turns out you can just do things

every solar farm in the uk ever:

1/ On increased costs for increased penetration of renewables: Jack is right, the variability of renewables is a problem, and it does become more difficult as their penetration increases. However, nuclear doesn’t coexist well in this environment, and the renewable buildout shows no signs of slowing down. A grid with high renewable penetration has very low prices at peak output. This means the nuclear operators of the future will need to think about how they plan to manage paying back their asset when, 30-40% of the time, the spot price is below even the rosiest production cost estimates (this is just a more extreme version of the NYISO example. I take the point that NYISO isn’t the most representative of the UK grid today, but it does give some rough colour to the expected grid dynamics in 10 years time). To survive in an energy landscape like this you need a generator that can provide response to changes in renewable output/demand and provide readily dispatchable capacity. Getting flex out of nuclear will make it even more expensive. Other solutions are likely going to be better than this across most reasonable time horizons (definitely gas but also hopefully some combination storage and synthetic fuels in the future when @harveyhodd and @CJHandmer succeed in their mission). Nuclear has a slightly better outlook if you don’t expect renewables to get built… However, to use the China example Jack mentioned: they may be constructing 32.5GW of Nuclear right now, but they commissioned 217GW of solar in 2023 alone (even at a baseload equivalent, the solar build-out dwarfs nuclear). This trend is broadly uniform across the globe. Prices are going to be very low, a lot of the time.

2/ On the cost of design vs cost of regulation: Agreed that regulation is broken, and accounts for a large proportion of the cost associated with design. However, even with a friendly or non-existent regulator, the costs we associate with regulation will not disappear, assuming the designer is doing their job properly. Nuscale spent ~$500 million on Engineering to support the design certification application (per EFI) including testing, engineering and simulations. Were the NRC applying ridiculous levels of scrutiny causing price inflation? Undoubtedly, but these costs shouldn't entirely disappear without a regulator, and if they do that’s also a problem. At the end of the day, even if the regulator works optimally, design and regulation will always be interwoven to a problematic extent in nuclear. Nuclear may have very small risk probabilities, but risk impacts can be catastrophic. Regulation will always hamper design speeds because of this (which is not necessarily a bad thing overall).

4/ On funding: The numbers above focused on funding R&D costs (and, as Jack mentions, the associated regulation costs of design). PPAs, CfDs and regulated returns are de-risking mechanisms to get financing to build a plant. They guarantee revenue, but they don’t fund the development phase. Equity and debt are still the only instruments meaningfully at work here providing capital to startups.

3/ On the scaling of regulation, insurance and decommissioning: Regulation can be improved in terms of how it scales in the UK, even more so in the US where the NRC use flat fees ($5.3m per year regardless of size)🤯. However, halving the size of a plant design does not halve the complexity, nor does it halve the risks. They reduce, but on a £/MW basis this is going to be a challenge. The site is also critical. Each site has its own unique risks for infrastructure (we learned this from trying to build solar anywhere possible). You cannot cut and paste a site from one location to the next. Flood zones, stability risks, and environmental hazards will always need to be designed for. One generic design approval does not mean a broad rollout is immediately possible (hence the need for a GDA, a site licence and an EIA. This is not a terrible framework). This same principle stands for insurance and decommissioning. Although for the latter a standardised process for a number of SMRs on a single site could certainly help.

current focus: digging holes

improve your life expectancy by having an in-house engineering team - great way to avoid 6 months of arguing over a 1,000-page EPC contract

Convenient truncation of the x axis but the point still stands

price → 0. value → ∞ :( oh no

Worth accounting for Opex. If opex was 0, plants in the US with paid off capital costs wouldn't require subsidies to stay alive.