Drive the Yabba Creek Road out past Imbil and you hit a construction zone that doesn’t look like much yet from the public lookout: some cleared ground, a few site sheds, a access road switchbacking up the range. That’s Borumba, and it’s meant to be carrying 2,000 megawatts of pumped hydro capacity by the early 2030s. It’s also, more or less, the load-bearing wall of Queensland’s entire energy plan. If Borumba slips meaningfully, the state’s story about itself changes.
Queensland has built its energy narrative on two ideas that reinforce each other politically even when they don’t always reinforce each other technically: big public pumped hydro, and a generation fleet that stays in government hands rather than being sold off to the AGLs and Origins of the world. I’ve spent a fair bit of this year going back through the CS Energy and Stanwell filings, the Powerlink transmission plans, and AEMO’s Queensland-specific modelling in the Integrated System Plan to work out whether that combination is actually coherent, or whether it’s two separate bets dressed up as one plan.
The pumped hydro bet, in megawatts and megawatt-hours #
Let’s be careful with that number, because Borumba gets quoted a lot of different ways. The project as currently scoped is 2,000 MW of power capacity with roughly 24 hours of storage duration at full output, which puts it in a different category entirely from a battery. A big grid-scale battery today might run two to four hours before it’s drained. Borumba, on paper, can run flat out for the best part of a day. That’s the entire point of pumped hydro as a technology: it’s cheap on a per-megawatt-hour basis once built, because the “battery” is just a dam full of water, not a warehouse full of cells that need replacing in fifteen years.
I’ve written before about where Borumba’s cost and timeline actually stand, and the short update is that the project hasn’t gotten any cheaper since. Queensland Hydro, the state-owned entity building it, has pushed the completion window out and the capital cost estimate up, which is not exactly a scandal in pumped hydro circles. Snowy 2.0 did the same thing, repeatedly, and I covered that saga too, right down to the tunnel boring machine that got stuck for the better part of a year. Big underground civil works are hard to cost precisely from a desktop study. The rock tells you what it’s going to tell you once you’re actually in it.
The smaller sibling project is Kidston, up near Georgetown, where Genex Power converted an old gold mine pit into a 250 MW pumped hydro scheme. That one’s operating now, which matters more than people give it credit for, because it’s the proof of concept that the basic engineering works in Queensland conditions. I’ve gone through Kidston’s status in detail previously – it’s smaller, it’s further from load, but it’s actually spinning.
Public ownership as the other half of the argument #
Here’s where it gets more interesting than the engineering. Queensland never fully privatised its generation fleet the way NSW and Victoria did. Stanwell and CS Energy remain state-owned, sitting alongside Powerlink as the publicly owned transmission operator. The current state government has leaned into that heavily as a point of difference, arguing public ownership lets Queensland build pumped hydro and transmission at a cost of capital private developers can’t match, and without needing to hit a commercial return hurdle on year one.
There’s a real argument there and I don’t think it’s spin. A state government borrowing through Queensland Treasury Corporation pays a lower interest rate than a private developer raising project finance for a multi-billion-dollar asset with construction risk. Over a 50-plus year asset life, that spread compounds into genuinely different economics. I’ve set out the broader version of this argument in the piece comparing public and private ownership models across the NEM, and Queensland is the best live test case going around.
But cheaper capital doesn’t fix schedule risk, and it doesn’t fix the fact that a government-owned builder still has to source turbines, tunnel boring crews, transformers and skilled labour from the same global and east-coast markets everyone else is competing in. Borumba’s slippage isn’t really a public-ownership problem. It’s a big-tunnel problem that happens to be owned by the public.
What’s actually keeping the lights on right now #
Coal still does the heavy lifting in Queensland today, and it’s worth being precise about what “heavy lifting” means here. Callide and Gladstone, Stanwell and Tarong between them still supply the majority share of Queensland’s annual energy, measured in megawatt-hours generated over a year, even as their share of installed capacity gets diluted by new solar and wind entering the system. That’s an important distinction my readers on this desk have heard me make before: capacity is what a plant could produce flat out, energy is what it actually produces averaged over time, and coal’s energy share stays stubbornly high because it runs around the clock while solar runs for maybe a third of the day on a good capacity factor.
AEMO’s Integrated System Plan has Queensland coal retiring progressively through the early-to-mid 2030s, a timeline I’ve gone over in detail in the ISP breakdown. The plan assumes Borumba and the broader renewable energy zone build-out land roughly on schedule to backfill that retiring capacity. If Borumba lands a few years late, and on current trajectory I’d put real odds on that: the state needs either more gas peaking, more batteries, or a a genuinely uncomfortable conversation about coal plant life extension. None of those options are free and none of them are quick to stand up either.
CopperString and the transmission side nobody outside Queensland talks about #
The generation and storage story gets most of the headlines, but Powerlink’s transmission build is arguably doing just as much heavy lifting, structurally. CopperString 2032 is the line intended to connect the north-west minerals province, think Mount Isa and the copper and critical minerals belt around it, into the main grid for the first time at meaningful scale. I’ve covered CopperString’s progress and the stakes around it before, and it’s genuinely one of the more underrated pieces of the Queensland plan. Without it, the renewable resources in the state’s west stay stranded, and the minerals processing industry up there stays reliant on diesel and isolated gas generation indefinitely.
Powerlink’s own transmission annual planning report, published alongside its submissions to the AER’s revenue determination process, lays out a build programme that’s ambitious even by current NEM standards, where every state is trying to build transmission roughly five years faster than anyone has previously managed. Queensland isn’t unique in that pressure. It’s just carrying it over a much bigger geographic footprint than Victoria or South Australia, which makes the logistics meaningfully harder.
Where the plan is honest and where it’s optimistic #
My measured, slightly unpopular view: I think Queensland’s public-ownership framing is doing more political work than engineering work. It’s a perfectly defensible policy position, and the cost-of-capital argument holds up under scrutiny. But it gets used rhetorically to suggest that public ownership alone de-risks a 2,000 MW pumped hydro scheme or a multi-billion-dollar transmission line, and that’s not how construction risk works. A state-owned entity digging a tunnel through variable geology faces the same geotechnical surprises as a private one. Queensland Hydro’s own cost revisions make that point better than I can.
Where the plan is genuinely strong is in sequencing logic. Queensland isn’t trying to do everything at once. Kidston proved the pumped hydro concept at small scale. Borumba is the big swing. CopperString unlocks the north-west. The Capacity Investment Scheme, which I’ve covered in detail on how it actually works mechanically, is underwriting a chunk of the renewables and storage that need to come in alongside all of it. That’s a coherent architecture, even if the individual dates on the Gantt chart keep moving.
A worked comparison worth sitting with #
Take a four-hour, 500 MW battery versus Borumba’s 2,000 MW, roughly 24-hour configuration. The battery delivers 2,000 MWh of total energy across its discharge window. Borumba, at full output for its design duration, delivers something in the order of 48,000 MWh. That’s not a knock on batteries, which do a different job, responding in milliseconds and covering the sharp evening ramp brilliantly, a comparison I’ve gone through properly in the pumped hydro versus batteries piece. It’s just a reminder of scale. Queensland isn’t building Borumba to shave an evening peak. It’s building it to carry the state through days of low wind and cloud cover, the way Hornsdale’s battery in South Australia was never meant to.
I’ll admit I used to be more sceptical of the duration figure before I went through Queensland Hydro’s own technical reports properly; the civil engineering logic for going long-duration on a scheme this size is sound, even if the price tag makes your eyes water.
What I’ll be watching next #
The thing I keep an eye on isn’t the headline completion date, it’s the interim milestones: whether the main dam wall contracts get let on schedule, whether Powerlink’s connection works for Borumba stay aligned with the generation timeline, and whether CS Energy’s coal retirement dates move in response. Those three threads have to stay roughly in sync or Queensland ends up exactly where NSW has, with a REZ build-out running behind its coal exit schedule, a mismatch I’ve tracked closely on the NSW side too.
Queensland’s bet is bigger, slower, and more expensive to get wrong than most states’. It’s also, if it lands, a genuinely different model for how a state runs its grid. I don’t think the jury’s in yet. Ask me again once the Borumba tunnel crews hit rock they weren’t expecting, because on a project this size, they will.
– Anjali Rao, Grid & Storage Correspondent
Photo by American Public Power Association on Unsplash