrolyzing water with electricity from a zero-carbon source – the process can cut emissions by over 90 percent compared to the coal-fired baseline. The furnace itself is a relatively modest engineering departure from a century of ironmaking practice. The true departure is upstream, in making available the enormous quantities of cheap, clean hydrogen the process demands.
Green steel costs meaningfully more than the coal-fired variety at the factory gate, but the gap barely registers by the time it is used in making expensive, high-precision products.
Producing a ton of iron this way takes on the order of 50 kilograms (110 pounds) of hydrogen, which in turn takes something like 3,000 kilowatt-hours of electricity to electrolyze at typical efficiencies – and that is before a single kilowatt-hour has gone into melting the iron in the electric-arc furnace or rolling the finished steel into useful shapes.
Thus scaling this process across even a modest fraction of global steel output would require electrolyzer capacity that dwarfs anything installed anywhere today. That is why the world’s first projects of this kind are sited where clean electricity is cheap and abundant, and demand for that power so far has been modest. Stegra (named H2 Green Steel at its founding in 2020) is farthest along the construction process at its site in Boden in northern Sweden. The key locational advantage: access to Scandinavian hydropower and wind.
The first phase, still under construction, pairs a 700-megawatt array of electrolyzers with a direct-reduction shaft furnace to make iron and a pair of electric-arc furnaces to transform the iron into steel, targeting 2.5 million tons of steel a year. (The potential second phase would double that capacity.)
The project has had its share of financing drama – a delayed disbursement of European Innovation Fund money and, this spring, a scramble to close a €1.4 billion financing round led by Sweden’s Wallenberg family investment vehicles.
But as of this writing, it remains on track to begin production within months. If it works at anything close to the promised scale, it will be the proof of concept that the entire hydrogen-steel thesis has been waiting for.
Stegra’s success (if it does prove a success) will be as much about economics and psychology as engineering. Mercedes-Benz, Scania and other companies locked in long-term purchase commitments years before the Boden plant produced a single coil. That was less because green steel penciled out as fiscally viable on a spreadsheet than because of the fear of missing out on something truly revolutionary for which their customers might be willing to pay extra.
The economics work out in favor of this thesis. Green steel costs meaningfully more than the coal-fired variety at the factory gate, but the gap barely registers by the time it is used in making expensive, high-precision products. Even if the first million tons of steel from the Boden plant cost twice that of conventionally made, high-quality steel fit for a Mercedes-Maybach, the extra few hundred dollars cost could be written off as a worthy marketing expense to sell a $100,000-plus green vehicle. And the claim of a truly green vehicle would be real.

Much the same logic applies to a refrigerator, a washing machine or an induction cooktop: the cost of the green steel would be a sliver of the retail price, which means the willingness of a handful of premium brands to pay for bragging rights can make the use of green steel financially viable.
There is also a good reason that Stegra is located in an economy regulated under the European Union’s Emissions Trading System, which establishes a carbon-emissions surcharge that adds some €75 per ton of CO2 to its competitor’s “dirty” products. It is no wonder, then, that Stegra is among the few steelmakers that would actively lobby for strengthening the EU’s climate policies.
Electrochemical Holy Grail
A second, even more radical green-steel technology bet skips hydrogen in favor of electrochemistry. Iron ore electrolysis works by running an electric current directly through molten or dissolved iron ore, plating out pure metallic iron much as electroplating deposits copper or chrome. No coal, no hydrogen, no combustion of any kind – just ore, electricity and, potentially, up to a 97 percent cut in emissions since the process sidesteps not only the use of carbon in chemical reduction but also much of the massive energy input involved in first extracting hydrogen gas from water and then using it to make iron.
Boston Metal, an MIT spinoff pursuing a high-temperature version of this technology called molten oxide electrolysis, is aiming to start up its first demonstration-scale plant this year, having raised well over half a billion dollars along the way despite an industrial accident at its Brazilian pilot facility in January 2026. Rival startup Electra is pursuing a lower-temperature “electrowinning” approach that extracts iron from an iron salt solution, skipping the melting step entirely. The company is building its own first commercial facility.
Both Boston Metal and Electra are starting at scales in the hundreds of tons a year, making Stegra’s million-ton ambitions look positively enormous by comparison. But the radical nature of using electrochemistry to produce iron is worth spending the extra time and money to get right. The resulting processes might more accurately be compared to battery manufacture than traditional iron or steel production: when the sun shines, Electra’s electrowinning process turns iron ore into iron; when it doesn’t, it could conceivably even release stored electricity by turning iron back into iron ore.
But long before electrochemical iron producers somehow enter the electricity generation business, they may find themselves selling byproducts. Both iron-electrolysis ventures have discovered that raw iron ore is not just iron oxide waiting to be stripped of its oxygen; the ore comes bundled with valuable impurities. Boston Metal is now chasing a side business extracting niobium, tantalum and tin from the same electrolytic cells that produce its iron, with vanadium, nickel and chromium recovery in its sights. This co-production strategy could thereby subsidize the core technology’s path to commercial scale in iron production, exactly as Brimstone’s alumina is meant to do for cement.
Think of this as the steel-sector equivalent of mopping up one’s own pollution after the fact. The approach is less transformative than a wholesale process change, but potentially useful as a bridge to radical decarbonization.
Blast and Capture Carbon
There is a third pathway to higher efficiency worth noting, one favored by incumbent iron producers unwilling to write off their massive investments in conventional blast furnaces. Here, the idea is to keep the furnace, inject some hydrogen along with the coke to economize a bit on carbon emissions, and bolt carbon capture technology onto the exhaust stack to mop up what remains. Nippon Steel, the world’s third largest steelmaker, is at the forefront of this approach, having pursued a version of it for years. Several European integrated steel producers are exploring similar hybrids.
Think of this as the steel-sector equivalent of mopping up one’s own pollution after the fact. The approach is less transformative than a wholesale process change, but potentially useful as a bridge to radical decarbonization that could smooth the phaseout of long-lived blast furnaces.
There is no getting away from the reality that adding carbon capture to conventional steelmaking will always cost more than processes that allow the CO2 to escape into the atmosphere. Unless steelmakers are paying a price for their own pollution, or a generous state begins to subsidize carbon capture en masse, or a less generous one forces steel producers to cut emissions by fiat, no steelmaker will voluntarily invest in cutting CO2 emissions.
Washington’s Cold Feet
In 2024, the U.S. Department of Energy selected Cleveland-Cliffs’s Middletown Works, an aging blast-furnace complex in Ohio, to receive up to a $500 million grant to swap its coal-fired ironmaking for a hydrogen-ready, direct-reduction plant paired with electric melting furnaces. But in mid-2025, Cleveland-Cliffs walked away from the free money.

The company’s stated reason doubled as an unusually candid admission about the entire hydrogen-steel thesis: there simply was not going to be enough affordable hydrogen available on which to run the plant – not on any timeline that made financial sense.
The Trump administration’s cancellation of several regional clean-hydrogen producing hubs, including leaving the Mid-Atlantic hub that would have supplied Middletown in bureaucratic limbo, only confirmed the point. Sweden’s SSAB had reached a similar conclusion even earlier, quietly abandoning its own $500 million DOE award for a green-steel project in Mississippi after running into the same hydrogen-supply wall – and before the Trump administration backed away from green technology.
The retreats on the Middletown and Mississippi projects are a pointed illustration of the chicken-and-egg problem that bedevils industrial decarbonization strategies the world over. Nobody builds a hydrogen-ready furnace without confidence in cheap reliable supply of hydrogen, and nobody builds a source of cheap hydrogen without confidence in furnace-scale demand. The Biden-era clean-hydrogen hubs were called hubs for that reason: it takes a cluster of industries to make sense of the initial investment push, and it typically takes government to coordinate such ventures.
Governments are, collectively, the world’s largest buyers of steel by way of the bridges, rail lines, transmission towers and public buildings they commission. Thus even failing a broader governmental strategy to help clean industrial hubs to get off the ground, the willingness to buy the green end-product would make a difference. And a growing number of governments have started using that purchasing power deliberately via “buy clean” standards that require a minimum share of green steel (and cement) in government-funded infrastructure.
A Tale of Two Tariffs
Few sectors are deemed to be of greater national strategic importance than iron and steelmaking. That is evident in Europe, where seemingly each country has its own, often oversized national steelmaker that no government is willing to abandon to the cold realities of market economics. It is evident in the United States, where tariffs on imported steel have now reached a punishing 50 percent, recently increased further by applying them to the full value of any imported product rather than just its metal content.
The EU’s Carbon Border Adjustment Mechanism took effect in January this year. It requires importers of steel, cement, aluminum, fertilizer and a handful of other carbon-intensive goods to purchase certificates offsetting the emissions embedded in imports.
The European Union, too, has introduced steel tariffs, but in a rather more rational way than the United States. The EU’s Carbon Border Adjustment Mechanism – CBAM to those in the know – took effect on the first of January this year. It requires importers of steel, cement, aluminum, fertilizer and a handful of other carbon-intensive goods to purchase certificates offsetting the emissions embedded in imports. The logic is simple: the EU’s carbon-emissions trading system establishes a price for domestic producers, while CBAM ensures that importers are not given any advantage. Without it, European companies would have incentives to move their operations abroad and sell cheaper products back into the EU, increasing global CO2 emissions in the process.
CBAM thus levels the playing field and then some, helping to propagate green policies beyond the EU. South Korean steelmakers may only export 5 percent of their wares to the EU, but those 5 percent would now be subjected to tariffs paid to Brussels. Why not keep that money in Korea and invest it in cleaner production at home?
Brussels would be more than fine with that. Indeed, it is this Brussels Effect the EU was after in the first place. It also rewards exactly the kind of investments Stegra, Boston Metal, Electra and their rivals are racing to make, though CBAM alone will not be enough.
European steelmakers have all run into similar problems that have bedeviled much of Europe’s broader clean industrial strategy: how to produce and procure clean power at a price that’s right. The key to Europe’s – or anyone’s – green industrial future is cheap, carbon-free electricity. That means power sector pricing reform as much as the physical buildout of the grid, and it necessitates coordination well beyond even green industrial hubs.
What Would It Take
Put the pieces together and the shape of a serious steel decarbonization policy is not all that mysterious. It takes cheap, genuinely clean electricity and, where hydrogen is the chosen route, cheap genuinely clean hydrogen. It takes policy change in the form of a substantial price/tax on carbon emissions like that set under the EU’s emissions trading system to make clean steel commercially rational rather than a philanthropic gesture. It would also take public procurement rules and long-term commitment-driven demand to carry early volumes. And it takes patient, risk-tolerant capital that bridges the years between a pilot plant proving a technology works and a bank financing the tenth commercial-scale furnace built on the strength of the first nine.

None of the individual pieces of this puzzle are weighty, unsolved engineering problems. Hydrogen-based direct reduction works: Stegra is building it at gigawatt scale. Iron electrolysis works: Boston Metal and Electra are busy proving it can run continuously rather than merely in a laboratory beaker. Scrap-based recycling has worked for decades. Methane emissions from coal mining can be largely contained for a trivial addition to the market price of steel.
* * *
Steel has been at the center of great technological disruptions before. The Bessemer converter, patented in 1856, took a material (steel) that was once so costly to make that it was displayed as jewelry and transformed it into the cheap backbone of railroads, skyscrapers and modern cities in a generation. The technologies now competing to decide what comes after coal and coke are no less consequential in their own way, even if the size of the pioneers is still measured in the low millions of tons of steel rather than the nearly two billion tons the world currently consumes each year.
What remains unsolved is not the chemistry, metallurgy or engineering, but the matter of who pays for bridging the gap between what clean steel costs today, and what it will cost at scale. Getting there may be as simple as not changing the rules of the game halfway between first investment decision and construction, along with recognizing the very real trade-offs between making a quick buck now and scaling the kind of technologies that can decarbonize one of the most fundamental building blocks of modern society.