Yep - there's a cost to build the infrastructure - the one, two or seven physical power plants and THEN there's the additional cost of ramping up a nuclear power industry (Australia has NONE) to run those plants, safely for the next 80+ years. The headline figures are, I suspect the cost to build.
Life-cycle costs are almost always factored into Nuclear Power's up front capital costs, that's a fundamental difference that's rarely discussed or distinguished. However, you are right that in terms of total logistics, that may not be correctly factored in.
Conversely, I can immediately point out this analysis is severely flawed because it does not at all factor in proven cost savings from series builds. *Conservative* models of nuclear construction show 15% savings reactor-over-reactor in series builds. This doesn't factor that in.
In fact a lot of their figures don't make sense. Even at a 5% industrial inflation rate (which would be high) Vogtle wouldn't cost 36.8 Billion in 2025 USD. It would cost 35.0. And industrial inflation was lower under Biden. It should be about $34 Billion.
Vogtle Unit 4 saved 30% on costs over Unit 3, ($131 per MWh versus $187 per MWh LCOE) and IF we had capitalized on that immediately, we could have started a third reactor for $95 per MWh according to Bechtel Corp and the US Department of Energy. That's an extremely high savings rate for NOAK builds.
Scientific studies have shown that series builds usually save about 15% reactor-over-reactor. Which means that if reactor 1 cost as much as Vogtle 3 ($29.3 billion AUD), a seventh unit as discussed in the article would be $7 billion $11.1 billion AUD.
Obviously inflation increases that number, but that's true of the whole economy, not nuclear specifically. However an overnight cost of $70 per MWh is extremely realistic and in-line with what we see in countries with successful series build programs (usually ~$56-87 per MWh).
Also I really want to know what their amortization rate and discount rate is for their study. It's not specified. This sounds like the whole study is extremely biased.