Around 4,600 megawatts. That’s roughly the combined nameplate of Loy Yang A and Bayswater, the two coal stations still carrying most of AGL Energy’s generation book, and the number that every other decision this company makes gets measured against.
AGL is the biggest gentailer in the National Electricity Market by generation capacity and by customer numbers, and it’s also the company with the most coal left to retire on a clock it keeps resetting. That combination makes it the single most useful case study in the entire transition. Not because AGL is doing anything wildly different to Origin or EnergyAustralia. Because it’s doing the same thing at a bigger scale, with a bigger balance sheet exposure, and a louder set of critics watching every move.
The coal exit dates keep moving, and that’s the point #
Loy Yang A in Victoria’s Latrobe Valley is now pencilled for closure by 2035. Bayswater in the Hunter Valley is scheduled for 2033. Both dates have shifted over the past several years, mostly earlier, occasionally with caveats attached about grid reliability and replacement capacity being ready in time.
That’s not unusual. Origin pulled Eraring’s closure date forward and then had to negotiate an extension with the NSW government when replacement firming wasn’t ready, a saga covered in detail in Eraring power station: how 2025 became 2027. AGL’s dates are different numbers on the same spreadsheet. The company has been explicit in its own investor materials that these are forecast closures tied to asset condition and market signals, not contractual guarantees carved in stone.
Here’s my issue with the way this gets reported. Commentators treat an updated closure date as either a broken promise or a bold new commitment, depending on which direction it moved. It’s neither. It’s a company managing an ageing thermal fleet against a replacement pipeline that isn’t fully built yet, and being reasonably transparent about adjusting the number as conditions change. I’d rather see honest recalibration than a fixed date that gets missed anyway, the way Eraring’s did.
Kanmantoo and the shape of the new fleet #
The Kanmantoo battery in South Australia, built on the site of a former copper mine south-east of Adelaide, is a comparatively modest addition at under 100 megawatts of storage. It won’t move the needle on AGL’s overall capacity mix. What it signals is the template: AGL buying or building grid-scale storage on brownfield industrial land rather than greenfield sites, which shortens planning timelines and sidesteps some of the community pushback that newer wind and transmission projects run into.
AGL has a broader battery and renewable development pipeline across several states, including large-scale battery projects attached to its existing generation sites at Loy Yang and Liddell. The Liddell site in the Hunter Valley, where AGL’s older coal station closed in 2023, is now earmarked for battery storage and other firmed capacity, a fairly literal version of replacing coal megawatts with storage megawatts on the same grid connection point. That’s the efficient way to do it. New transmission capacity is the scarcest resource in the entire NEM right now, a point this masthead has made before in Who really pays for the transmission build-out, and reusing an existing connection is worth more than people give it credit for.
Consumer energy resources: the quieter bet #
While the generation story gets the headlines, AGL has been building out a consumer-facing business around rooftop solar, home batteries and orchestrated demand response. The company has pushed products that bundle solar, storage and electric vehicle charging under its own retail brand, betting that the household side of the ledger becomes as important as the wholesale side over the next decade.
This is the sensible read of where margins are heading. As rooftop solar capacity keeps climbing, past the nineteen gigawatts discussed in Energy Supply Australia’s rooftop solar coverage, the wholesale market gets squeezed in the middle of the day and the value shifts to whoever controls flexible load and storage behind the meter. AGL’s consumer energy push, and the broader policy architecture around it covered in The Consumer Energy Resources roadmap: what it means on your roof, is the company trying to own a slice of that shifted value rather than watching it go to third-party aggregators and virtual power plant operators. Whether AGL can out-compete nimbler VPP operators on that front is a genuinely open question, one I don’t think the market has settled yet.
What the AER and AEMO data actually show #
AEMO’s generation data and the AER’s wholesale market reporting both show coal’s capacity factor sliding gradually across the NEM, not because the plants are being shut down overnight but because rooftop solar and utility-scale renewables are pushing thermal generation out of the middle of the day, forcing it into a shoulder role it wasn’t designed for. That’s expensive wear and tear on old boilers, and it’s part of why Loy Yang A and Bayswater’s economics look worse each year even without a closure date attached.
AGL’s own investor disclosures acknowledge this cycling pressure explicitly. It’s the same mechanical story playing out at EnergyAustralia’s Yallourn, discussed in EnergyAustralia: the gentailer counting down to Yallourn’s exit, and it’s a big part of why none of the major coal exit dates across the NEM feel especially stable from one annual report to the next.
Where AGL sits against Origin #
I’ve written before, in AGL versus Origin: two gentailer strategies compared, that the two companies are running slightly different plays with the same hand of cards. Origin has leant harder into gas peaking and an LNG-linked upstream position through its Australia Pacific LNG stake. AGL has leant harder into batteries and firming attached to its own retiring coal sites, plus a bigger consumer energy resources push.
Neither bet is obviously wrong. But AGL carries more coal-closure risk on its own balance sheet, simply because it owns more coal capacity outright. Origin’s exposure to Eraring is one station; AGL is managing two large stations in two different states with two different regulatory environments and two different sets of community and union politics attached. That’s a heavier lift, and it’s reflected in how cautiously AGL has talked about firm closure commitments compared to some of its peers.
The firming gap nobody’s fully closed #
Batteries are brilliant at smoothing short, sharp swings. They are not a like-for-like replacement for the sustained output a coal unit can hold for days during a prolonged heatwave or a cold snap with low wind. That gap, between battery duration and the multi-day firming job coal currently does, is the genuine unresolved question in AGL’s transition plan and in every other gentailer’s plan too.
AGL’s answer so far leans on a mix of batteries, demand response, its consumer energy fleet, and continued gas where needed, roughly the menu laid out in Gas peakers versus big batteries: the firming maths in 2026. Whether that menu scales fast enough to cover both Loy Yang A and Bayswater’s retirement on the current dates is the single biggest question an AGL shareholder should be asking at the next annual general meeting. AEMO’s Integrated System Plan modelling, published on AEMO’s website, assumes a build-out pace for storage and transmission that the industry itself has repeatedly flagged as ambitious.
The honest scorecard #
None of this makes AGL a villain or a hero. It makes it a very large company trying to run two businesses at once, a shrinking thermal generation fleet and a growing distributed energy platform, while the regulatory and market settings underneath both keep shifting. The Clean Energy Regulator’s annual data, available through its published reports, will keep tracking how fast the emissions intensity of AGL’s generation portfolio actually falls, and that’s the number worth watching more closely than any press release.
My own read, and I’ll admit it’s a minority one among some of the louder transition commentary, is that AGL’s incremental recalibration of closure dates is more honest than the market wants to admit. A company that nails a closure date to the wall five years out and then has to walk it back under reliability pressure, the way Eraring’s owner did, hasn’t actually served anyone better than one that keeps saying “here’s our current best estimate” and updating it as the batteries and transmission actually get built.
Whether AGL’s consumer energy bet pays off before the coal fleet runs out of road is the question I’d want answered at the next results briefing. Right now it reads like a company hedging its own uncertainty rather than hiding it, which, in this sector, counts as a point in its favour.
– Marcus Wren, Editor
Photo by Alexandru Boicu on Unsplash