Seven sites. Eleven gigawatts. Operational by the early 2040s. Those are the headline numbers from the Coalition’s nuclear policy, which was taken to the 2025 federal election and lost — but which Peter Dutton’s successor has since indicated the party intends to prosecute as a long-term alternative energy vision. The policy didn’t win government, but the argument hasn’t gone away, and it deserves a fair hearing on the arithmetic rather than on the vibes of either side.
Let me be direct about what this piece is and isn’t. It’s not a referendum on whether Australia should build nuclear. That debate will run for years regardless of what I write on a Monday morning in Perth. What I want to do is work through the cost and timeline claims specifically, because both have been badly served by advocates and critics alike.
Agree on the units first #
The short version is: the cost estimates for the Coalition’s plan range so widely that the range itself tells you something. The opposition’s own pre-election modelling, prepared by consultancy Frontier Economics, put the capital cost across the seven-site programme at somewhere north of $300 billion in total system terms when you account for the supporting infrastructure, the grid reconfiguration, and the opportunity cost of retaining ageing coal assets as bridge capacity. Critics, including CSIRO’s GenCost analysis — which the Australian Energy Market Operator draws on for its integrated system planning — have consistently placed the levelised cost of electricity from new nuclear well above that of utility-scale wind and solar firmed with storage.
Supporters counter, not entirely without reason, that levelised cost comparisons are sensitive to the discount rate and the assumed capacity factor, and that comparing a dispatchable baseload technology against variable renewables without accounting for full system integration costs is an apples-and-oranges exercise. Fair enough. AEMO’s own Integrated System Plan does attempt whole-of-system costing, and its current trajectory doesn’t include nuclear — partly because the lead times take any new plant outside the planning horizon where modelling has much reliability.
That lead-time problem is, in my view, the more serious of the two objections. Cost is genuinely contested and technology costs do move. Timeline is less elastic.
What the international record actually shows #
The Coalition’s plan leaned heavily on small modular reactors for some sites and conventional large-scale pressurised water reactors for others. The SMR argument is that a newer, factory-built design avoids the bespoke civil engineering cost blowouts that plagued Hinkley Point C in the United Kingdom and Vogtle in the United States. That’s a reasonable hypothesis. It remains a hypothesis. As of mid-2026, no SMR has been commissioned and connected to a grid at commercial scale anywhere in the world, though projects in Canada, the UK, and the US are at various stages of licensing and construction approval.
The conventional reactor record is not encouraging on timelines either. Vogtle Units 3 and 4 in Georgia — the most recent large reactors to enter service in a Western regulatory environment — took roughly sixteen years from licence application to commercial operation and came in at more than double their original budget. Hinkley Point C, still under construction when I’m writing this, has seen its cost estimate revise upward repeatedly from the original figure EDF quoted when the contract was signed in 2016.
The Coalition’s policy said the first Australian reactor could be online by 2035. That claim attracted immediate scepticism from engineers and energy economists, including from people with no ideological stake in the outcome. The CSIRO has flagged that even in an optimistic scenario, a first plant before the early 2040s would require regulatory frameworks, skilled workforce pipelines, and supply chains that don’t currently exist in Australia, and building all of that in parallel is a project-management ambition without a clear precedent in comparable jurisdictions.
The regulatory gap is not a small thing #
Australia has no nuclear power regulatory regime applicable to commercial reactors. The Australian Radiation Protection and Nuclear Safety Act 1998 actively prohibits nuclear power plants. Repealing that prohibition, establishing a regulator with the technical capacity to licence and inspect commercial reactors, and working through state-level planning approvals would, on any realistic reading, consume most of a decade before a sod is turned. The Coalition acknowledged this and said a government would move quickly on the legislative framework. Moving quickly in Commonwealth and state planning law is a relative concept.
I’ve watched this kind of sequencing argument play out before — not with nuclear, but with the early offshore wind licensing regime that took years longer than the initial government timetable suggested it would. The regulatory machinery for genuinely novel infrastructure in Australia is not fast, and pretending otherwise doesn’t serve anyone’s planning. If you want to understand how that firming and dispatchable capacity debate plays out in the meantime, the tension between pumped hydro and big batteries is worth reading alongside this.
What the gap years actually mean for the grid #
Here is the part of the nuclear debate that gets less attention than it should. Whatever your view on nuclear’s long-run merits, the NEM’s coal fleet is retiring. Eraring closed its final units in 2025. Liddell went before that. The remaining capacity in New South Wales and Queensland is ageing and, in several cases, operating beyond original design life. AEMO’s operational planning is built around a grid that needs firming capacity this decade, not in 2040.
The implicit logic of the Coalition’s position — that committing to nuclear would reduce investment in renewables and storage in the interim — created a coherence problem the party never fully resolved publicly. If the first reactor is fifteen or twenty years away and coal plants are exiting in five, the gap has to be filled by something. Gas? More batteries? Demand response? Interconnectors? None of those options were ruled out in the policy, but the sequencing was never spelled out with the clarity the question deserved.
Labor’s position — large-scale renewables backed by storage and transmission — has its own cost and integration questions, and I’m not pretending otherwise. The broader transition picture is genuinely complex. But at least the timeline arithmetic is internally consistent: wind and solar can be built in two to four years; the NEM needs capacity in two to four years.
The workforce and supply chain argument #
One part of the nuclear case I think deserves more charitable treatment than it usually gets is the long-run workforce argument. The Coalition and its advisors were right that once built, nuclear provides long-duration, high-skill employment in regional areas — the kind of economic anchor that a coalfield town losing its mine doesn’t get from a solar farm with a handful of permanent operations staff. That’s a real consideration. It doesn’t resolve the timeline and cost questions, but it’s not nothing, and dismissing it entirely is lazy analysis.
The counterpoint is that building the workforce for nuclear from scratch in Australia would itself take fifteen years of university programmes, apprenticeships, and regulatory licensing — which loops us back to the same sequencing problem.
What a serious cost-benefit process would look like #
The Coalition, before the election, proposed an independent authority to assess the business case for each site. That’s actually the right instinct, even if the timeline framing around it was optimistic. A genuine, independent, technically credible assessment of each proposed site — using the Department of Energy’s infrastructure frameworks and AEMO’s system planning data — would produce something more useful than the duelling press releases we’ve had so far. The AER has the analytical muscle for parts of this. AEMO has the system modelling capability. What’s been missing is a terms-of-reference that both sides of the debate could accept as genuinely independent.
I’ll admit I’m not holding my breath. The nuclear debate in Australia has become as much a cultural and political identifier as an energy question, which makes the technical work harder to hear. I noticed the same thing last year watching a panel discussion at the Australian Energy Week conference in Melbourne — half the room had made up their minds before the first slide, in both directions.
The question worth keeping open is simpler than the politics: if the cost projections for firmed renewables continue to fall and SMRs reach commercial operation somewhere in the world in the next five years with costs near the lower end of current estimates, does the calculus shift? Probably yes, at the margin. The trajectory of Australia’s energy consumption will have a lot to say about how much additional dispatchable capacity the system actually needs by the time any reactor could realistically come online.
In the meantime, the argument will continue. My suggestion is to agree on the units before the argument, and to be honest about which numbers are settled and which are projections with wide error bars. On the Coalition’s nuclear plan, quite a few of them are the latter.
— Tom Fitzgerald, Baseload & Fuels Correspondent
Photo by Vilmantas Bekesius on Unsplash