A conversation on nuclear power, the bill nobody should pretend away, and why Australia should set a date anyway
Nick Hossack in conversation with Tony Davis
Nick Hossack sat down with Tony Davis, Principal at Perseus Management and co-host of Australia Inc., to put a question to him that usually gets a slogan for an answer: should Australia lift its nuclear ban? Davis says yes. He also says nuclear is expensive and slow, that its waste has no answer yet, and that nothing nuclear will keep the lights on in the 2030s. His case rests on something harder to price: whether a country can set a long-term goal and hold to it across governments.
NICK: Tony, before we start, you have an interest to declare.
TONY: I do. Rolls-Royce is my old company, and I was proud to be part of it. When I was a senior executive there, it was building a civil nuclear business, and I had colleagues in France whose control systems were running reactors across that country. I’m still a shareholder.
NICK: So, the headline. Should Australia lift the ban?
TONY: Yes. It costs nothing, and the current position doesn’t hold together. More than two-thirds of Australians support nuclear-powered submarines, and we’ve committed about four point six billion dollars over ten years to expand the reactor factory in Derby, run by Rolls-Royce, that will supply the reactors for our own AUKUS submarines. Meanwhile it’s illegal to build a power reactor on land. Lifting the ban lets engineers and investors price an option we’ve forbidden anyone to build.
NICK: How big a job is lifting it?
TONY: It’s bigger than it sounds. Two federal acts, the radiation protection act and the environment act, bar the approval and licensing of nuclear power plants, and states including New South Wales, Victoria and Queensland have bans of their own. The Coalition’s plan would repeal the federal prohibitions and open the government’s underwriting scheme to nuclear alongside coal and gas, but it needs the Senate, which means Labor or the Greens. The first obstacle is political, and it comes before any engineering.
NICK: The critics’ strongest point is cost.
TONY: On the evidence they’re right about where things stand today. CSIRO’s final GenCost report this July has solar and wind, with storage and firming, as the least-cost route to net zero at 141 to 152 dollars a megawatt hour, and finds no new technology below a hundred dollars by 2050. By then it puts large reactors slightly above gas with carbon capture and small modular reactors as the most expensive of all. It also says a first-of-a-kind premium will likely apply on our first plant and suggests 120 per cent for a large reactor and 92 for a small one, halved for the second and gone by the third. CSIRO cites research putting the global average nuclear cost overrun at 120 per cent. Overruns aren’t confined to nuclear. Snowy 2.0, the pumped hydro project, has gone from $2 billion in 2017 to a target cost of $22.2 billion, more than eleven times the original, with a $25 billion scenario and completion now in late 2029. Nobody should sell nuclear as a saving. The argument is for an option, with its overrun risk priced up front.
NICK: And the small modular reactors everyone is excited about?
TONY: They’re early. The best-known Western project, NuScale’s in Utah, was cancelled in 2023. The live test is Canada’s Darlington, four 300-megawatt units at about twenty-one billion Canadian dollars. Rolls-Royce signed a contract in April to start design work and order components for three units at Wylfa in Wales, aiming for power in the mid-2030s, with a final investment decision still to come. In the Czech Republic it has an early works contract for the first of up to six units, and that one is engineering only, with no equipment supply. None of these designs is running yet.
NICK: What about the waste?
TONY: It’s not trivial, and we don’t have an answer yet. Australia has no high-level waste facility and no chosen site. The AUKUS treaty makes us responsible for the submarines’ spent fuel, and the government says disposal won’t be needed until the 2050s. Our siting record is poor even for low-level waste: the national facility has failed more than once, with Muckaty in the Northern Territory dropped in 2014 and Napandee, chosen in 2021, dropped in 2023. AUKUS doesn’t solve it for power either. A submarine core is sealed and fuelled once, so that is one batch per boat. A power reactor is typically refuelled about every eighteen months, so a civil fleet produces spent fuel continuously. Finland picked its deep repository site in 2000, started building in 2004, and this week became the first country to grant an operating licence for one, though it must still pass commissioning checks before it buries any fuel.
NICK: Given all that, why set a goal at all?
TONY: Because energy security is a real issue and we sit on the raw material. Geoscience Australia puts our gas reserves at about sixteen years at current production, most of it exported, and conventional reserves keep falling because we produce more than we find. Coal lasts longer, but AEMO says around fifteen gigawatts of coal and gas will retire over the next decade, and it has data centres rising from five to thirty-four terawatt hours by 2035-36, about thirteen per cent of grid demand. Against that, we hold more than a quarter of the world’s identified uranium, we’re the fourth largest producer, and we have no power reactor of our own. The OECD’s nuclear agency says closed fuel cycles with recycling could stretch the world’s uranium supply to thousands of years.
NICK: That sounds like a moonshot argument.
TONY: It is. In 1962 Kennedy said America would go to the Moon in a decade, “not because they are easy, but because they are hard.” They landed in 1969. Apollo had one agency and a blank cheque, which we can’t copy. What carried it across three administrations was a goal with a date, and that’s what we’re missing.
NICK: It’s still ten to fifteen years away at best. What keeps the lights on until then?
TONY: An integrated system, and engineering, not religion. AEMO’s plan has almost fifty gigawatts of grid-scale storage and hydro and seventeen gigawatts of flexible gas by 2050, with coal running until 2049. I’d sweat the coal we have, build the gas and storage, and keep high-efficiency coal on the table if it can win on equal terms. AEMO’s August outlook sees no reliability gaps before 2030, so the case for nuclear is a post-2030 case. With lead times of ten to fifteen years, the decision has to come now.
NICK: When could we actually have power?
TONY: The IAEA’s benchmark is ten to fifteen years from first considering nuclear power to the first plant, and the UAE did it in twelve, starting in 2008. Decide in 2028 and that puts first power somewhere between 2038 and 2043. A proven design won’t spare us the first-plant premium either, because CSIRO applies it to our first plant however well the design has done overseas.
NICK: And who pays for the first plant?
TONY: Nobody has said. Britain has put public money behind its program, with 2.6 billion pounds allocated in its 2025 spending review and a National Wealth Fund loan of up to 599 million pounds on commercial terms. The Coalition’s answer is underwriting on the same terms as coal and gas. Nobody invests in a sixty-year asset the next government can cancel, so a bipartisan deal has to come first.
NICK: Buy, hold or sell — on nuclear?
TONY: Buy the option, with conditions: lift the ban, set a date, build the bridge with everything we’ve got, and be honest about the bill. Anyone selling nuclear as the answer to the 2030s, or refusing even to price it, I’d ask for the evidence.