Hydrogen = Heavy Industries Hype

The world produced more hydrogen last year than at any point in its history, and it burned almost none of it as a fuel. That is not a failure of hydrogen. It is the answer to a question the market spent the better part of a decade asking the wrong way. The question was never whether hydrogen would replace fossil fuels across the economy. It was always narrower, and more useful: which fuel would hydrogen replace, and where. Three years on from the peak of the hype cycle, we finally have enough data to answer it properly, and the answer sharpens rather than dilutes our conviction on where the capital should go.

Hydrogen is a large, essential and growing industrial commodity. It is not, and was never going to become, a fuel that displaces anything on the energy transition side of the ledger. That distinction — feedstock versus fuel — is the one the last three years of project cancellations, cost blowouts and hard engineering constraints have now settled. What is left standing is a narrower and, in our view, considerably more investable thesis: hydrogen as the industrial fuel of the future for heavy industry, long-haul shipping and aviation, while electrons take over everything that can plausibly run on a wire.

The Market That Wasn't

Global hydrogen demand reached almost 100 million tonnes in 2024 and is expected to have surpassed that milestone in 2025. The growth is being driven by demand for industrial products that use hydrogen as a feedstock, not by the successful implementation of energy and climate policy. Demand remains concentrated almost entirely in the sectors that have used hydrogen for the better part of a century: refining, ammonia, methanol and fossil-based direct reduced iron. Demand for the new applications the hype cycle was built on — biofuels upgrading, new industrial uses, mobility, power generation, synthetic fuels — is growing, but from a base so low it represents under one per cent of total demand.

The composition is telling. More than seventy million tonnes of that base is grey hydrogen, made from unabated natural gas. The clean hydrogen the last decade of policy and investment was built to deliver is, for now, a rounding error against the industrial base it was supposed to transform. The refining sector alone illustrates the scale involved: the global petroleum refinery hydrogen market was valued at USD 148.7 billion in 2025, with refining hydrogen demand reaching 43 million tonnes in 2023 and continuing to grow through 2025, projected to reach USD 257.9 billion by 2035. This is not a market in decline waiting to be disrupted. It is a mature, capital-intensive industrial market that happens to have very little to do with the energy transition.

Concentration risk sits alongside scale. Ammonia and methanol alone represent roughly half of global hydrogen-based consumption, and the Middle East plays a critical role in global markets for both. The conflict in the Middle East has already disrupted not only oil and gas flows but hydrogen-based product supply chains, particularly fertilisers and chemicals such as ammonia, urea and methanol — a reminder that hydrogen's industrial base carries the same geopolitical exposure as the rest of the energy complex, notwithstanding its clean-fuel branding.

Inside the Refinery: What Hydrogen Actually Does

It is worth being precise about what refinery hydrogen is for, because the popular framing — that hydrogen reduces refinery emissions — is not quite right. Refiners use hydrogen for two core processes. Hydrotreating, also called hydrodesulfurisation, combines heated oil feedstock with high-purity hydrogen over a catalyst bed to strip sulfur and nitrogen out of diesel, gasoline and jet fuel. This matters because sulfur poisons the catalysts used downstream in catalytic reforming, and because fuel-quality regulations require ever-lower sulfur limits in the finished product. Hydrocracking, the second core process, cracks heavier hydrocarbons in the presence of hydrogen into lower molecular weight, higher value products, letting a refiner convert more of each barrel into diesel and jet rather than low-value residue.

Both processes are fundamentally about fuel quality and yield economics — cutting sulfur oxide emissions once the fuel is burned in a vehicle, and improving the commercial value of a barrel of crude — rather than cutting the refinery's own carbon footprint. The refinery's own emissions problem sits mostly upstream, in how the hydrogen itself is produced, since steam methane reforming of natural gas emits CO₂ at the point of hydrogen manufacture, not at the point of use.

There is a genuine decarbonisation angle here, but it is a retrofit at the margin rather than a sector-wide transformation. Based on projects that are operational, under construction or past final investment decision, more than 2 million tonnes per annum of low-emissions hydrogen is expected to be consumed in refineries and industrial facilities by 2030. Set against 43 million tonnes of refining demand in 2023 alone, that is a meaningful start, not a transformation. Indian Oil Corporation's own target for green hydrogen facilities across all its refineries is set for 2047. That is the timeline the industry itself considers realistic: decades, not years.

A useful historical detail often lost in the current debate is that refinery hydrogen use is not a recent innovation bolted on for emissions purposes. Cheap hydrogen first became widely available to refiners in the early 1950s as a by-product of catalytic reforming. Modern refineries still run largely on this same recycled-hydrogen logic. The current decarbonisation question is not whether refineries will start using hydrogen — they always have — but whether the marginal tonne they buy in over the next decade is grey, blue or green.

The Reckoning: What 2025 and 2026 Confirmed

The cost curve for green hydrogen moved in the wrong direction. Electrolyser system costs rose by a median of 57 per cent between 2022 and 2024, according to BloombergNEF's Electrolyzer Price Survey — the opposite of the steady cost-decline curve that underpinned most 2020 to 2022 forecasts. Current green hydrogen production sits at USD 3.00 to 6.00 per kilogram, two to five times the cost of grey hydrogen, with electrolyser capital costs of USD 500 to 1,800 per kilowatt-electric and renewable power accounting for 60 to 70 per cent of the final cost.

The project pipeline has shrunk to match. The IEA's Global Hydrogen Review 2025 tracks more than 200 committed low-emissions hydrogen investments globally, yet potential 2030 production from announced projects fell to 37 million tonnes per year, down from 49 million tonnes estimated only a year earlier — a downgrade of roughly a quarter in twelve months. The cancellations are global: in Europe, Repsol froze 350 megawatts of Spanish projects, Shell halted its Aukra blue hydrogen project, and a planned Germany-Norway pipeline was shelved. In Australia, bp exited the Asian Renewable Energy Hub, Trafigura exited its Port Pirie project, and Fortescue wrote off projects, all after feasibility studies revealed prohibitive renewable electricity costs.

The pattern that separates the survivors from the write-offs is the one that matters most in energy infrastructure: bankable, long-term offtake — not subsidy or sentiment. The USD 8.4 billion NEOM Green Hydrogen project in Saudi Arabia was 80 per cent complete as of February 2026, and it survived where comparable projects failed because it is de-risked by a 30-year exclusive offtake agreement with Air Products. The largest single green hydrogen cancellation in history, by contrast, failed to secure sufficient industrial offtakers covering 75 per cent of production. This is the same discipline we apply to every behind-the-meter power investment we underwrite: the project that has a contracted buyer gets built, and the project that has only a policy tailwind does not.

Electrolyser manufacturing is now running well ahead of the projects that would actually use it. Global annual electrolyser manufacturing capacity is estimated at 61 gigawatts, with a further 16 gigawatts under construction — a substantial oversupply against the volume of projects that have reached final investment decision. The US Department of Energy's Hydrogen Shot Initiative, which targets a green hydrogen cost of USD 1.00 per kilogram by 2031, is itself an implicit admission of how far current costs sit from viability. None of this changes the physics of what hydrogen and its derivatives are good for. It simply confirms that the timeline for green hydrogen to reach cost parity with grey hydrogen at scale is measured in the 2030s, not the 2020s, and that capital deployed on any other assumption is capital deployed on hope.

The Great Sort: Electrons for the Many, Hydrogen for the Few

Strip away the hype and what remains is a sensible division of labour between two energy carriers, decided by physics and unit economics rather than ideology. Electrons win comprehensively wherever a battery, a wire or a heat pump can do the job: commuter and passenger vehicles, light industry, buildings and general-purpose transport. Hydrogen and its derivatives win where electrons physically cannot do the job at acceptable cost — and those are precisely the sectors where genuine hard-to-abate demand sits.

The commuter and light industrial side of this split is already happening at scale. Global electric car sales exceeded 20 million units in 2025 — one in every four new cars sold worldwide — and the IEA's Global EV Outlook 2026 projects 23 million units this year, taking the global share to 28 per cent. China alone saw electric cars capture more than half of all annual car sales for the first time in 2025, and electric heavy-freight truck sales tripled to over 200,000 units. None of that growth required a hydrogen refuelling network or a cost breakthrough still years away. It required batteries that kept getting cheaper, on a curve hydrogen has conspicuously failed to replicate.

In heavy industry, hydrogen-based direct reduction is a real decarbonisation pathway for steel, and the Stegra industrial-scale plant due to start in 2026 will be a key test. But even hydrogen advocates now frame this as a smaller residual problem once scrap volumes rise and electric arc furnace capacity grows. Some of the most credible near-term demand is not green at all: steel producers with cheap by-product hydrogen, such as Shanxi Jinnan Iron and Steel Group in China, are already running large fleets of hydrogen fuel cell heavy trucks on hydrogen priced at or below USD 3.60 per kilogram without subsidy, precisely because the hydrogen was a by-product rather than a purpose-built green molecule.

In long-haul shipping, the 2026 engineering consensus has settled on green ammonia — a hydrogen derivative — rather than pure hydrogen for any route over 2,000 nautical miles, because the volumetric energy-density penalty of hydrogen itself is too high for long voyages. Maritime shipping underpins roughly ninety per cent of world trade and is responsible for close to three per cent of global carbon emissions, which is why iron ore majors including Fortescue and Rio Tinto have been converting or trialling ammonia-fuelled and dual-fuel vessels ahead of the IMO's 2050 decarbonisation targets.

In aviation, direct hydrogen combustion or fuel-cell aircraft remain a genuinely future technology. The nearer-term and commercially relevant role is hydrogen as feedstock for e-SAF, or e-kerosene, where renewable hydrogen combines with captured CO₂ to produce a drop-in fuel usable in today's aircraft without any change to the airframe or engine. EU policy is forcing this pathway onto a fixed schedule, with ReFuelEU Aviation requiring minimum sustainable aviation fuel shares at EU airports rising from 2 per cent in 2025 to 70 per cent by 2050.