top of page

Thumb on the Scale

Updated: Apr 30



  • This blog post was first published on 9 March, 2024.

  • On 26 March, WePlanet Australia published a response to that blog post by energy commentator Simon Holmes à Court: "Balancing the Scale: An open discussion on the cost of new nuclear". You can find that guest post here.

  • This blog post was temporarily removed on 31 March.

  • The current, updated version was published on 30 April.


Towards the end of 2023 the pioneering US Small Modular Reactor company NuScale Power abandoned its first nuclear power station project in Idaho. In Australia, this was apparently viewed as too good of an opportunity to waste.

Just one of many such headlines.

First, clarifications.

This article is, explicitly, a public response to the GenCost team at the CSIRO, regarding their presentation of nuclear technology capital costs.

It revises and improves upon the original from a month ago, however the entire intent remains the same: to inform lay readers of the GenCost team’s error.

It isn’t an analysis of LCOE or other present value calculations. It isn’t unfounded criticism. It isn't about a peripheral issue which can be responsibly explained away.

GenCost numbers and charts are released with the well-earned authority of the CSIRO and get held aloft by various stakeholders to confirm their biases and claims. The importance of either getting the inputs right on nuclear technology, or excluding it altogether, was comprehensively addressed in prior submissions. Since then, unfortunately, the inconsistent treatment of nuclear technology has instead worsened, and this article will explain.

Next, NuScale.

The Australian Parliament’s Standing Committee on the Environment and Energy held its Inquiry into the prerequisites for nuclear energy in Australia in 2019. When the CCO of NuScale Power submitted clarification to this inquiry that his company’s rigorous cost estimate, at the time, for the SMR design they were working to deploy was AUD$5,248 per kilowatt (kW), at the prevailing exchange rate, he CC’d the letter to both AEMO and the CSIRO.

Does this mean GenCost should have used $5,248 per kW? No.

But $5,248 per kW was an overnight cost figure, and this was still important in 2023, in the GenCost consultation draft:

“Overnight capital/construction costs” …for all of these technologies, anyway.

Why was it important? Recall the definition supplied by Dominion Engineering for Overnight Construction Cost/Overnight Capital Cost:

Overnight Construction Cost (OCC) – OCC is the cost of construction if all costs were incurred at once. In addition to the base costs, this OCC includes contingency costs (i.e., additional or unexpected costs during construction) and owner’s costs. Owner’s costs are the costs borne by the owner (exclusive of financing) including land, permitting, operator training, and taxes.

As most people know, financing involves paying back borrowed money, over time, plus interest – somewhat like when they choose a car loan over paying cash. Of course, power stations are never paid for and built “overnight”, but the OCC measures costs without the complexities of timeframes and interest rates.

So, before it fell through in 2023, the 6-module, 462 MW Carbon Free Power Project had a reported price tag of USD $9,336,000,000 (page 193 of this city council document). This figure divided by 462,000 kW gives USD $20,208 per kW. In 2023 AUD, that's almost thirty thousand per kW, at the “prevailing exchange rate” of 0.7 used by GenCost (however, GenCost sets the cost at AUD $31,100 (see below)).

This is from the same GenCost consultation draft document.

GenCost shouldn’t use this “$31,000 per kW” figure either, because it comes directly from the Gross Total Project Cost (page 186 of this city council document), the total price for a project with an economic lifespan of forty years. So it was a shame to see that they did use it, and even graphed it alongside actual OCC figures to illustrate a special "explainer article", which was shared heavily on social media.

Why did the CSIRO put the Total Project Cost on this chart? No really, why? (This chart has been annotated. Scroll right to see the original without annotation.)

That leaves us (and anyone seeking to cite the Carbon Free Power Project for nuclear technology costs) asking, "then what was the Overnight Capital Cost?"

This exact cost isn’t public, but, for now, with the right resources it can be estimated. Fortunately, such work was recently performed by an MIT graduate, peer reviewed and published in Renewable and Sustainable Energy Reviews.

Stewart and Shirvan used the US Economic Energy Data Base (EEDB) to estimate the 2018 USD values for the "direct" and "indirect" first-of-a-kind overnight costs of NuScale Power’s basic design. One of the outputs of their follow-up sensitivity analysis was the overnight capital cost of a half-size power plant with up-rated modules, similar to the CFPP, and noted by the authors as reflective of changes made by the vendor “to meet the needs of a particular customer”. The result is a combined direct and indirect cost estimate, adjusted for inflation, equivalent to AUD$13,159 per KW.

We’re not quite there yet, though. To pin the Overnight Capital Cost down, we’ll consult one more source, a 2020 report from the OECD-NEA on controlling nuclear power plant construction costs. Direct and indirect costs are the lion’s share of OCC, but remember, it includes owner’s and contingency costs as well.

Adapted from Sections 2.2.1-2.2.2 (pp 36-39) of Unlocking Reductions in the Construction Costs of Nuclear: A Practical Guide for Stakeholders, OECD-NEA 2020.

The MIT paper indicates that contingency costs are included in their direct cost estimations. So, adding owner’s costs to their 6-module cost estimate, in line with the OECD-NEA’s guidance, yields an estimated OCC, in 2023 AUD, of $15,481 per kW for NuScale’s 6-module first-of-a-kind plant.

Finally: nuclear energy for Australia after 2030

Now, is this the figure GenCost should use? No.

This OCC estimate isn’t for GenCost, it’s for you.

GenCost’s projections have nuclear technology capital costs falling, by 2037, to a minimum value of 35% of 2023’s starting cost. (Long before then, Overnight Capital Costs for projects in Canada and beyond will become available.)

So if the chart below started somewhere within plus or minus 50% of our $15,481 per kW OCC estimate, it’s worth pondering what costs could be in the 2030s and beyond, post-inflation and with the benefit of learning rates to “Nth-of-a-kind cost” levels (we don’t have access to their modelling).

GenCost’s cost projection chart, annotated with this article’s OCC estimate as a ±50% range (green).

Maybe they’ll be quite high. Maybe high enough to still make nuclear technology the most expensive option in GenCost’s analysis. But maybe they won’t be so high – and mistakes like this Total Project Cost error can’t happen again.


  • AEMO commissioned GHD for a set of overnight capital cost estimates in 2018, including a fictional non-light water SMR (page 88), which was applied by GenCost to generic SMRs

  • The CSIRO excludes analysis of conventional large reactors; an explanation was recently offered here (from 46 minutes 40 seconds)

  • The “prevailing exchange rate” is set at 0.7 USD, consistent with GenCost (page 16)

  • Genuinely curious readers can find this article’s OCC calculations as a GoogleSheet.

Further reading:


bottom of page