Fortescue quietly wound back its green hydrogen ambitions through 2024 and 2025, shelving projects in Queensland and overseas and writing down the capital already spent. Origin Energy stepped away from the Hunter Valley ammonia plan a little earlier. Andrew Forrest kept the flame burning longer than most, but even he trimmed the targets. None of that means the molecule is dead. It means the industry spent several years trying to force green hydrogen into jobs it was never going to be good at, while mostly ignoring the two or three jobs it’s actually suited for. Worth sorting out which is which before the next investment cycle repeats the mistake.
I’ve sat through enough hydrogen roadshows in Perth conference rooms to have a rough catalogue of the pitch: hydrogen for home heating, hydrogen for passenger cars, hydrogen blended into the gas network, hydrogen trucks, hydrogen as grid storage competing with batteries. Most of that has fallen away as the efficiency maths caught up with the slide decks. What’s left standing is narrower, duller, and, in my view, considerably more investable than the full rainbow version ever was.
Agreeing on what green hydrogen actually costs #
Start with the physics, because the economics follow directly from it. Making hydrogen by electrolysis takes roughly 50 to 55 kilowatt-hours of electricity per kilogram of hydrogen produced, depending on the electrolyser technology and how hard it’s run. That hydrogen holds about 33 kilowatt-hours of energy per kilogram on a lower heating value basis. So before you’ve compressed it, stored it, transported it or converted it into anything useful, you’ve already lost something like a third of the input electricity to the splitting process alone.
Every additional step costs more. Compress it for pipeline or truck transport and you lose more energy to that compressor. Convert it to ammonia for shipping and you lose more again, then lose a further chunk converting it back to hydrogen at the other end if that’s what the buyer actually wants rather than ammonia itself. Run it through a fuel cell to make electricity again and you’re down to round-trip efficiencies that batteries beat by a wide margin, which is exactly why the AEMO Integrated System Plan treats hydrogen as a firming afterthought rather than a backbone, and why gas peakers versus big batteries: who actually wins the firming job barely mentions it as a contender for grid balancing this decade.
None of this is an argument against hydrogen. It’s an argument for being honest about where those energy losses matter and where they don’t.
The jobs hydrogen can’t win #
Anywhere there’s a direct electrical alternative, hydrogen loses on cost, full stop. Heating your home with a reverse-cycle unit beats a hydrogen boiler by a factor of roughly three to four on primary energy use, because the heat pump moves heat rather than making it from scratch, and because you’ve skipped the electrolysis loss entirely. Driving a battery electric car beats a hydrogen fuel-cell car for the same reason. You can see this in the hydrogen refuelling network that largely failed to materialise in Australia while home battery uptake, however patchy, kept climbing; see home battery payback: does it stack up yet for where that economics landed.
Grid-scale storage is the same story, just with bigger numbers. A battery charges and discharges at something like 85 to 90 per cent round-trip efficiency. Hydrogen storage, taken through electrolysis and back through a fuel cell or turbine, is lucky to clear 35 to 40 per cent round trip. You can build the hydrogen loop to store energy for weeks rather than hours, which batteries genuinely can’t do yet at reasonable cost. That’s the one niche argument for hydrogen as storage, long-duration seasonal balancing rather than daily cycling. But nobody’s built that at scale in the NEM, and the economics depend on electricity being essentially free for long stretches, which is a big ask outside South Australia on a windy spring afternoon.
Where the chemistry actually needs the molecule #
Here’s the distinction that gets lost in the broader hydrogen debate: some industrial processes don’t want electricity, they want hydrogen as a chemical feedstock or reducing agent, and there’s no way around that with a heat pump or a battery.
Ammonia production is the clean case. Roughly 180 million tonnes of ammonia get made globally each year, overwhelmingly via the Haber-Bosch process, and overwhelmingly using hydrogen sourced from steam-methane reforming of natural gas – what the industry calls grey hydrogen. Swap that grey hydrogen for green hydrogen made with renewable electricity and you haven’t added a new product to decarbonise; you’ve decarbonised an existing global industry that was never going anywhere. Fertiliser demand doesn’t disappear because the grid gets cleaner. That’s a genuinely different proposition to inventing a hydrogen car market from scratch.
Steelmaking is the second real case, via direct reduced iron using hydrogen instead of coking coal to strip oxygen from iron ore. This is still mostly pilot-scale globally, and Australia’s iron ore majors have been cautious rather than evangelical about the timeline – rightly, given the capital intensity and the fact that Pilbara ore grades complicate the direct reduction chemistry more than Brazilian or Swedish ore does. But the demand pull is real: European and Japanese steelmakers are under genuine regulatory and customer pressure to decarbonise, and green iron, not green hydrogen for its own sake, is the product Australia could plausibly sell into that.
Why Pilbara ports and the Hunter keep coming up #
The projects that have survived the shakeout share a pattern: cheap, abundant renewable resource next to port infrastructure and an export customer who already wants ammonia or green iron for reasons unrelated to hydrogen hype. The Pilbara’s combination of some of the best solar and wind co-location in the country, existing iron ore export infrastructure, and Japanese and Korean buyers under their own domestic decarbonisation mandates, is about as good a starting position as this industry gets anywhere in the world.
Gladstone and the Burnett region in Queensland get talked about for similar reasons: existing ammonia and explosives manufacturing, port access, and proximity to renewable zones including the broader build-out tracked in Powerlink’s big bet: CopperString and the north-west minerals grid. The project economics there still depend heavily on the cost of renewable electricity landing well under the $40 to $50 per megawatt-hour mark on a sustained basis, which is why the final investment decisions keep slipping rather than arriving on schedule.
That slippage is the part spruikers tend to skip over. ARENA’s own technology readiness assessments and funding rounds for hydrogen electrolysis have consistently flagged capital cost as the binding constraint, not the chemistry. Electrolyser costs have come down since the early 2020s but not nearly as fast as some of the 2021-era roadmaps assumed, and the gap between announced hydrogen hub capacity across the country and hydrogen hub capacity actually under construction remains wide enough to walk a Mack truck through it, Pilbara road train style.
The subsidy question nobody likes answering plainly #
Green hydrogen for ammonia and green iron is not commercially viable today against grey hydrogen and conventional blast-furnace steel, not in Australia, not anywhere. It’s viable at some future point if either carbon pricing on competitor products bites hard enough, overseas buyers pay a durable premium for low-carbon inputs, or governments underwrite the gap with production credits, as the federal Hydrogen Headstart programme and various state schemes have attempted to do.
My honest judgement, and this is the mildly unfashionable bit, is that the subsidy bridge is the right call for ammonia and green iron specifically, and was the wrong call for the broader hydrogen economy vision that got pitched alongside it. Backing an export industry that decarbonises someone else’s existing demand, using Australia’s genuine comparative advantage in renewable resource, is a defensible industrial policy. Backing hydrogen cars and hydrogen home heating networks, which several state and federal pilot programmes briefly did, was closer to picking a technology because it sounded futuristic rather than because the units added up. The Clean Energy Regulator’s data on where hydrogen-related Australian Carbon Credit Unit methodologies and guarantee-of-origin certificates have actually been used tells the same story: concentrated around industrial feedstock applications, thin everywhere else.
What’s actually likely by the early 2030s #
A handful of ammonia and green iron projects reaching final investment decision, mostly in the Pilbara and one or two Queensland or South Australian sites, running on dedicated renewable generation rather than grid power, selling into Japanese, Korean or European buyers with their own decarbonisation mandates. Not a hydrogen refuelling station on every highway. Not a hydrogen boiler in every home. Not hydrogen trucks replacing diesel fleets at any meaningful scale, because battery electric trucking has moved faster than most 2020-era forecasts assumed, in much the same way rooftop solar and home batteries outpaced expectations on the residential side – see the pattern described in is rooftop solar quietly breaking the grid? for how quickly a cheaper, simpler technology can eat a more complicated one’s lunch.
The companies still standing in this space, including Fortescue’s remaining projects and the ammonia proposals around Gladstone, aren’t making a bet on hydrogen as a concept. They’re making a narrower bet on specific customers who need a specific chemical input and are willing to pay something for it being green. That’s a smaller industry than the 2021 slide decks promised. It might also be the only version of it that actually gets built.
I’ll admit I was more sceptical of the whole hydrogen push a few years back than the ammonia case deserved. It’s easy to lump a genuinely sound industrial chemistry argument in with the noisier stuff around it. The lesson, same as it ever is with energy technology, is to work out which application the physics actually supports before arguing about the subsidy. Agree on the units first; the politics sorts itself out a lot faster after that.
– Tom Fitzgerald, Baseload & Fuels Correspondent
Photo by Andreas Felske on Unsplash