Zero. That’s how many commercial nuclear reactors are under construction in Australia right now, and how many have ever operated for grid power in this country. Compare that to roughly 40 gigawatts of wind, solar and storage capacity either built or contracted through state schemes and the Commonwealth’s own tender process, and you start to see why the nuclear debate generates so much heat relative to its actual footprint on the National Electricity Market. So does Australia need nuclear? I’ve sat with the numbers for a while now, and my answer is no, not for the reasons either camp wants to argue about.
Let’s be precise about what’s actually being proposed. The federal Coalition’s policy, detailed in its own costings release and picked apart since in outlets across the spectrum, points to seven sites including the retiring coal stations at Liddell and Mount Piper, with large-scale reactors slated for the 2030s and small modular reactors flagged as a bridge. I’ve covered the cost and timeline mechanics of that plan in detail elsewhere on this site, and nothing in the past year has shifted the fundamentals: the earliest credible generation date sits beyond 2035, and the CSIRO’s GenCost modelling, updated again this year, still has nuclear landing well above firmed renewables on a levelised cost basis even when you load in generous transmission and storage assumptions for the wind-and-solar case.
The capacity that’s missing isn’t nuclear-shaped #
Here’s the mechanical problem nobody on either side likes to sit with. AEMO’s Integrated System Plan, the document that actually governs how the NEM gets planned, models an orderly coal exit running through the early 2030s and replaces that capacity with renewables, storage, transmission and a modest gas firming fleet. I wrote about what that plan actually says in detail a while back, and the thing that jumps out is timing, not technology. The ISP’s preferred pathway needs replacement capacity landing in roughly the same window coal plants are closing. Eraring’s exit, delayed once already, and the saga around Yallourn’s own countdown both show how tight that math already runs.
A reactor commissioned in the mid-2040s doesn’t help with a coal closure in 2033. It’s not a competing technology at that point. It’s a different problem entirely, arriving a decade after the gap it was meant to fill.
Cost comparisons keep getting re-litigated, and that’s fine #
I won’t rehash the full levelised-cost argument here because I’ve made it at length before, but the short mechanical point still holds: nuclear in Australia would require building an entirely new regulatory, workforce and supply-chain capability from scratch, something countries with existing nuclear industries don’t have to do. The UK’s Hinkley Point C blew through its original budget and timeline by a wide margin, and that’s a country with an operating nuclear regulator, an existing skilled workforce and decades of reactor operating experience to draw on. Australia has none of that. The CSIRO and AEMO’s own GenCost draft, released again this year, is about as close to a neutral umpire as this argument gets, and it still has nuclear’s levelised cost sitting well above wind, solar and storage once you account for the first-of-a-kind premium.
Where I think the nuclear advocates have a fair point, and it’s a point worth conceding, is on land use and transmission. A reactor sits on roughly the same footprint as the coal plant it replaces and can often reuse existing transmission connections. Renewable Energy Zones require new poles and wires across greenfield corridors, and the fights over routes, like the paddock dispute tangled up in VNI West, are real and they’re costing time. That’s a genuine advantage for nuclear on paper. It just doesn’t close the cost and timeline gap by anywhere near enough.
What the moratorium actually does and doesn’t do #
Commonwealth and state legislation bans nuclear power generation in Australia, a position that traces back to amendments made to the Australian Radiation Protection and Nuclear Safety Act and mirrored in state environmental protection legislation through the late 1990s and 2000s. Overturning that isn’t a cabinet decision; it requires federal legislation plus matching repeals in at least the states where plants would sit, and Labor governments in Victoria, Queensland and WA have all said publicly they won’t shift. Even if a future Commonwealth government secured the numbers in both houses, getting four or five state parliaments to agree inside the same term is its own political project, separate from the engineering one.
That’s before you get to insurance, waste storage licensing, and the fact that no Australian university currently runs an undergraduate nuclear engineering program at the scale a dozen-reactor fleet would need. Small modular reactors get wheeled out as the answer to the workforce and timeline problem, and I’ve gone through the actual state of that technology in a separate piece. The plain version: there isn’t a single SMR operating commercially anywhere in the world at the cost and schedule its spruikers promise, and the handful of projects furthest along, in Canada and the US, keep slipping their own dates.
The gas comparison nuclear advocates avoid #
This is where I’ll own the contrarian bit properly. The strongest technical argument against an all-renewables NEM isn’t that it lacks baseload, a term that’s doing less work than people think now that storage duration keeps extending. It’s that firming capacity, whatever form it takes, needs to be dispatchable on short notice, and I’ve set out why gas peakers keep winning that argument against batteries on long-duration events even as batteries eat into the shorter end of the market. Nuclear doesn’t help here either. Reactors run best flat out, near their rated output, for as long as possible. They’re a poor technical match for a grid that needs capacity to ramp hard twice a day around the solar duck curve. France’s own fleet, the most-cited nuclear success story globally, has had to adapt its operating profile for exactly this reason as rooftop and utility solar spread across Europe, and that’s with sixty-plus years of institutional experience behind it.
So the honest technical case for nuclear in Australia isn’t really about firming at all. It’s about having a large, centralised, dispatchable asset that doesn’t need weather. That’s a legitimate thing to want. It’s just not what the current grid is short of. What the ISP and AEMO’s own planning documents say we’re short of is transmission, storage duration, and project delivery speed, three things a nuclear build makes worse before it makes better, given the land acquisition and approval processes alone would consume years that renewable projects are already most of the way through.
Where the public argument gets distorted #
Polling on nuclear support in Australia moves around depending on how the question is framed, and I’d treat any single number with some suspicion. What’s harder to dispute is the gap between stated support for “nuclear power” in the abstract and support for a specific reactor at a named site. Local opposition at Liddell and Mount Piper has been vocal since the sites were named, and that’s a predictable pattern repeated in every country that’s tried to build near existing communities rather than in genuinely remote locations. Remote siting solves the opposition problem and creates a transmission and workforce-housing problem instead, which is exactly the kind of trade-off CopperString and the New England REZ have had to work through for renewables projects already underway.
I’d also push back gently on the idea that this is purely an engineering argument dressed up as politics. It’s genuinely both. The Capacity Investment Scheme, which I’ve walked through in detail separately, has already locked in contracts for a large share of the next decade’s firmed capacity build. Unwinding or pausing that pipeline to make room for a nuclear fleet that won’t generate for fifteen-plus years isn’t a neutral act. It’s a bet that the thing you’re waiting for will definitely arrive, against a backlog of major infrastructure projects in this country that mostly don’t.
My actual read #
Nuclear isn’t a bad technology. France, South Korea and increasingly the UAE show it can run a grid reliably at scale. But “can it work somewhere” and “does Australia need it now” are different questions, and conflating them is where most of this debate goes wrong on both sides. The renewables lobby sometimes pretends cost curves settle the whole argument when transmission delay and firming duration are the live constraints. The nuclear lobby sometimes pretends a moratorium repeal is the hard part when workforce, insurance and thirty-year construction realities are the actual wall.
My own view, and I’ll own this as an editorial judgement rather than a neutral summary: Australia doesn’t need nuclear to solve the problem it currently has, which is a 2030s capacity gap. It might be worth keeping the door open for the 2040s and beyond as storage technology plateaus and the easy wind and solar sites get used up. But spending political capital on the moratorium fight right now, while the REZ build-out in NSW is already running behind the coal exit it’s supposed to match, looks to me like solving next decade’s maybe-problem at the cost of this decade’s certain one. That’s not a each-way bet dressed up as caution. It’s a judgement that the opportunity cost is too high, and I think the data backs it.
Anyway, reactors don’t change their mind on a five-day forecast, which is more than you can say for most of the arguments made for and against them.
– Marcus Wren, Editor
Photo by Moritz Kindler on Unsplash