The industry’s own numbers show why the debate over 100, 170, or 300 million tonnes by 2050 matters less than it sounds. Roughly 700 megawatts of electrolysers are actually running today, producing around 110,000 tonnes of green hydrogen a year, a fraction of even the most conservative target on the table.
Global electrolyser capacity in operation today stands at around 700 megawatts, producing an estimated 110,000 tonnes of green hydrogen annually, according to analysis published by BloombergNEF’s Michael Liebreich. Whether the eventual 2050 market turns out to be 100 million tonnes, 170 million, or 300 million, current output represents somewhere between 0.06% and 0.1% of any of those figures. That gap, more than the specific number analysts settle on, is the more useful fact for judging where the hydrogen sector actually stands.
The debate over the size of the future market is real, and forecasts have moved substantially. The IEA’s October 2021 Net Zero by 2050 Roadmap called for 212 million tonnes of hydrogen by 2030, 150 million of it low-carbon; the agency’s own subsequent updates cut that to 150 million tonnes total, with only 70 million low-carbon. The widely cited political target of 90 million tonnes of clean hydrogen by 2030, referenced at the 2022 Hydrogen Energy Ministerial, traces back loosely to the same IEA modelling exercise rather than to any independent assessment of what was achievable.
What has driven the repeated downward revisions is more instructive than the destination number. BloombergNEF’s own cost outlook for green hydrogen has moved in the wrong direction over the past few years, not the right one. In 2021, BNEF forecast that renewable hydrogen would fall below $1 per kilogram in most modelled markets by 2050 and undercut natural gas on an energy-equivalent basis in 15 of 28 markets studied. Its most recent projection instead puts 2050 costs between $1.60 and $5.09 per kilogram, with cost parity against grey hydrogen achievable only in China, India, and Texas. BNEF attributes the reversal to rising electrolyser manufacturing costs and to the limited economies of scale that follow from stagnant demand and a wave of cancelled large-scale projects, a description that cuts directly against the “cost reduction through scale and industrial learning” argument that continues to justify new investment in the sector.
The manufacturing side of the industry illustrates the same disconnect between capacity and demand. Global electrolyser manufacturing capacity expanded from 6.8 gigawatts to 33.7 gigawatts between 2022 and 2024, with BloombergNEF projecting 52.6 gigawatts by the end of 2024, yet actual orders in 2023 struggled to reach 2 gigawatts. Electrolyser manufacturers’ share prices have fallen as much as 95% from their 2021 peaks, and Plug Power, one of the sector’s longest-established names, has told investors it may not survive to its 25th year without new financing after 24 years of losses. The concentration of that overbuilt capacity matters too: China now accounts for somewhere between 60% and 68% of global electrolyser manufacturing capacity, spread across more than 350 domestic makers who have cut alkaline electrolyser prices by roughly 73% since 2021 in a price war one industry analyst has described as pushing manufacturers to lose money simply to hold market share, with as much as 70% of those makers expected to disappear through consolidation. The “massive expansion of…electrolyser deployment” that closing the gap to 100 million tonnes would require is, in the country doing most of that expanding, already an oversupplied and financially distressed industry rather than a scaling success story.
The practical questions the sector has to answer, on electrolyser capital cost, on renewable power availability at the scale required, and on offtake commitments from steel, chemicals, shipping and aviation firms that have so far signed far more letters of intent than binding, cash-backed contracts, are the same regardless of whether the eventual 2050 figure lands at 100 million tonnes or 300 million. Every recent revision to that figure, from the IEA’s, from DNV’s, and implicitly from BNEF’s own cost work, has moved in one direction, and the pace of actual deployment has not moved to meet even the lower end of the range.
