EnergyAustralia’s Tallawarra B gas peaker, down near Lake Illawarra, is now into its second winter of proper dispatch, and the pattern repeats itself with almost boring regularity. It sits idle for weeks. Then a cold front parks itself over the southeast, wind generation across three states falls away at once, and for two or three days the plant runs close to flat out. Nobody at AEMO’s control room in Sydney finds this remarkable. It’s the plant doing the one job it was built for, which is not to generate cheap energy, but to be there.
That job has a name in the industry, and it’s the word you’ll keep hitting if you read anything about the grid’s transition: firming. It gets used loosely, sometimes to mean batteries, sometimes gas, sometimes pumped hydro, occasionally as a synonym for ‘backup’ that flattens three quite different technologies into one. Worth pulling apart properly, because the differences matter more than the headline word suggests.
Start with the units, because this argument usually goes wrong here #
A wind or solar farm is sized in megawatts of capacity, and everyone quotes that number because it’s the one on the press release. But megawatts only tell you the ceiling. What actually keeps the lights on second by second is megawatt-hours delivered exactly when the system needs them, which AEMO’s five-minute dispatch engine is constantly reconciling against demand: a process I’ve written about separately at how the NEM dispatches power every five minutes. Firming is the answer to the gap between those two numbers. Wind and solar are cheap and plentiful in aggregate, but their output at any given half-hour is a function of weather, not of what the grid happens to want. Firming capacity is whatever fills that gap on short notice: batteries, pumped hydro, gas turbines, or in a pinch, demand response.
The confusion starts because all four of those things get called ‘firming’ in the same sentence, as if they’re interchangeable. They’re not. A big battery can hit full output in under a second and hold it for one to four hours before it’s flat. A pumped hydro scheme can run for eight, ten, sometimes days if the dam’s full, but takes longer to spin up and costs a great deal more to build. A gas peaker sits in between on speed but can run for as long as the fuel supply and the contract allow, which, on a still, cold week with low wind, is precisely the scenario where batteries run out and pumped hydro schemes are still waiting on their tunnel boring machines to finish, a point I made when writing about Snowy 2.0’s long construction road.
Why the sun makes this worse, not better #
It would be tidy if the problem were simply ‘renewables are intermittent, therefore we need backup.’ The sharper version of the problem is the shape of the day. Rooftop solar now pushes minimum daytime demand on the grid down to levels that would have seemed implausible a decade ago, a trend I went through in detail in rooftop solar and the midday minimum demand problem. Around the middle of the day, when the sun’s doing most of the work, there’s often too much supply and prices go negative. Then the sun drops away in late afternoon just as everyone gets home, switches on the air conditioner, and starts cooking dinner. That evening ramp, not some flat average shortfall, is where firming earns its money. It’s a narrow, predictable window most days, and an unpredictable multi-day slog during a heatwave or a cold snap with no wind. Different problem, different technology, arguably different market.
Where gas actually sits in that stack #
Gas turbines have three things going for them that batteries and pumped hydro, for all their virtues, don’t yet match at scale. They can run for as long as there’s gas in the pipe, which matters on the multi-day events rather than the daily ramp. They can be built relatively fast compared with a new pumped hydro cavern. And critically, an open-cycle gas turbine can provide the kind of grid-forming inertia and fast frequency response that a system with less spinning synchronous plant increasingly needs: a separate but related argument I laid out in grid-forming inverters: why the grid suddenly needs them.
What gas doesn’t have going for it any more is cost of energy. Gas peakers are expensive to run per megawatt-hour, because east coast gas prices haven’t behaved themselves for years, and I’ve covered that mess more than once at has the Australian gas industry been written off too early? The economics of a peaker were never about cheap energy. They’re about being paid to sit around most of the year doing nothing, then earning its keep in the handful of hours that actually matter: a business model that only really works if someone underwrites the standing cost, which is exactly what the Capacity Investment Scheme is now doing for a slab of new firming capacity, gas included, as I set out in how the Capacity Investment Scheme actually works.
Batteries versus gas versus pumped hydro, without picking a favourite #
I’ve gone through the head-to-head numbers before, in more detail than fits here, at gas peakers versus big batteries: who firms the grid? and pumped hydro versus batteries: what actually firms the grid. The short version, without repeating myself word for word: batteries have won the fast, shallow end of the market decisively. Anything under about two hours, and increasingly out to four, a battery beats a gas peaker on cost and on speed, and the build times aren’t comparable – months versus years. Where the argument gets genuinely interesting is the deep end: the multi-day, low-wind, high-demand event that AEMO’s Electricity Statement of Opportunities flags as the scenario the grid is actually short for, not the daily evening peak. That’s the window pumped hydro was designed for, and it’s also where gas, faults and all, still does something nothing else on the grid does yet at the scale required.
The bit I think gets oversold on both sides #
Here’s my actual read, and it’s not a comfortable one for either camp. Gas advocates like to talk about firming as though it justifies a large, permanent gas peaking fleet running for decades. It doesn’t, not at the volumes some of the CIS contracts imply. Batteries are eating the shallow end of the firming task faster than most 2022-era modelling assumed, and every gigawatt-hour of storage that gets built pushes gas further towards a genuinely thin sliver of extreme-event running hours, which is a terrible economic proposition for a plant that costs hundreds of millions to build. Some of the new OCGTs being underwritten right now may end up running at levels closer to an emergency reserve than a peaker, and someone will eventually have to explain that to whoever’s carrying the capital cost.
But the renewables-only crowd who wave away gas as a rounding error are kidding themselves too. Australia doesn’t yet have anywhere near enough long-duration storage built, permitted, or even fully financed to cover the multi-day low-wind events that show up in AEMO’s own reliability forecasts. Until pumped hydro schemes like Snowy 2.0 and Kidston are actually delivering, and until batteries reliably stack four-hour blocks back to back across a whole state, something has to carry that tail risk. Right now that something is still mostly gas, and pretending otherwise doesn’t make the physics go away.
What to actually watch #
The number that matters isn’t gigawatts of gas capacity contracted, it’s hours run per year. A peaker that runs 40 hours a year is doing exactly the job firming implies: rare, expensive, essential. A peaker running 1,500 hours a year has quietly become baseload with better PR, and that’s a different conversation about cost and emissions entirely, one closer to the arguments I’ve made about the Safeguard Mechanism and who ends up carrying industrial gas costs. Keep an eye on AEMO’s published dispatch data and the AER’s state of the market reporting, because that’s where the actual running hours show up, not in the announcement press release.
For what it’s worth, I was out at Perth Observatory a fortnight back for one of their public viewing nights, waiting on a break in cloud that never quite came. Someone standing next to me, an engineer as it turned out, made the observation that firming and stargazing have the same enemy: you don’t get to choose the weather, you just plan for the nights it won’t cooperate. Fair point, and about as tidy a summary of the whole debate as I’ve heard.
The grid doesn’t need gas to win the argument, and it doesn’t need gas to disappear either. It needs enough of the right kind of firming, in the right duration bucket, priced honestly for the hours it actually runs. Whether the current wave of CIS-backed gas contracts gets that duration bucket right is the question worth revisiting in a couple of years, once the running-hours data is actually in.
– Tom Fitzgerald, Baseload & Fuels Correspondent
Photo by Energie-portal.sk on Unsplash