Four gigawatts of coal capacity sitting on the shore of Lake Macquarie, and the closure date attached to it has now moved twice in three years. That’s the plain fact sitting under every argument about Eraring. Origin Energy first flagged an exit around 2025, then under pressure from the NSW government agreed to run on to August 2027, and the industry has spent the time since arguing about whether that’s cautious or just slow. I’ve gone back and forth on it myself, and I reckon the honest position is: cautious, with a reasonable chance of slipping again.
Eraring is the biggest power station in the National Electricity Market, full stop. Four units, roughly 2,880 megawatts of nameplate capacity on the books, built in the early 1980s on the edge of Lake Macquarie near Dora Creek. On a hot February afternoon it still does more heavy lifting for NSW demand than any other single asset in the state, coal, gas or otherwise. That’s precisely why nobody (not Origin, not AEMO, not the NSW government) was comfortable just letting the original retirement date stand without a very close look at what would actually fill the gap.
How 2025 became 2027 #
The short background, because it matters for understanding the current deal: Origin announced an earlier closure horizon for Eraring back in 2022, flagging it could shut as soon as 2025, seven years ahead of the plant’s original scheduled life. That announcement landed hard in Canberra and in Macquarie Street, because AEMO’s own modelling at the time didn’t have enough firm replacement capacity and transmission ready in the Hunter-Central Coast corridor to absorb the loss cleanly. We covered the mechanics of that slippage in detail at the time – see Eraring power station: how 2025 became 2027, and the short version holds: NSW Treasury and Origin negotiated a deal in 2024 that keeps units running to August 2027, with the state carrying downside risk on energy market revenue and taking on some of the closure costs including worker transition.
That arrangement wasn’t free, and nobody pretended it was. It’s effectively a contingent payment structure: NSW compensates Origin if wholesale prices and capacity payments don’t cover the cost of keeping four ageing coal units turning over for an extra two years, in exchange for guaranteed capacity through a period AEMO flagged as tight. Origin’s own ASX filings describe it as a deed with the NSW government rather than a subsidy in the ordinary sense, and I’d push back gently on anyone calling it a straightforward handout. It’s closer to a reliability insurance premium the state decided was cheaper than blackouts.
The battery going in next door #
The part of the Eraring story that gets less airtime than the closure date is what Origin is actually building on the same footprint. There’s a large battery energy storage system under development at the Eraring site, designed to use existing grid connection infrastructure, transformers, switchyard capacity, transmission easements, rather than starting from a paddock with nothing. That’s not a small advantage. Grid connection queues around NSW are long enough that a project which can reuse an existing connection point jumps years of process that a greenfield battery in, say, the New England REZ would still be working through. We’ve written before about how slow some of those connection pipelines move; see New England REZ: where the wind-solar-storage build actually stands for the contrast.
Origin has staged the battery in phases, with an initial tranche already delivering grid services and further capacity planned as the surrounding units wind down. The logic is straightforward: retire the coal units progressively rather than all at once, and have storage absorbing some of the firming load as each unit comes offline, rather than waiting until the last turbine stops and hoping the replacement capacity has turned up on schedule. It’s a sensible staging approach, and it’s one of the few parts of this whole saga I’d call genuinely well designed rather than merely adequate.
Why NSW couldn’t just let it close #
Go back to basics on the units, because the Eraring argument gets muddled when people skip this step. Eraring alone has historically supplied somewhere around a quarter of NSW’s electricity generation in a typical year. There is no single replacement asset anywhere near that scale queued up to take its place by 2027, not Kurri Kurri, which we profiled in Kurri Kurri: the peaker that took the long way round and which is a peaking gas plant built for hours, not a baseload substitute; not the Waratah Super Battery, which does real work smoothing fault conditions and local congestion but isn’t sized to backfill a quarter of state demand; and not the Central-West Orana or New England REZs, both of which are behind the schedule AEMO’s Integrated System Plan originally assumed.
That mismatch, between how fast coal wants to leave and how fast replacement firm capacity and transmission can actually be built, is the central tension running through NSW’s whole transition right now. We’ve tracked it closely in NSW’s coal exit is outrunning its REZ build, and Eraring is the sharpest single example of it. AEMO’s most recent Electricity Statement of Opportunities flagged reliability gaps opening up in NSW through the back half of this decade if generation retirements run ahead of new capacity, which is more or less the exact scenario the Eraring deed was built to avoid.
What it’s costing, and who’s actually carrying it #
Agree on the units first, because this is where public commentary tends to get sloppy. The NSW deed is a contingent support arrangement, not a straight subsidy cheque, and the amount paid out depends on how wholesale electricity prices and capacity market revenue perform over the extension period. If prices stay reasonably strong, taxpayer exposure could end up fairly modest. If the market softens, or if unplanned outages at an ageing plant push availability down and costs up, the bill to the state rises. Origin’s own investor materials have been consistent on this point: the structure caps Origin’s downside while NSW absorbs volatility risk in exchange for certainty of supply through a period it judged too risky to leave to the open market.
The NSW government’s own public statements, and the state’s energy department commentary, have framed the deal as cheaper than the alternative of forced load shedding or having to rush emergency gas peaking capacity into the grid under time pressure. That’s a defensible call. Whether it was the cheapest possible path is a different question, and one that probably won’t be settled until the plant actually closes and someone tallies up the final number against what a faster REZ and transmission build might have cost instead.
The coal exit arithmetic, honestly stated #
Here’s where I’ll put a flag in the ground rather than sit on the fence. I think two years was roughly the right amount of extra runway, not because Eraring is cheap to keep operating, an ageing coal unit rarely is, but because the alternative of letting the original 2025 date stand would have meant NSW entering summer peak season without anywhere near enough firm dispatchable capacity behind the solar and wind that has gone in. AEMO’s modelling backed that read at the time, and nothing that’s happened with REZ transmission progress since has made me think the state could have absorbed the loss any faster. Readers who want the broader national version of this argument should look at Coal exits versus replacement builds: who’s actually ahead, which runs the same maths across every state closure on the table.
Where I part ways with some of the louder commentary is on the idea that this extension proves coal retirements generally should slow down. Eraring is a special case precisely because of its size and its location relative to Sydney demand. Smaller, more isolated coal units elsewhere in the NEM don’t carry the same systemic risk if they close on the original schedule, and treating every closure date as needing a Eraring-style rescue misreads the specific problem that was actually being solved here.
What to watch between now and August 2027 #
Three things will tell you whether this extension holds or slips again. First, battery commissioning progress on site, because every megawatt of storage actually energised and passing AEMO’s registration tests reduces the argument for yet another extension. Second, transmission augmentation progress feeding the Hunter and Central Coast, since a chunk of the reliability case for keeping Eraring running rests on network capacity that still needs to be built. Third, and this is the one I’d watch most closely, plant availability at Eraring itself. Ageing coal units don’t fail gracefully, and an unplanned extended outage on one or two units in the next eighteen months would sharpen the argument for an earlier wind-down rather than a later one.
I spoke to someone on the NSW energy planning side a few months back, off the record, who put it bluntly: the 2027 date is a planning assumption everyone is working to, not a guarantee carved in granite. That’s probably the fairest way to think about it. The Clean Energy Regulator and AEMO both publish regular updates on generator retirement notices and system reliability outlooks, and anyone wanting the primary data rather than secondhand commentary should go straight to AEMO’s published Electricity Statement of Opportunities and the retirement notice register the market operator maintains under the National Electricity Rules.
The bigger pattern this sets #
Eraring won’t be the last coal plant where the gap between announced closure and replacement readiness forces a government-brokered extension. EnergyAustralia is working through its own version of this at Yallourn, which we’ve covered in EnergyAustralia: the gentailer counting down to Yallourn’s exit, and the same pattern of closure dates moving to match transmission and storage readiness, rather than the other way around, is likely to repeat at every major coal closure left on the NEM calendar this decade.
What’s actually novel about Eraring is the battery sitting on the same site as the coal units it’s helping replace. If that staged build-down works the way Origin has planned it, pairing battery capacity growth with unit retirements rather than treating them as separate projects, it’s a template other gentailers closing large coal plants would do well to copy rather than reinvent. Whether NSW gets a second extension past August 2027 is still an open question. My guess, for what it’s worth, is that it won’t need one, but I’ve been wrong about Eraring’s calendar before, and so has everyone else covering this plant.
– Tom Fitzgerald, Baseload & Fuels Correspondent
Photo by Maksym Kaharlytskyi on Unsplash