For a few years, green hydrogen was sold as a silver bullet — a clean, storable fuel that could decarbonize almost everything, from steel mills to family cars. Then 2025 arrived with a different message. BP walked away from a gigawatt-scale hydrogen project in Oman, Air Products wrote down as much as $3.1 billion on cancelled U.S. plants, and the International Energy Agency reported that the global project pipeline had shrunk by about 25% in a single year [source: IEA, 2025] [source: Air Products, 2025]. This is not a story about batteries or about pulling carbon out of the air. It is a narrower, more grounded question: after all the hype, where does hydrogen made from water and renewable electricity actually make sense — and where was the promise oversold?
What "green" hydrogen even means
Hydrogen is not an energy source you dig up; it is an energy carrier you have to make. Almost all of the roughly 100 million tonnes (Mt) the world uses each year is "grey" hydrogen, produced by stripping it from natural gas or coal in a process that releases large amounts of carbon dioxide [source: IEA, 2025]. "Blue" hydrogen is the same fossil process fitted with carbon capture to bury some of those emissions. "Green" hydrogen is the newcomer: it splits water into hydrogen and oxygen in a device called an electrolyzer, powered by renewable electricity, so that — in principle — the only by-product is oxygen.
The appeal is obvious. If the electricity is clean, green hydrogen offers a way to decarbonize industries that are otherwise very hard to electrify directly. The catch, which the last two years have made unavoidable, is that making hydrogen this way is still far more expensive than making it from fossil fuels, and building the industry has proved slower and harder than its boosters promised.
The hype years, 2020–2023
Between 2020 and 2023, green hydrogen became a centerpiece of national climate strategy. The European Union wrote it into its plans for hard-to-decarbonize industry and set ambitious production and import targets. The United States, through the 2022 Inflation Reduction Act, created a clean-hydrogen production tax credit — Section 45V — worth up to $3.00 per kilogram, alongside billions of dollars for regional "hydrogen hubs" [source: U.S. Treasury, 2025]. Gulf states and Australia announced export mega-projects meant to ship hydrogen and its derivatives to Europe and Asia.
On paper, the pipeline was enormous. Developers announced enough projects that, if every one were built, low-emissions hydrogen could have reached tens of millions of tonnes a year by 2030. But announced capacity and financed capacity are two very different things — and the gap between them is where the reality check begins.
Announced versus built: the correction of 2024–2026
Start with the layer that matters most: what was announced versus what was actually financed and running. Based on all announced projects, the IEA estimated that low-emissions hydrogen could reach 37 Mt per year by 2030 — but that figure was itself a downgrade from the 49 Mt the agency had projected just a year earlier, and roughly half of announced projects now face delayed start dates [source: IEA, 2025]. The project pipeline shrank by about 25% in a year [source: S&P Global, 2025].
Set those announcements against measured reality. Installed water-electrolysis capacity reached only about 2 gigawatts (GW) worldwide by the end of 2024 [source: IEA, 2025]. Low-emissions hydrogen still makes up less than 1% of total hydrogen use, and even in 2025 its production was on track for only about 1 Mt [source: IEA, 2025]. The capacity that actually reached a final investment decision (FID) — the moment money is genuinely committed — was, in 2024, no higher than in 2023 [source: IEA, 2025]. Announced pipelines are a measure of ambition; FID and installed capacity are a measure of conviction, and the two have diverged sharply.
The cancellations put faces on the numbers. In December 2025, BP withdrew from the 1.5 GW Duqm green-hydrogen project in Oman, which had been designed to produce about 150,000 tonnes of hydrogen a year; the company said a deep review had concluded that large-scale green hydrogen did not meet its cost-of-capital threshold under current conditions [source: Hydrogen Insight, 2025]. Earlier, in February 2025, Air Products — one of the industrial-gas giants that had bet heavily on hydrogen — exited three U.S. projects and took a write-down of up to $3.1 billion, explicitly blaming, for its cancelled New York green-hydrogen plant, both a regulatory change that made its hydropower ineligible for the 45V credit and slower-than-expected demand for hydrogen in transport [source: Air Products, 2025].
One caution about reading these headlines: a wave of cancellations is not the same as collapse. The same IEA review that documented the shrinking pipeline still expects low-emissions hydrogen to grow strongly from a low base, and capital spending actually rose — investment in low-emissions hydrogen reached $4.3 billion in 2024, an 80% jump, and recent FIDs could push it toward roughly $8 billion in 2025 [source: IEA, 2025]. The honest reading is neither "hydrogen is dead" nor "hydrogen won"; it is that a speculative pipeline is being filtered down to the projects that can actually pencil out.
The cost reality
Why do so many projects fail to pencil out? Cost. This is the difference between the felt promise of steep future declines and the measured price today. In Europe, the levelized cost of green hydrogen sits around $4.5–6 per kilogram, against roughly $1.5–2.5 for grey hydrogen from fossil fuels — and green hydrogen still runs two to three times the cost of blue [source: IRENA, 2024]. A subsidy like the U.S. $3/kg credit can close much of that gap, which is precisely why projects lean on it so heavily.
The optimistic case has always rested on costs falling fast as electrolyzers scale up and renewable power gets cheaper. Some of that is real — but forecasts have grown more sober. BloombergNEF, which once projected dramatic declines, revised its outlook upward, with green hydrogen expected to stay above roughly $3.74/kg even through 2050 in a recent analysis [source: BloombergNEF, 2025]. Two big components drive the price: the electrolyzer itself, which accounts for perhaps 40–60% of the levelized cost, and the electricity to run it, another 30–50% [source: BloombergNEF, 2025]. Cheap, abundant, genuinely additional renewable power is the whole game, and it is not yet cheap or abundant enough in most places. The projected cost curve and the measured cost today are not the same number, and 2025 was the year the market stopped confusing them.
Where it makes sense — and where it probably doesn't
Almost lost in the boom-and-bust noise is a genuine consensus among analysts about where green hydrogen actually belongs. Today's hydrogen demand is overwhelmingly industrial — oil refining, ammonia for fertilizer, methanol, and steelmaking — and nearly all of it is made from fossil fuels, so there is a large, real emissions-cutting prize in cleaning up those existing uses [source: IEA, 2025]. These are the "hard-to-abate" sectors: places where hydrogen is a feedstock or where direct electrification is genuinely difficult. Green steel, clean ammonia and fertilizer, refining, and hydrogen-derived shipping and aviation fuels are where most independent analysis says the molecule earns its keep [source: IEA, 2019].
The flip side is where hydrogen was oversold. For heating homes and for passenger cars, multiple studies conclude hydrogen is an inefficient detour compared with heat pumps and battery-electric vehicles, which use far less energy to deliver the same service [source: Nature Energy, 2022]. That distinction matters, because much of the early hype blurred it — pitching hydrogen as a universal fuel for everything rather than a targeted tool for the sectors that have no better option.
Even in its right sectors, green hydrogen faces a chicken-and-egg problem. Producers will not commit to giant electrolyzer plants without buyers signed up for the output; buyers will not switch their factories to hydrogen without a reliable, affordable supply and the pipelines and storage to move it. Weak, uncertain demand has become the single most cited reason projects stall — and, unlike a single company's balance sheet, that missing infrastructure and offtake is a collective-action problem no one developer can solve alone.
The policy reset
The final layer is policy, and here it is easy to confuse correlation with causation. Subsidies clearly correlate with where projects get built, but generous headline credits did not, by themselves, cause a durable industry to appear — the rules attached to them, and the demand beneath them, decide what actually gets financed.
The United States illustrates the whiplash. Treasury and the IRS finalized the 45V rules in January 2025, offering up to $3/kg for hydrogen whose lifecycle emissions stay at or below 4 kg of CO₂-equivalent per kilogram, governed by three pillars — new ("additional") clean power, deliverability, and time-matching of that power to production [source: U.S. Treasury, 2025]. Barely months later, the 2025 budget law known as the One Big Beautiful Bill Act moved up the credit's expiration, ending it for projects that begin construction after the end of 2027 rather than in the 2030s [source: U.S. Congress, 2025]. A subsidy finalized and then curtailed in the same year is its own kind of reality check.
The European Union has its own rulebook. Its delegated regulation on renewable fuels of non-biological origin (RFNBO), in force since 2024, defines when electrolytic hydrogen counts as truly renewable through criteria for additionality and for time- and location-matching of the electricity used [source: European Commission, 2023]. Those rules were meant to guarantee that hydrogen does not simply divert existing clean power — but as Europe looks likely to miss its targets, the Commission has moved to relax and delay several requirements to keep projects viable [source: European Commission, 2025]. That, too, is telling: when a regulator loosens the rules to keep an industry alive, it is conceding how fragile the economics still are.
Overhyped, or just early?
So which is it — was green hydrogen overhyped, or is it simply early? The evidence supports a careful "both." The hype was real and, for some uses, misplaced: hydrogen was never going to heat most homes or power most cars, and treating announced pipelines as if they were built capacity set the stage for disappointment. But the underlying case for green hydrogen in heavy industry has not gone away; it is being tested, slowly and expensively, in exactly the sectors where alternatives are scarce.
The reality check of 2025–2026, then, is not a verdict that green hydrogen has failed. It is the market and regulators doing what hype could not: separating the projects that make physical and economic sense from those that never did. A smaller, more focused industry aimed at steel, ammonia, refining, and clean fuels may prove more durable than the sprawling everything-everywhere vision it is replacing.
Watch three things from here. First, how much of the shrinking pipeline actually reaches final investment decision and operation — FID and installed gigawatts, not press-release ambitions. Second, whether costs resume falling as electrolyzer manufacturing scales, or settle at a stubborn premium over fossil hydrogen. Third, whether real demand materializes in the hard-to-abate sectors, backed by the infrastructure and offtake contracts that turn a good idea into a financed plant. The silver-bullet story is over. The more interesting question — how big a role a right-sized hydrogen industry will play — is only now being answered.