Pull up the AEMO Integrated System Plan and the first thing you notice is that almost nobody who quotes it in a press release has read past the executive summary. I say this having sat through two separate industry briefings this year where a chart from the document was shown, confidently, upside down in its interpretation. Not the numbers – the framing. That matters more than it sounds, because the ISP is not a prediction. It’s a plan built on a scenario, and which scenario you pick changes the story substantially.
The document itself, published by AEMO roughly every two years with updates in between, models the lowest-cost way to keep the National Electricity Market reliable as coal retires and demand shifts. It is, at its core, an optimisation exercise across four candidate development paths, with the “Step Change” scenario usually doing the heavy lifting in public commentary because it’s the one closest to current policy settings. That’s a reasonable thing to lean on. But treating Step Change as a forecast rather than a planning assumption is where a lot of the public debate goes sideways.
Four scenarios, one headline #
AEMO runs Progressive Change, Step Change, Green Energy Exports and Hydrogen Superpower as parallel futures, each with different assumptions about industrial electrification, export demand and the pace of the consumer energy resources rollout: rooftop solar, home batteries, EVs. Step Change assumes Australia meets its legislated emissions commitments without a dramatic export-driven demand surge. It is, by AEMO’s own description, the scenario consistent with current and announced government policy, not the most ambitious or the most conservative of the four.
When somebody tells you “the ISP says we need $20 billion in new transmission” or similar, ask which scenario they’re quoting. I’ve made the mistake myself of letting that distinction blur in conversation, and a colleague pulled me up on it at a conference dinner in Melbourne last year, rightly. Let’s be careful with that number, because the transmission capital estimate under Hydrogen Superpower is materially different from Progressive Change, and conflating them in a soundbite does nobody any favours.
What the plan actually prioritises #
Strip away the scenario modelling and the ISP’s practical output is an “optimal development path”: a sequenced list of transmission projects, renewable energy zones and storage additions that AEMO’s modelling finds cheapest overall, weighed against reliability standards set by the AEMC. The REZ framework gets a lot of attention in our coverage of Renewable Energy Zones: the plan behind the pushback, and for good reason: the ISP is essentially the document that says which REZs get built and when, at least in AEMO’s preferred sequence. New England, Central-West Orana, the various Queensland zones: their priority ordering comes substantially from this planning process, even though state governments and private developers make the final calls.
Storage gets a similar treatment. The ISP’s capacity numbers for batteries and pumped hydro are frequently quoted as targets, when they’re really modelled outcomes of a least-cost pathway given certain assumptions about gas prices, coal retirement dates and technology costs. Anyone who has watched the actual pipeline (Snowy 2.0’s long slide past its original budget, Borumba’s staged approvals in Queensland) knows that modelled dates and delivered dates are different animals. We’ve written at length about where Borumba Pumped Hydro: where Queensland’s 2GW bet actually stands, and the gap between the ISP’s assumed timeline and the construction reality there is instructive rather than damning.
The capacity versus energy distinction AEMO insists on #
One thing the ISP does well, and one thing lost whenever it gets summarised for a general audience, is keeping power and energy separate. A gigawatt of new battery capacity sounds impressive in a headline. But a battery rated at 1GW might only deliver that output for two hours before it’s drained, meaning its energy contribution, the MWh that actually keeps the lights on through an evening peak, is a fraction of what the headline figure implies. The ISP’s reliability modelling works in MWh terms precisely because capacity alone doesn’t answer the question of whether the grid stays up at 7pm in July.
Worked example, roughly: if the NEM needs an extra 500MW of firm capacity to cover an evening peak lasting four hours, that’s 2,000MWh of energy that has to come from somewhere – batteries, pumped hydro, gas peakers, or imports via interconnectors. A 500MW battery with two hours’ storage covers the first two hours and then needs backup for the rest. This is exactly the kind of distinction we’ve tried to unpack in Pumped hydro versus batteries: what actually firms the grid, and it’s the same reason gas peakers keep featuring in the ISP’s near-term reliability gap analysis even as the long-term trajectory points away from gas.
Where the plan gets tested against reality #
The ISP’s optimal development path assumes transmission gets built roughly on schedule. That assumption has not held up well across the current build-out: community opposition, landholder negotiations and supply chain constraints on specialised transformers and conductor have all pushed timelines out, something AEMO itself has acknowledged in subsequent Draft and Final ISP updates where committed project dates slip relative to the prior edition. Our piece on Who really pays for the transmission build-out goes into the cost allocation fights that are a direct downstream consequence of this.
Coal retirement timing is the other place where the plan and the paddock diverge. The ISP models retirement dates based partly on generator notices of closure and partly on economic modelling of when ageing plant becomes uneconomic to keep running. Eraring is the case study everyone reaches for, and Eraring power station: how 2025 became 2027 and why covers how a plant’s actual exit date moved against an earlier retirement assumption once reliability modelling and commercial negotiation caught up with each other. The ISP isn’t wrong to model these things; it’s just that a model built on notices of intention has to be revised when intentions change.
A measured scepticism about the headline capital figure #
Here’s my actual gripe, and it’s a mild one. Every time a new ISP or its updates drops, the headline becomes a single capital expenditure figure – tens of billions in new transmission and generation investment required by a certain year. That figure is real, in the sense that it’s what the modelling produces under its stated assumptions. But it gets reported as though AEMO has handed down a bill, rather than published the cheapest pathway AEMO’s engineers could find given the inputs they were given by government and industry. If the inputs change – slower EV uptake, faster REZ delivery, a different gas price trajectory – the figure moves. Treating it as fixed makes for a cleaner headline and a worse public conversation.
I’ll admit I used to take the headline number at face value myself, back when I was newer to this beat. A few years of watching projects slip and the Draft ISP revise its own numbers changed that. It’s not that AEMO gets it wrong: the 2024 Integrated System Plan documentation is genuinely rigorous about its assumptions and sensitivities if you read the technical appendices. It’s that the public conversation strips out every caveat AEMO included on purpose.
The consumer energy resources wildcard #
One variable the ISP has had to revise upward repeatedly across successive editions is the pace of rooftop solar and battery uptake. Each update has had to acknowledge that households are adopting distributed energy resources faster than the previous plan assumed, which changes minimum demand profiles in the middle of the day and shifts where the grid’s actual stress points sit. We’ve covered the sharp end of that in Rooftop solar grid stability: the honest read, and it’s a decent example of how a planning document has to keep chasing a moving target rather than setting one.
The AEMC’s reliability standard sits underneath all of this as the actual legal benchmark AEMO is modelling against: the expected unserved energy criterion that defines how much risk of blackout is considered acceptable. The ISP exists to find the cheapest way of meeting that standard, not to produce a wishlist of preferred projects. That’s a narrower, more technical document than the one usually described in Senate estimates or on cable news, and it’s worth reading it as what it is.
What to actually take from the next edition #
When the next ISP update lands, the useful exercise isn’t to find the biggest number in the executive summary. It’s to check which scenario produced it, compare the committed and anticipated project lists against what’s actually under construction, and watch whether the storage mix assumption has shifted toward longer-duration options given how the two-hour battery fleet has grown relative to pumped hydro. None of that makes for a punchy headline. It does make for an accurate one.
I still ride out most Sunday mornings along one of the REZ corridors near a transmission easement not far from home, mostly for the quiet, partly out of habit from this job. You end up noticing the actual steel going into the ground rather than the modelled version of it. The gap between the two is smaller than people assume, most of the time. Not always.
– Anjali Rao, Grid & Storage Correspondent
Photo by Matthew Henry on Unsplash