Eight. That’s how many coal units have left the NEM since 2022 across NSW, Victoria and South Australia, against a replacement build that AEMO’s own Integrated System Plan modelling assumed would be running years ahead of where it actually sits today. Four of those units came out of NSW alone. The question running through every AEMC consultation paper and every gentailer investor call this year is whether the sequencing has quietly come unstuck, and whether anyone with the authority to slow it down is willing to say so out loud.
I don’t think the answer is as simple as either camp wants it to be.
The pro-coal crowd says we’re ripping out dispatchable capacity on an ideological timetable with nothing firm to replace it. The pro-transition crowd says the closures are investor decisions made on commercial grounds, not government fiat, and the grid has absorbed worse. Both of these things are true at once, which is usually a sign the real story is more boring and more specific than either slogan allows.
The closures were never really a policy choice #
Start with the plain mechanical fact that gets lost in this argument constantly. Coal plant owners are closing stations because the economics of running fifty-year-old thermal assets into a market with rising rooftop solar penetration and falling daytime prices don’t work anymore. Origin brought forward Eraring’s planned exit date, then walked part of that back under government pressure, which tells you something about how fragile the commercial case for keeping ageing coal open actually is. It only survives with a subsidy attached. EnergyAustralia is working through its own countdown on Yallourn, a station it has flagged will close well before the end of the decade.
Nobody is being forced out early by a minister with a pen. They’re being pushed out by a market that no longer pays enough for midday generation from an asset built for a grid that doesn’t exist anymore.
Where the replacement maths actually breaks down #
Here’s the number that should worry people more than it does: transmission connection queues, not generation capacity, are the actual bottleneck. AEMO’s capacity numbers for committed wind, solar and storage look fine on paper. The problem is getting that capacity connected to the grid at the pace the closure schedule assumes, which is a transmission and REZ story, not a generation story. I’ve written before about the gap between NSW’s coal exit and its REZ rollout, and the gap hasn’t closed, it’s shifted shape. Central-West Orana is making progress. New England is moving more slowly than its own timetable promised. Transmission build-out, as I’ve argued before, is fundamentally a question of who actually pays for the wires, and that argument isn’t settled either.
So when people ask “are we replacing coal fast enough”, the honest framing isn’t megawatts of nameplate capacity. It’s megawatts of nameplate capacity that can actually get power from a wind farm in the New England Tablelands to a substation in Sydney’s west without sitting in a connection queue for three more years.
The reliability debate nobody wants to finish #
This is the bit that deserves a harder look than the trade press usually gives it. AEMO’s reliability standard, the famous 0.002% unserved energy criterion, has been tested in real conditions over the past few summers and the grid has mostly held. Mostly. There have been lack-of-reserve notices, price spikes into the market price cap, and moments where batteries and demand response did genuinely heavy lifting that a few years ago would have fallen to a coal unit running flat out. I covered this territory in more detail in the piece on coal closures and the reliability debate nobody’s settled, and nothing that’s happened since has resolved it cleanly either way.
What’s changed is the margin for error. Every unit that closes without its firm replacement fully connected shrinks the buffer AEMO has to work with on a genuinely bad day: low wind, hot, a transmission line down for maintenance. That’s not a reason to keep clapped-out coal plants running past their use-by date. It is a reason to be honest about how thin that buffer has become.
Batteries are carrying weight they weren’t built for #
Big batteries have been the standout performer of the transition so far, and I’ll give them that credit freely. The Waratah Super Battery in NSW has become the shock absorber the grid needed during exactly the kind of volatile event that used to be coal’s job. But most grid-scale batteries running today are built with two hours of storage, sometimes four. That’s brilliant for smoothing a price spike or catching a sudden drop in wind output. It’s a different job entirely from replacing the sustained, multi-hour output of a 700 megawatt coal unit through a still winter evening. I’ve made this argument before in looking at whether big batteries are being built for the wrong job, and the closures happening now are testing exactly that mismatch in real time, not in a modelling scenario.
Gas peakers fill some of that gap, and the firming maths between gas and batteries is a live argument I don’t think has a settled winner yet. It depends entirely on how many hours of firming you actually need on the worst night of the year, a question AEMO’s planners wrestle with differently to how a gentailer’s trading desk wrestles with it.
What the money is actually saying #
Follow where the capital’s going and you get a cleaner answer than any press release. Private capital, superannuation funds included, keeps backing storage and transmission-adjacent renewables projects at a pace that outstrips government funding rounds. That’s a vote of confidence in the long-term economics of the replacement fleet. But the same capital is notably cautious about anything requiring a transmission connection beyond the next few years, because connection risk has become the thing that kills project returns, not technology risk, not even price risk. Developers will tell you privately that a signed power purchase agreement means less than a confirmed connection date these days. That’s a quiet admission the bottleneck isn’t generation, it’s wires, and it undercuts the simple story that renewables “aren’t ready” to replace coal. They’re ready. The grid isn’t ready to carry them to where the coal used to be.
My read, and where I think the politics gets it wrong #
Here’s my mildly unpopular view, having watched four coal closures announced, debated and walked back over three winters of covering this beat: the closure dates aren’t the problem. The problem is that governments and regulators keep treating transmission and REZ delivery as though it runs on the same calendar as generation investment, when every piece of evidence from New England to VNI West says it doesn’t. Projects like VNI West are stuck in route disputes measured in years, not months, while the coal units they’re meant to help replace are on fixed retirement schedules measured in the same years. Matching those two calendars is the actual policy task, and it’s being handled with far less urgency than the generation build gets.
I’ll admit there’s a cricket analogy sitting right here about declaring too early with wickets in hand, but I’ll leave it: some comparisons write themselves and don’t need me forcing them.
The AEMC’s work on transmission access reform and connection process redesign is the quieter, less photogenic fix that actually matters more than any single battery announcement. The AEMO Integrated System Plan lays out the generation and storage volumes needed with reasonable clarity. What it can’t force is faster easement negotiations, quicker native vegetation approvals, or state planning departments moving at the pace the retirement schedule demands. The Clean Energy Regulator’s latest data on committed renewable capacity shows the pipeline is there. The Clean Energy Regulator’s quarterly carbon market report backs that up on the generation side. It’s the connection side still playing catch-up.
So, are we going too fast? #
Not in aggregate, no. But in specific REZs, on specific transmission corridors, closures are arriving before their replacements can plug in, and that mismatch is where the real reliability risk sits, not in some grand abstract “renewables can’t do the job” argument that doesn’t survive contact with the AEMO data. The fix isn’t slowing coal exits back down to protect ageing plants that don’t want to run anyway. It’s treating transmission delivery with the same political urgency as the retirement dates themselves, which right now, it plainly isn’t getting.
Next time someone tells you the grid’s fine because the generation numbers add up, ask them what date the relevant REZ actually finishes construction. That’s usually where the conversation gets honest.
– Marcus Wren, Editor