Atomic Pancake Theory

The UK’s first SMR project looks to be a four-pack of AP300s. Why build four small reactors over one big AP1000? Let’s explain with a delicious metaphor

Independent developer Community Nuclear Power (CNP) recently announced the UK’s first Small Modular Reactor (SMR) project, stealing thunder from the government-driven SMR efforts that we’re still waiting on. CNP said it has targeted building four AP300 reactors on a site in Teeside, northeast England, amidst an industrial cluster that will provide local demand for not just power but heat.

The AP300 is basically just Westinghouse’ flagship AP1000, but “downsized,” with one loop instead of two. This is actually an advantage as it will share many components with big brother. But it also raises the question: why don’t CNP just build one big AP1000 instead of four small AP300s?

In fact, this is exactly what Princeton energy prof Jesse Jenkins asked: why build four small units versus one big one?

Yes, it will be more expensive to build four AP300s over a single AP1000. By how much? It’s hard to tell exactly. Westinghouse CEO Patrick Fragman gave an overnight cost estimate of $3,400/kW when the AP300 was rolled out in 2023. This means a single reactor would cost less than a billion dollars. That is both massively lower than the overnight cost of Vogtle and higher the the ‘should cost’ of the “nth of a kind” AP1000 as calculated by a team from MIT lead by Koroush Shirvan. I’ve heard the AP300 been described as a way to make “a third of the power of an AP1000 at half the cost.” The truth is in there somewhere, but we won’t know until we build it.

So let’s go back to Jenkin’s question: if it’s so hard to build one big one, why would anybody want to build four small ones instead?

To answer that question, imagine if you are making pancakes. Of course, it’s more efficient to make a few big pancakes over a bunch of small pancakes. But consider: When would you choose to make multiple small pancakes instead of one large one? I can think of a few scenarios:

(1)When you’re not completely confident: We all know about the first pancake curse. It’s always burnt or doughy and the subsequent ones come out better and better. 4 SMRs is enough to get significant learnings through serial production and mitigates first of a kind risk.

(2)When you only have a small pan: It’s asking for disaster to make a big pancake in a small pan. The pan here can stand in for any number of bottlenecks in nuclear construction from labor shortage (remember the UK still has two huge EPR builds going concurrently) to limited capital.

(3) When you could run out of gas: You might hedge against uncertainty by trying to cook a couple of pancakes quickly just in case your stove goes cold. In this case, building four small instead of one big gives the developer an off-ramp after building one or two reactors.

(4) When you need to share the pancakes: Just as one big pancake is harder to share. So the output from one big nuclear reactor harder to spread around. First of all, just managing the refueling outage would be a pain. This project aims to sell a lot of power and heat to local offtakers in the industrial cluster around North Teeside. Smaller reactors might make sense.

B.F. Randall ⚛ ⛏ ⚡

of Mining Atoms added a 5th and 6th reason: “Faster to cook” and “easier to pay for.” Indeed, it would be an advantage when signing Corporate Power Purchase Agreements for the power to be available sooner rather than later, and as we are no longer living in a zero interest rate environment, “time is money” is truer than it used to be.

Of course, to truly deliver on the dream of the SMR, something that can be factory manufactured and practically plopped on site, you can’t have something like an AP300, which has the same components as the AP1000 but fewer of them. But that dream SMR could be quite a bit further into the future. In the meantime, there are legitimate reasons to build SMRs that are basically just the same “recipe” but smaller.

CNP cites Westinghouse’s mature supply chains in their technology selection and there is also the hope that the existing licenses given to the AP1000 can transfer to the AP300 or at least pave the way. The Teeside site is expected to be fully-licensed by 2027, and commercial operations is projected to commence in the early 2030s.

Of course, it’s way too soon to tell how a project like this would fare. We are years away from a final investment decision and as of now don’t even know who are the private parties behind this project in terms of investments. And I don’t claim to have any insider information: this post is just my speculation of what could make a project like this work.

THE ELEMENTAL TAKE

I still believe that for big grids you need big reactors. Barring some kind of crazy technical innovation that takes us out of the realm of the traditional light water reactor, there’s simply no overcoming the incredible economy of scale.

But with increasing sale to private, industrial offtakers who aren’t grids, 300MW might be a much more suitable size. The North Teeside area that might be served by this 4 reactor project will be doing very energy intensive things such as making hydrogen. But even in that case, a gigawatt is just too large.

In the end, we need reactors of all shapes and sizes. Here’s to wishing CNP and Westinghouse the best of luck. But we also look forward to announcements from the UK SMR competition, where Westinghouse is also in contention alongside UK local champion Rolls Royce with their 470MWe UK SMR and the crowd favorite BWRX-300.