Fortescue walked away from its Gibson Island green ammonia project in Queensland in 2024. Woodside shelved its H2Perth and H2TAS proposals around the same time. The South Australian government’s hydrogen power station near Whyalla, once the centrepiece of its Hydrogen Jobs Plan, was deferred when the costs came back higher than the business case could carry. Three announcements, three different companies, one pattern. The short version is that green hydrogen didn’t turn out to be the universal solvent everyone was spruiking around 2021, and the industry has spent the past couple of years finding out, expensively, where it actually works.
That’s not a story about a technology failing. Electrolysis is not new — it’s been done in industrial chemistry for the best part of a century. It’s a story about a lot of very confident numbers on investor slide decks meeting the actual cost of renewable electricity, transport, storage and compression. I’ve been following the Australian hydrogen pipeline since the first National Hydrogen Strategy came out, and the gap between the 2020 hub map and what’s actually under construction now is instructive. So let’s agree on the units first, then work out what survives contact with reality.
What green hydrogen actually is, and what it isn’t #
Green hydrogen is hydrogen gas split from water by electrolysis, using renewable electricity to run the current. That’s the whole trick. Grey hydrogen comes from steam-reforming natural gas, which is cheaper but carries the associated carbon dioxide, most of which currently goes straight to air. Blue hydrogen is the grey process with carbon capture bolted on — Santos’s work at Moomba, discussed in our piece on Santos’s Moomba carbon bet, is the local example of that path, and it comes with its own argument about how much carbon actually gets captured.
Green hydrogen’s appeal is simple: no gas feedstock, no carbon dioxide stream to deal with, just water, electricity and a stack of membranes. The problem is that electrolysis is not free. You lose energy converting electricity into hydrogen, more energy compressing or liquefying it for transport and storage, and — if you’re planning to burn it in a turbine or a fuel cell to make electricity again — a further chunk converting it back. Do that whole loop and you can lose something like two-thirds of the original electricity before it reaches a socket again. Batteries, for comparison, return roughly 85 to 90 cents of every kilowatt-hour you put in. That arithmetic is the entire reason hydrogen lost the argument over daily grid firming almost before it started, a point we’ve made before in our comparison of pumped hydro versus big batteries.
Why it never made sense for your car or your gas stove #
The hydrogen-for-everything pitch a few years back had it running domestic heating, replacing petrol at the servo, and firming the grid overnight. Most of that has quietly gone away, and for good reason. Battery electric vehicles beat hydrogen fuel-cell cars on cost and infrastructure simplicity — you plug into a wall, not a specialised high-pressure bowser, a point covered in our explainer on electric cars versus petrol cars. Blending hydrogen into the gas network for home heating turns out to embrittle old steel and cast-iron pipe at anything beyond a small percentage, and even at full substitution the household appliance retrofit bill dwarfs the emissions saved. CSIRO’s GenCost work, the standard reference the sector argues over every year, has consistently shown green hydrogen sitting well above the cost of direct electrification for these small, distributed, low-temperature uses.
None of that makes hydrogen a bad technology. It makes it a technology with a fairly narrow lane, and the industry spent a few years pretending the lane was a six-lane freeway.
Where the physics actually cooperates #
The honest case for green hydrogen sits in places where there is no easy electrical substitute — where you need a molecule, not just a current.
Ammonia is the clearest example. Australia already makes a lot of it, mostly for fertiliser, using hydrogen stripped from natural gas. Yara’s Pilbara operation at Karratha has explored producing part of its ammonia feedstock via renewable hydrogen, with ARENA support behind the feasibility work — swapping the hydrogen source rather than reinventing the plant. That’s the kind of project that survives a downturn in hydrogen enthusiasm, because the ammonia market already exists, the customer is already buying, and you’re changing one input, not building a whole new value chain from scratch.
Steelmaking is the second genuine case. Conventional blast furnaces use coking coal as both fuel and chemical reducing agent. Direct reduced iron processes can use hydrogen instead of coal to strip oxygen from iron ore, and several Australian and international steelmakers have run pilot-scale trials on exactly that basis. It’s early, it’s capital-intensive, and Australia’s iron ore is not always the easiest chemistry for hydrogen-based reduction — but if the country wants to keep exporting value rather than just raw ore, this is one of the few plausible routes, and it’s the kind of heavy-industry emissions problem the Safeguard Mechanism is explicitly designed to keep pressure on.
Shipping fuel and high-temperature industrial heat — glass, cement, some smelting processes — round out the list. In each case the alternative to hydrogen or its derivatives (ammonia, methanol) isn’t a battery. It’s still burning something. That’s the test I’d apply before believing any hydrogen pitch: if there’s a battery or a heat pump that does the job more efficiently, hydrogen is competing against physics it can’t win. If there isn’t, it’s worth a proper look.
The Australian pipeline, after the correction #
The 2019 National Hydrogen Strategy mapped out a network of regional hydrogen hubs — Pilbara, Gladstone, Hunter Valley, Bell Bay in Tasmania, Darwin among them — on the assumption that Japan and South Korea would be lining up to buy shiploads of green hydrogen or ammonia within a few years. That export assumption has moved a lot. Offtake agreements that looked firm on paper in 2021 have in several cases been renegotiated, delayed, or quietly not renewed, as Asian buyers ran their own cost assessments and found the same electrolysis-plus-shipping-plus-reconversion arithmetic that Australian developers were running into at home.
What’s left standing tends to be smaller, closer to an existing industrial customer, and less dependent on a hydrogen export ship that doesn’t exist yet. The projects still moving are mostly attached to fertiliser plants, mining haul-truck trials, or steel pilot programmes rather than standalone hydrogen-for-hydrogen’s-sake ventures. That’s a healthier place for the industry to be, frankly, even if it’s a smaller headline number than the gigawatt-scale announcements from a few years back. It’s a similar recalibration to what we’ve tracked with SunCable’s Australia-Asia PowerLink — big export ambitions meeting the slower, harder reality of financing, offtake and construction timelines.
The Capacity Investment Scheme’s quiet hydrogen exposure #
It’s worth watching how much of the Commonwealth’s Capacity Investment Scheme support ends up flowing toward hydrogen-adjacent firming projects, versus batteries and pumped hydro, as tender rounds proceed. AEMO’s own Integrated System Plan modelling has hydrogen playing a modest role in long-duration seasonal storage in the 2040s scenarios, not as a routine daily firming tool — a distinction that gets lost whenever a minister stands in front of an electrolyser announcement. My read, for what it’s worth, is that governments have been more useful backing hydrogen where there’s a captive industrial customer already lined up than backing it as speculative grid infrastructure. The Whyalla deferral is the cautionary tale on the latter; Yara’s ammonia work is the better model of the former.
A cider-cellar sort of judgement #
I keep a batch of cider going most winters out past Bickley in the hills behind Perth, and the thing you learn fermenting anything is that some processes just have an efficiency ceiling you can’t argue your way past — you can rush it, but you lose flavour and yield doing so. Hydrogen has the same honesty problem. No amount of policy enthusiasm changes the round-trip loss of splitting water and putting it back together. That’s not a reason to write the technology off — plenty of chemical processes with modest efficiency are still worth doing because there’s no better alternative — but it is a reason to be suspicious of anyone still pitching hydrogen as a grid-storage or transport fix in 2026. That argument had its moment, and the market has mostly moved on from it, the same way the debate over baseload coal closures moved from ideology to scheduling once the actual retirement dates started arriving.
Where green hydrogen genuinely earns its place is narrower and less exciting than the 2021 pitch decks promised: fertiliser, steel, shipping fuel, high-heat industrial processes that have no electrical alternative. That’s still a meaningful decarbonisation task — Australia’s ammonia and steel-adjacent exports are not small numbers — it’s just not the everything-fuel some of the early spruiking implied. The projects surviving the correction are the ones that were honest about that from the start.
Whether the remaining hubs get built at the scale originally announced, or whether they shrink further as financing costs bite, is the thing I’ll be watching over the next couple of budget cycles. The molecule still has a job. It’s just a smaller job than advertised.
— Tom Fitzgerald, Baseload & Fuels Correspondent
Photo by Andrew Van Hofwegen on Unsplash