Policy & Markets

Has the gas industry been written off too early?

20 November 2026 · by Marcus Wren
7 min read·1519 words·Updated 20 Nov 2026

11 gigawatts. That’s roughly the gas generation capacity still sitting in the NEM’s registered plant list, according to AEMO’s generation information data, and on the coldest mornings of this winter just gone a decent slice of it was running flat out while Twitter energy commentary insisted gas was basically finished. Two things can be true at once. Gas is shrinking as a share of annual generation. And gas is still the thing the system operator reaches for when the wind drops and the sun sets on a still, cold evening in July. The gap between those two facts is where most of the public argument about gas gets sloppy.

I want to make the contrarian case carefully, because it’s easy to sound like an apologist for an industry that genuinely does need to shrink. It does. Nobody serious disputes that gas generation’s share of the NEM’s energy mix is heading down over the next decade, and it should. But “shrinking” and “finished” are different words describing different things, and a lot of the commentary around the Australian energy transition has quietly swapped one for the other.

What the closure timetable actually shows #

Start with the coal exits, because that’s really what’s driving the gas conversation. Eraring’s closure has been pushed out to 2027, a deal struck between Origin Energy and the NSW government that I’ve covered in detail elsewhere on this site – see Eraring at 2027: the deal that bought NSW two more winters. Yallourn in Victoria is pencilled in for 2028. Every one of these closures removes dispatchable capacity from a grid that is adding enormous amounts of weather-dependent generation but comparatively little firm capacity to match it.

Batteries are arriving fast, and I’ve written favourably about what the Waratah Super Battery and others are doing for system strength. But a battery built for a two-hour firming job in a capacity market dominated by four-hour and overnight shortfalls is not the same animal as a gas peaker that can run for a week straight if a cold snap sits over Victoria and South Australia simultaneously. I made this argument in detail in Gas peakers versus big batteries: who actually wins the firming job, and nothing this year has moved me off it. Batteries win the short, sharp dispatch windows. Gas still wins the sustained, multi-day shortfall. Pretending otherwise because batteries are cheaper to build per megawatt is the kind of error that looks fine on a spreadsheet and ugly on a cold Tuesday in the Latrobe Valley.

Follow the money on new gas investment #

Here’s where the “written off” thesis runs into actual capital allocation. Woodside sanctioned Scarborough. Santos remains central to east-coast supply arguments around Narrabri and Moomba, disputes I’ve traced in Santos at the centre of the east-coast gas standoff. The Kurri Kurri peaker in the Hunter eventually got built, years late and over budget, which I wrote about at the time in Kurri Kurri: the peaker that took the long way round, but it got built, by a Commonwealth-owned entity, Snowy Hydro, because the market signal said firm capacity was needed and nobody else was queuing up to supply it fast enough.

Follow the money and you don’t see an industry in terminal retreat. You see an industry contracting in one dimension – share of annual energy, while remaining stubbornly necessary in another: dispatchable backup during system stress. Capital doesn’t flow toward things universally regarded as obsolete. It flows toward things with a cash-generating future, even a narrower one. Gas peakers are increasingly being built and valued as insurance assets, not baseload assets, and insurance is a business model, not a eulogy.

The east-coast supply problem is real and gets worse before better #

None of this lets the gas industry off easily, though. The east-coast gas market has a genuine supply-availability problem heading into winter 2028, something the AEMC and AEMO have both flagged repeatedly in their respective reports, and I’ve set out the mechanics of that timeline in Winter 2028 is when the east-coast gas question gets answered. Domestic reservation policy, LNG export contracts signed a decade ago, and declining Bass Strait output are colliding at roughly the same point on the calendar. That’s not a renewables-versus-gas story. That’s a gas-versus-gas story, where east-coast generators compete with LNG exporters for the same molecules, and consumers sit downstream of both.

So the honest version of “has gas been written off too early” has two halves. As a generation fuel for firming, gas is being under-built relative to what AEMO’s own Integrated System Plan modelling implies the NEM will need through the 2030s. As a commodity supply chain on the east coast, gas has real structural tightness that no amount of enthusiasm for peakers fixes, because you still need the gas to burn in them.

Where the consensus gets it wrong #

The consensus view, roughly, runs like this: renewables plus batteries plus transmission equals the whole solution, and gas is a transitional nuisance best minimised. I think that’s wrong on timing, not on direction. Transmission builds are running years behind schedule – I’ve covered VNI West’s ongoing arguments with its own route map in VNI West: a transmission line still arguing with a map, and the REZ rollout gap in NSW in NSW’s coal exit and the REZ rollout: a widening gap. Every month a transmission project slips is a month the system leans harder on whatever dispatchable capacity already exists behind the meter, which in most states still means gas turbines sitting at the end of a local feeder.

My mildly unpopular view, for what it’s worth: the industry most likely to be quietly grateful for every transmission delay and every battery project that slips its connection date is gas generation. Each slippage extends the runway on assets that would otherwise be stranded. That’s not a conspiracy, it’s just how option value works when you’re the backup plan and the primary plan keeps missing deadlines.

Hydrogen blending and carbon capture are the industry’s weakest argument #

Where I think the gas lobby overreaches is in selling hydrogen blending and carbon capture as a long-term decarbonisation pathway for existing gas infrastructure. Green hydrogen only makes commercial sense in a narrow set of applications right now, mostly heavy industrial feedstock rather than combustion blending into the existing gas network: a distinction I went through in Green hydrogen: where it genuinely makes sense. Carbon capture at scale on gas generation remains unproven in Australia at anything like the volumes claimed in some company investor decks. I’d treat both as a sideshow to the real argument, which is about gas as a firming fuel for the next 10 to 15 years, not a 2050 decarbonised gas network.

This is the bit the gas industry’s own marketing gets wrong, in my read. The stronger case for gas isn’t “gas can be made green.” It’s “gas is the least-bad insurance policy available while the grid builds out the storage and transmission it actually needs,” and that’s a less flattering pitch but a far more defensible one. A cricket tragic might call it the number eleven batter who isn’t there to score runs, just to not get out while the real work happens at the other end.

What the Safeguard Mechanism does to the arithmetic #

The Safeguard Mechanism adds a genuine constraint here that doesn’t get enough airtime in the “gas forever” camp’s arguments. Baselines are tightening for covered facilities, and I covered the mechanics of that squeeze in The Safeguard Mechanism’s baseline problem for heavy industry. Gas-fired generators above the threshold face the same declining baseline trajectory as any other covered facility, administered by the Clean Energy Regulator. That doesn’t kill gas peaking, which typically runs at low capacity factors and may sit under relevant thresholds depending on output, but it does put a ceiling on how much any gas generator can lean on sustained high-output running without wearing a growing compliance cost. The regulatory settings available through the Clean Energy Regulator’s published guidance are worth checking directly rather than taking a company’s framing of its own exposure at face value.

So, written off too early or not? #

My read: yes, on the specific and narrow point of dispatchable firming capacity through the back half of this decade. No, on the broader claim that gas has a long-term growth story in Australian electricity. Those aren’t contradictory positions, they’re just two different time horizons answering two different questions.

The industry’s defenders oversell gas as a transition fuel with decades of runway. The industry’s critics undersell how much the next five winters still depend on turbines that burn methane when the wind doesn’t blow. Both camps would do better services to the public argument by being specific about which horizon they’re talking about. Until transmission and storage genuinely catch up to the pace of coal closures, and the ISP’s own modelling suggests that gap persists into the early 2030s, gas remains less a legacy industry than an unglamorous utility holding the line. Whether that’s a comfortable place for investors to sit for another decade is a separate question, and one I suspect several boardrooms in Adelaide and Brisbane are still working through.

– Marcus Wren, Editor

Photo by Nikola Johnny Mirkovic on Unsplash