Ep116: Tom Samson "Small Modular Reactors: Nuclear's Next Act?"

Michael Liebreich So, Tom, thank you so much for joining us here today on Cleaning Up.

Tom Samson It's a pleasure, Michael, great to see you again.

ML So, what we always do is, get started by you describing what you do, in your own words, because I'm just gonna get it wrong, and mangle things. So, tell me exactly what you do.

TS Well, I'm the Chief Executive Officer of Rolls Royce SMR Limited, which is a company that's been set up by Rolls Royce, and it's now, as well as Rolls Royce, it's got three other shareholders. So, I run that business for those shareholders, and I lead a team that is now about 600+ people across the UK, that is now designing the technology that's going through the application process in the UK, to bring our new Rolls Royce factory-built SMR to market.

ML Right, now, our audience, they're all smart, you can assume they will know about climate change, and most of them know a lot about energy. But they are not nuclear; they've also been listening to episodes on hydrogen, where SMR stands for something completely different. So, let's actually make sure we don't lose anybody with the acronyms; I'm gonna bring you up short when you use acronyms, I'm afraid. So, SMR, in the nuclear context, means what?

TS Small Modular Reactor. And in the context of nuclear, SMR is quite a broad church. There's everything from a few megawatts, and we're probably the largest SMR in the market, we're getting over 470 megawatts electric power from our unit. So, we're at the high end of the scale, but it covers a whole range of technologies that are defined as SMR because they are significantly smaller than the gigawatt programs have been. But it does include a range of different fuel types, different megawatt capacities and different delivery models.

ML So, we had Julia Pyke on the show, and she was talking about the gigawatt scale. So, those are the big Pressurized Water Reactors, nuclear as it currently exists on the electricity system. So, when you talk about 470 megawatts electric, that is just under half the size of a big nuclear power station as we know it?

TS Actually, it's more like, no, a third of the size, or a quarter the size, of the units that are being built at Hinkley. The Hinkley units are about 1600 megawatts, so they're large gigawatt. AP1000 units from Westinghouse are about 1200 megawatts. So, we're a fraction of the scale of the current gigawatt fleet of reactors, which tend to be between 1200 and 1600 megawatts. So, ours is small by comparison, but relatively large compared to the wider, broader church of other SMR technologies.

ML How small does the SMR category go down to? Because there are some very small nuclear reactors that are being discussed. But those, at some point, there's a division between SMRs and micro-reactors, correct?

TS Yeah, you'll get down, when you start getting down to deployable on a truck or deployable in a more mobile environment, I think they would be categorized as micro-reactors, in the kind of one to two megawatt scale, but then everything in between that and where we are is probably the broad church of SMR reactors, that tend to be more land-based, probably. Maybe, that's the distinction; I'm sure the micro-reactors are more agile and mobile than say an SMR is, which tend to be grounded and fixed location.

ML Because Rolls Royce also has a design, does it not, for something that's around the five megawatt mark that might go on board ships or to be used in very agile, potentially military situations? Is that right?

TS Yeah, well, my colleagues in our defense business have been making a much smaller reactor for space applications, that could also [be used] in defense applications, it could also have a crossover to commercial applications. And that is a much, much smaller range of five megawatt capacity.

ML The background here is to create something, then, let's call it the SMR, it is not those very small, not space, not military, not shipping; it is for land-based. And usually SMRs are proposed to produce electricity. There's a few proposals to use them for heat and district heating in China, I believe, but you're really talking about electrical output, mainly, correct?

TS Well, actually, I think it does go broader than that, Michael; I think we absolutely are looking at on-grid applications, where you plug the SMR into the grid, and you make traditional electricity. But I think there is a broader opportunity set, looking at off-grid applications, where SMRs can be used to, be dedicated to powering data centres, for hydrogen production, for other energy intensive usage applications as well. So, I think that that is a broad categorization that should consider SMRs as addressing both traditional grid power, but there are a growing number of off-grid applications that are very exciting as well.

ML Okay, but even those off-grid, you just mentioned two which we'll come back to, data centres and hydrogen. Those would both be primarily electrical output that you'd be using?

TS Primarily, although we've been doing a number of studies on hydrogen production at that scale, which would include [a] combination of heat - thermal heat - as well as electricity to get higher performance, hydrogen production from say, a solid oxide electrolyzer. So, it doesn't have to just be electricity, we can combine it with heat, and then [inaudible] as well, there is opportunities for us to use the back-end heat for district heating applications as well, which is not something we've done extensively in the UK, for example, but in many of the Central European applications, that is a huge driving force for the energy transition; is what solutions can be readily adapted to provide both electricity, but also district heating solutions.

ML We talked about this on the first episode we did on nuclear, which was with Kirsty Gogan, very early on in Cleaning Up's history. We talked about how some people would think, wait a minute, district heating with nuclear, what could possibly go wrong? Having steam from a nuclear power station going into your home... But of course, it's all completely... it's a secondary fluid-loop, this is not radioactive steam in any way, shape, or form and never could be, right?

TS Correct, correct. I mean, it's way further, further down the steam extraction chain, and indeed, the district heating schemes would be entirely independent and isolated networks that would sit outside of the nuclear power plant, and just transfer heat through an exchanger.

ML So, there's all these different designs of Small Modular Reactors, SMRs. And there's been this real kind of pre-Cambrian explosion of different types of designs, different types of fuels. There's something like 50, or 60 different designs a few years ago, when I last looked in detail. What are the main design choices? Because I want to get on to your particular design, but I just want to give our audience some sense of the range, because they've got different trade-offs on all sorts of different dimensions. So, what are the different groups of SMRs that you see out there? And then we'll come on to, why have you chosen the design that you are pursuing it Rolls Royce?

TS I think you can probably put it into two categories. There's some Gen Three Plus, let's call it, which is the evolution of existing technologies, and we would probably put ourself in that category. And then I think you can call the Gen Four technologies, which are more innovative nuclear reactions, more innovative nuclear fuels, that either don't exist today or haven't been commercially exploited at scale. And so, I would look at the SMR landscape in those two categories, and we are very much in the existing, proven technology end of the spectrum, where we've actually genuinely, by design and by intention, not chosen to create a different type of nuclear reactor or nuclear fuel. We're using pressurized water, or light-water reactor technology with standard enriched uranium, for the reason that it allows us to bring a solution to market much more quickly.

ML Are all the Gen Three, are they all pressurized water, or are there some Gen Three that take other conventional or existing technologies forward?

TS Yeah... Not to spend too long talking about my competitor technologies, Michael, if you'll forgive me for that, but GE Hitachi have got a Boiling Water Reactor design, so that's a different existing technology, but a different type of reactor to a Pressurized Water Reactor. NuScale has a Pressurized Water Reactor, but they've integrated it with a number of other components, such as the Steam Generator, into an integrated reactor design, so they have a different technology, and risk profile. But I think GE Hitachi and NuScale are using existing Gen Three Plus technologies; I think the newer design that EDF are exploring, I think is a traditional PW, Pressurized Water Reactor with standard fuel. So, I think those are the more near-term deployable SMR solutions. And then you have other companies that are developing solutions that are using more innovative fuels like X Energy, TerraPower, and so they are more into the Gen Four categories.

ML Those would be Molten salt (MSR), and then you've got the traveling-wave reactors (TWR), very innovative. But presumably, the more innovative you are, the harder it is to get through the whole process of certification?

TS The process of certification is much more risky and complicated and will take longer, but more importantly, many of these other designs will require a new form of architecture and infrastructure in-country, in terms of fuel handling, fuel disposal, management of fuel, management of waste, management of the various aspects that are needed to accommodate technologies that don't exist today. And that's one of the reasons why we've chosen to work with standard enriched uranium, a standard Pressurized Water Reactors, because it's very well known; it's already existing within the architectures that exists today, and therefore, it's much more easy for customers to adopt that solution without having to put in place all the investments to be able to accommodate a new design and a new type of fuel.

ML So, what's hopefully emerging then for the audience is, understanding of where you've chosen to play, which is about a third of the size of these very big plants. So, it's 470 megawatts versus more than a gigawatt, and it's using, at least on the surface, conventional technology; pressurized water, it's conventional fuel, and you're taking advantage of all the supply chains, presumably for components as well as for fuel? So, if that's a correct characterization of where we've got to so far, then my question is, why does this transform the economics? Because surely, you make things cheap by making them big in nuclear; you've gone the other way, so you're going conventional, but small. Why is that a better solution economically?

TS Well, I think what we're demonstrating is, you make things lower cost by building them on a much smaller scale. But what we've done is not just made a small, stick-build, build-it-in-a-field, small reactor; we've actually extensively used modularization to transform how a nuclear power plant can be built. And so, we build it in a factory, not in a field, and that's the big differentiator between what we're bringing to market with our Rolls Royce SMR, is we're using not only our nuclear capabilities, which we've got in Rolls Royce from our submarine and defense business, which we've had for six decades; we're combining it with our production, manufacturing expertise, and how we can make highly complex, heavily-engineered products in a factory environment, so that you are creating a product, not a project. And so, that kind of productionization, commercialization, commoditization of a product is what allows us then to drive down the costs, improve on the schedule, and allow for the inbuilt learning curve to flow from unit to unit within a factory environment. And that's really the major transformational aspect to help us both reduce the schedule; and by reducing schedule, you better manage cost. And by using proven technology - and in fact, effectively commercially off-the-shelf products, such as our turbine generator island, which is a commercially available product we can buy from many, many companies in the world today - we lower the innovative risk that we're introducing by bringing nuclear technology forward. And we proudly proclaim not to have the world's biggest, or the world's first, in anything that we're bringing into the market.

ML And for the audience that are not familiar, we should probably do a quick recap of what Rolls Royce does, and also importantly, what it doesn't do. This is not - if anybody out there is listening and hoping that we're going to be talking about those lovely cars - it's not that Rolls Royce. That I think is part of BMW if I'm not wrong? But this... which businesses are you going to draw on to bring that manufacturing, technology, and know-how out of?

TS So, Rolls Royce today operates in five business sections: we have our civil aerospace business, and we deliver a large portion of the wide-bodied aircraft engine market as Rolls Royce, and you'll see many Rolls Royce logos on the sides of large, wide-body airplanes if you fly long haul; we have a defense business, both in the US and in the UK and Europe, that provides defense products, including the reactor technology for our submarine fleet in the UK, which is where our nuclear heritage originates from within Rolls Royce; we also have in an Friedrichshafen in Germany, a power systems business that makes diesel generator sets, that power of variety of applications including land-based and marine applications; and then we have two new businesses that are focused on the energy transition, electric flight, and we have an electrical business that is providing electric propulsion systems, and we've actually flown the world's fastest electric aircraft in 2021; and then on top of that, we now have Rolls Royce SMR, which is a separate company but still majority-owned by Rolls Royce PLC. But we also, we are part of that Rolls Royce family. You're absolutely correct, the Rolls Royce motorcar has the license to the brand which is hugely powerful, but those cars are made by another company, under a separate business and nothing to do with Rolls Royce PLC any longer, apart from the brand.

ML Right. So, the aerospace and the other energy businesses, that gives you the heritage to be able to manage a supply chain of components, but also ultimately, presumably, to manage service once you've got some of these out there with customers?

TS Yeah, I mean, our contracting model will be quite different than the ones we use in those other businesses, and we're looking to deliver a turnkey nuclear power plant, that's what we're bringing to the market, which I think is quite a big differentiator; we have done all the integration, we're providing a single contract to our customers, that will then allow us to build them a fully integrated nuclear power plant that we then hand over, and they would then operate that over the life of the asset, which will be at least 60 years. So, that's the model that we've developed, but you're absolutely right: the manufacturing capabilities, the production line dynamics, the nuclear engineering, manufacturing, and materials experience has all been built into the IP, as we've created Rolls Royce SMR, that gives us the ability to do this technology forward. And indeed, we're in the midst of that regulatory approval process right now with the UK regulators, to take this design forward for that UK certification.

ML I want to push on this question of modularity, because it's really at the heart of why you think this is going to work, and it's going to get sufficiently cheap to be economically viable, where I certainly have referred to the gigawatt scale plants as having been tested to economic destruction; I may be being ever so slightly unfair, but I'm not sure. But the modularity is really the secret sauce here, but as I've understood modularity for SMRs, I always thought that it was kind of, you had multiple of them to get up to gigawatt scale. But your version of modularity is that you want to make what's really quite a big nuclear power station, 470 megawatts - I mean, it's not really small at all, it's mid-sized, I would call it - but you want to make it in modules. So, that's what your version of modularity means, correct?

TS Yes, I mean, you referred to there something that's very similar to the NuScale model, where they have 50, 60, 70 megawatt units that are all lined up to get to a gigawatt, or however many megawatts you want to get, by putting multiple small reactors side-by-side.

ML Or 200, or 150 megawatts - these are the normal sizes that get described as SMRs.

TS Right. So, what we've done, and the reason why we've reached this capacity, is we've actually tried to drive as much product into the factory as possible, to enable us then to maximize that modularization benefit, and to enable us to utilize the factory environment to do as much of the work as possible. And indeed, what that's constrained by, is road transportability. So, you can do the work in the factory, and you can transport the modules, or the biggest components, then directly to site to be assembled. So, you're reducing the need to stick-build, as they call it in the industry, which basically means connecting everything together in a site, because it's too big to transport to the site. By making something that's big enough to take advantage of road transportation limits, you can maximize the activity in the factory, transport everything as you can to the site, and then the site focuses then an assembly of modular products that have been maximized in terms of the production in the factory space. So, that's what we've brought to market, is that heavily modularized solution, where we try to minimize the need to do work on site, in a field, in the open air, by maximizing what can be built into modules that are then delivered to site for final assembly; and products, the biggest single vessels that we have, which are then defined by road transportation limits. That's how we actually arrive at our 470 megawatts, Michael, that's the biggest reactor diameter we can transport by road, and therefore that's as much fuel as we can get in, to transport it by road, is determined by that length. And that fuel in that Pressurized Water Reactor creates about 1350 megawatts of thermal energy, that converts into about 470 megawatts of electrical. So, it's really bounded by road transportability constraints and maximizing then the work that's performed inside a factory.

ML  So, you're gonna have a lot of containers - if I've understood it right - a lot of containers. But then each of them can go on the back of a truck, each of them delivered by road, and then they get assembled. How many containers are we talking about?

TS  So, we're still in the design stage of the numbers of modules that will be produced in our module factories, but it's going to be in the range of 1500 to 2000 modules; and that includes everything from modular structures, host stairwells, and supporting structures. We're trying to drive modularization to create supporting structures for the actual building, so the building is actually the structures themselves. And so, that's roughly the range, depending on the actual detailed design of how many modules we'll need to be delivered for a single unit.

ML So your reactor vessel, which is the kind of one of the most critical parts, that's going to be made all off site, put on the back of a truck, and it arrives, right? And that's very different from the gigawatt scale, where there's lots of welding that has to go on on site, correct?

TS Correct. And the same for other components as well.

ML And the same for other components. What about the actual container? I mean, there's still an enormous amount of construction to make the site ready: the concrete, the base plates, the services, and then the all-important pressure... the containment structures, they still have to be built on-site, do they not?

TS So, we're maximizing how those aspects of the design are modularized as well, so we're working, for example, we've done a lot of work with Laing O'Rourke, who's our civil partner; they have a civil module factory in the UK already, where they pre-manufacture civil modules, so we adopt those into our design. So, we minimize the need for concrete pouring on site by exploiting those civil module solutions. But the containment, for example, is a steel containment that will be produced on-site in a number of steel plates, that will be then welded together and delivered to site as modules. So, that's how we overcome the containment piece. But every part of our design, we look at it to determine, can modularization improve cost, schedule and performance by being adopted? And that's how we've looked at maximizing modularization on that basis across the plant.

ML So, two thousand containers, gets assembled on site. If we have any children listening, I've got the perfect description of what you're doing: it's a Lego nuclear reactor. It's not really small, modular because of that, it's modular because you're going to build it from pieces on site.

TS Absolutely, absolutely. There was an article in this weekend's Financial Times that talked about modular construction build, and used the Lego analogy in terms of what can be possible with Small Modular Reactors, as well. So, that's a great analogy, Michael, that's exactly what this is: it's a Lego kit of parts that comes to site, and is connected and assembled on site.

ML Darn it, there's me thinking I was really original, and the FT has got there first. Still, I got to publication first, I'm recording today, not... Anyway, that's a really good description. But, now what it leaves of course, is the really difficult questions about, how much cheaper will this be than the gigawatt scale power stations? Because, they are, I've called it tested to economic destruction; we've got Hinkley, which cost in today's money, about £120 per megawatt hour, $140 / $150 per megawatt hour, the resulting electricity. We've got Sizewell, we're supposed to believe is going to be so much cheaper, and we had Julia Pyke on the show. I challenged her; she wouldn't answer, she couldn't answer because she's negotiating, but you know, I'm thinking £80 to £90 pounds, maybe $100 / $110 per megawatt hour is probably believable. What price point are you going to deliver power at?

TS Well, I think... I'm trying, I'll try and give you as complete an answer as I can, Michael, given that we're also hoping to move into negotiations on a number of fronts, in terms of delivery of these units. But what I would say is - and Julia mentioned this in her podcast, as well - that cost of capital does play an important part. So, for example, whether we could access a Regulated Asset Base model, using the Nuclear Financing [Act] for SMRs, would affect the cost of capital. If we didn't access that, and we had to use a CfD, and we were then looking at a weighted average cost of capital, of say 9%, then the numbers would be higher.

ML We need to explain CfD: so, CfD would be a Contract for Difference, where the only way that, not you, but your customer gets remunerated, is by selling electricity, or other services. But you'd have to finance the whole build period without any revenues; unlike the Regulated Asset Base model, where you get some payments throughout, during the construction period. So, just for those listeners who are not completely au fait with all of the acronyms.

TS And there's two benefits to the Regulated Asset Base model, as you say, you can achieve a return on your equity and your debt during the build phase, which for gigawatt is even longer, but it's still a number of years for SMR. So, that reduces the need to fund interest during construction. But also, depending on the structure of the Regulated Asset Base, there's a role to be played for us as a contractor to take risks, there's a role to be played for equity to take risks, and there's a role to be played for the customer to share some risk. And so, depending on how that risk allocation is structured, and depending on how much government support there is behind that Regulated Asset Base model, the costs of capital in a Regulated Asset Base model could be significantly lower than a Contract for Differences model, which would traditionally rely upon project finance type debt and equity.

ML OK, so give us the range though. If you say, right, you get a really good Regulated Asset Base model at one end, and you have to go for a pure Contract for Difference, CfD base, at the other end, what is the range of power price that you're going to be able to deliver?

TS So, I think in the Regulated Asset Base space, we're going to be between £40 and £50 per megawatt hour, and I think in the Contract for Differences space, we're probably going to be between £60 and £70. Probably around £70 for the first units, and as we get down that learning curve, subsequent units then coming in around the £60 or below mark, as we come down that cost curve. And so, that's the benefit of the learning curve being applied unit to unit as we build a fleet of these things, for a global deployment that will allow us then to drive down the costs.

ML Would that be for your first unit?

TS Yeah, so the range covers, somewhere... The first units will be at the upper end of that range, and then second, third, and fourth units would start to come down, and we would then drive towards the bottom end of that range or even lower, once we got through nth of a kind deployment, and we've started to deploy multiple units. There is other differences as well, Michael, that's worth highlighting, such as: if we're building these things in other jurisdictions, in other countries, then cost of labour, wherever the factories are located, could have an influence. So, there's a number of factors, but that gives you the sense. That, importantly, is the premise on which this technology has been developed; it's been developed as a low-cost solution, and we've now got the underpinning and the evidence to back up those costs, that give us the confidence to enter into contracts, hopefully in the near term, to deliver these units on the market.

ML I'd be very interested to know, if you can share, what is that evidence? Because the question is, why should anyone believe you, when nuclear has got just such a desperately poor, frankly, track record? There's a tremendous academic at Oxford called Bent Flyvbjerg, and he's just publishing a book, he's done so before, and he looks at different projects and how much their cost overruns are typically. And nuclear, the number is 120%. I mean, the only thing worse for cost control, than nuclear, frankly, is the Olympics; where each Olympic city says oh, it's going to be cheap and marvelous and ends up being expensive and marvelous. Why should anybody believe your numbers?

TS Well, I think if we were proposing to build the same type of solutions that were the basis for that evidence, then that would be a fair challenge, Michael. But the design that we've developed, the purpose behind our design, is to come up with a much lower cost, and more deliverable with greater certainty, solution. And that's why the factory build concept, that's why our modularization concept, that's why sizing this to maximize the use of commercially available products in the market, helps us to then not only deliver a lower cost solution with greater certainty, but build up that evidence base as we go forward, to provide that certainty. And I don't underplay the complexity and the risks involved in doing this, and there needs to be appropriate contingencies made; but again, these are much more digestible units, in the kind of two, two and a half billion pounds range for these first units. That's a different proposition than taking that kind of risk on a £25 plus billion program. So, our ability to demonstrate control and manage risk on a much smaller scale is what gives us the ability to have that confidence. And I can I can tell you, we've been scrutinized by the UK Government, as we've gone through and secured our grant; we've been scrutinized by investors who've put equity into this business, in November 2021 alongside Rolls Royce; we've been scrutinized as we access the UKRI grant funding; and we'll be scrutinized, again, by our shareholders before we sign these contracts. But we have got an evidence base and we're working with the supply chain, we're working with partners, to solidify and improve that evidence base over the coming months.

ML But, you know, the scrutiny is great. But of course, Hinckley was scrutinized, Vogtle in the US was scrutinized, Flamanville was scrutinized, Olkiluoto was scrutinized, Taishan was scrutinized. And manufacturing things in factories is no guarantee against cost overrun; setting up a new factory... even our good friend, Elon Musk setting up his factories, it was very lucky that he had access to infinite free capital at the time, because he was always enormously late and enormously over cost as well.

ML I think the challenges are real, and that's why the analysis, the engineering, the confidence has to be demonstrated with contracts, with supply chain evidence, with parties who are prepared to share the risk with us in delivering these products. So, all those things are things that we're doing today, to enable us to bring this program to market. We're actually taking responsibility for an integrated design: that isn't something that has happened in all gigawatt programs, and isn't happening in all SMR programs. So, taking accountability and responsibility for that integrated approach, managing the interfaces, being the design authority, and controlling the design decisions across the whole plant, specifying the numbering system, specifying the kit of parts that will be used inside the modules: these are all elements that will massively de-risk delivery, and provide greater price certainty to enable us to move forward. Again, not without risk, not without execution risk, not without supply chain risk, but we have to then make sure we have the strength of those underpinnings, and the evidence to give us the confidence to come forward and enter into these contracts.

ML So Tom, you've given range of electricity costs between £40 and £70 pounds per megawatt hour, depending to a certain extent on how its funded and regulated. But if I was to say, right now, I will buy one of these, but you - I want you, Rolls Royce SMR - to take the risk of any cost overruns that push the price above £70 per megawatt hour. So, you take that risk. I'll put the money down, I'm good for it, but I want you to take the risk of any overrun above £70. Would you sign that contract today?

TS I think we would, Michael. Yeah, absolutely. That's why we're in this business, that's what our job is, is to deliver that cost certainty, and deliver those projects. Now, we wouldn't necessarily do that project for a fixed price, lump sum number; we may have a fixed target, we may have a target element, we may have contingency, and we try and do that as clearly and as transparently as possible, all of which would build up to those numbers that I referred to. So, I don't think we would be uncomfortable with taking that risk on board, we've just got to convince our shareholders of that strategy, and we've got to demonstrate to them the confidence, as I mentioned, that underpins that. But one thing I would say to you, Michael, is there aren't any other contractors that we're aware of in the SMR world that are offering that turnkey solution. So no, regardless of how much risk is being taken - and I've said to you we would take risks. Now, what that risk looks like, and the detail, is obviously a matter of negotiation - but we're bringing to market an integrated solution, and a turnkey solution, to deliver a fully integrated nuclear power plant. That's a big differentiator in this space. And that in itself, is a demonstration of our commitment to taking on that risk.

ML I think the question is not about the turnkey solution, so much as the risk of the costs not being £70, but being £80, being £90; the start date being so late, which drives up the cost of electricity as well, those sorts of things. And do you have sufficient balance sheet to stand behind that, because I'm not aware that you do, by the way?

TS Good Rolls Royce SMR doesn't, so we're gonna have to manage that in conjunction with our shareholders and with the first customers to try and make sure we define something that's going to work for all parties. And I don't want to disclose, or get into the nitty gritty details of what that might look like, Michael, but I can tell you, we wouldn't be coming to the table offering an integrated turnkey solution unless we were prepared to have that open and transparent conversation on how that risk is allocated. Otherwise, we're not a distinct and different solution in the marketplace.

ML Let's talk a little bit about the role of these SMR, the Rolls Royce SMRs, so the 470 megawatt SMR. Let's talk about the role of that in the energy system. Because even at £40, £50, £60, £70 per megawatt hour: that is higher than the cost of other sources of clean electricity, which are perhaps less problematic from the perspective of fuel proliferation, waste management... there's just fewer headaches with some of the other solutions, which produce electricity more cheaply. So, why would anybody use your solution?

TS I think, and I'll talk in general terms about nuclear, because I think that's the challenge that you're asking, and I think it has to be a low-cost deliverable nuclear solution to be in that marketplace. But what nuclear offers is quite different. The quality of nuclear power is quite unique in the clean energy space. Now, you talk about other alternative sources of energy; they're not often compared on a like-for-like basis with what is equivalent firm power that's providing available, dependable, always on, clean power 24/7. That in itself is a valuable element to any balanced energy mix. And so, in fact, the greater the dependency on intermittent renewables, the greater the reliance there will be on having a form of clean, dependable, available power. So, I think that there has to be that factor built into it. I think renewable marginal costs of additional turbines, on additional wind farms, those costs have come down dramatically over the last 15 years, and rightly so, to prove the value of that type of solution. And we expect nuclear costs when we start to roll out fleets of SMRs to come down as well in price, as we build up that supply chain, and we transfer that knowledge and experience in the factory environment. So, I think that we have to make sure we're comparing apples with apples.

ML Well, Tom, we have to do one other thing. I agree with that, but we have to also do one other thing, which is: the nuclear has to play nicely with the renewables, right? Because when you talk about dispatchable, and people use the word baseload and so on; the problem with that, is that the very cheap renewables, they out-compete nuclear when it's windy when it's sunny, and the nuclear gets kind of shifted to have to load-follow. And as soon as it does that, your capacity factor falls, and your costs go up. Because the £70 - if it's that - is on the basis of it working 365, 24 hours a day, or as near as possible; so, 90%, 95% of that. So, how do you integrate what you're doing into a grid, which we all know, is going to be dominated by, frankly, cheaper, but variable renewables?

TS So, I agree, it's a great point, Michael. And it's a perennial debate we have within our industry, in terms of the energy transition, is how do you get the balance right? And I think that is the key debate and the key topic. I think that what I would say is, on the comparison, and getting the balance, right: we've got to think of this number - and I use it a lot in my thinking - 8760. 8760 is the number of hours in a year. We've got to have generation available 8760 hours a year, and that's essential for customers, for stability; it's essential to ensure that we've got a grid that we can rely upon. And I think that's the first thing we have to think through. Nuclear power, though, will operate for 18 to 21 months between refueling outages, at that level of availability, as you say, maybe 90%, 95% with the fuel outages, which gives you 8760 a year and in an outage year, slightly less than that. But that is dependable power. I think we need to recognize that as we have more and more renewables, the intermittency impacts make it more and more important that we've got a clean source of power that is available 8760.... I'll come to your point about the balancing aspects, in terms of interruptions and changes in demand profile; and I don't subscribe to the baseload-is-dead philosophy; I think there will always be an element of base load - you may say it will be higher or lower than it's been historically - but I think there will always be a chunk of power that has to be met 8760 hours of the year. But, to your point about balance, nuclear can be used, as I say, for a range of other mechanisms. And whilst there might be times when the grid would like the nuclear asset to fluctuate, you could choose to fluctuate the power in an SMR. It may not be the most cost effective thing to do, and a more cost effective way to deal with that interruption in demand might be to divert that power to produce hydrogen, ammonia, synthetic fuels or other products that can be then produced in a much more dispatchable way in conjunction with an SMR. But, I do think that we're kind of using the wrong lens to look at that challenge, because we are going to need baseload. I mean, I think the UK...

ML You say we're going to need baseload, but the fact is, we're not, because we're going to have a huge amount of wind.... Let me just be very clear, because: baseload, the word comes from the demand side. It says, there's some minimum level of demand always, and then it switches to the supply side and says, and therefore there are some large, cheap plants that will run 24/7, never get switched off, and they're cheap; that's where it comes from. But the fact is... there is no plant that is going to be expected to run 8760 hours a year, because what that would mean is switching off vast amounts of, in the UK's case, wind, in other countries, solar, for many, many hours a year, unless they happen to coincide with a peak. Now, the fact is, that is proven low-cost electricity that we know... and at the margin, by the way, it's got zero marginal costs, no fuel at all, unlike yours. So... you're going to be squeezed down to 1000 hours, 500 hours, unless you can find a better role in the energy system.

TS I don't agree, Michael, with that analogy.

ML Are you saying, Tom... Let me ask you this: are you saying that unless you can sell all your electricity as close to 8760 hours per year, you cannot be built? Is that what you're saying? Because I think that's a problem.

TS Well, if you only generate power for 60% of that time, then the cost of that power has to be increased accordingly to reflect the fact you're not using it 100% of the time. You're right: renewable wind, renewable energy can operate for 30% of the time, and still be economical at that point - £50, £60 a megawatt hour - because it's being used whenever it's being produced. And so, that allows it to operate on an economical basis with only 30% or 40% or 50% availability. Now, my question to you, Michael, is: is there a direct correlation between that demand and supply on a second by second basis, today? There isn't. So, do you have a full match between the demand for clean energy, which is 24/7, and the supply of clean energy, which is 24/7? That doesn't happen today.

ML So, first of all, let me answer that: no, of course it doesn't match, it needs to be matched. There are a number of ways of matching it, right? That's why, you know...

TS What do you do to match it?

ML So, first of all, there's demand response, and we'll have a lot more of that; we should come back to that because in the context of hydrogen, ammonia...

TS Hydrogen and ammonia require energy to produce, so you've already created a bigger problem in trying to generate those products to begin with.

ML Hold on, I'm talking about matching; what do you use to match? So, what we'll be using is demand response; we'll be using interconnections; we'll be using overcapacity; we'll be using various forms of storage but what we won't be doing... Because here's the thing that I'm pushing on: if your logic is that there are times when there isn't wind, and therefore we need nuclear; but therefore, we're going to run it 365 days a year, and therefore we don't need the wind; I can tell you, you will sell no power stations that way, because the fact is we've got the wind already, very successful and cheap, and we're going to have a load more of it, because it is currently being sold and built. So, then...

TS But Michael that doesn't address the fact... Look, let's take 2022, right? Solar and wind in the UK accounted for 78 terawatt hours of electricity production, right? And that's from an installed renewable base of, say, 30 gigawatts, okay? The total energy consumed, electricity consumed, in 2022, was about 256 terawatt hours. So, that's 30% of the UK's energy - not megawatts, not capacity on windy days - but over the 12 months, about 30% of our energy production came from those renewable sources. Which means that 70% of the energy came from gas, coal, biomass, nuclear, hydro, storage, interconnectors.

ML Mainly gas and nuclear.

TS So, how are you going to replace that 70%, even if you build more wind capacity, which is only available for those hours...

ML Where I'm going with this is... And by the way, that 70 terawatt hours is going to increase enormously as we build out the existing, the already committed offshore wind that we've committed, right? And by the way, I won't go into things like Xlinks, where I'm, for disclosure, I'm an investor. But you know, Xlinks can bring in renewable power - dispatchable, same stuff as you sell - for £48, without all of the brain ache of nuclear; different brain ache, to do with security and so on, but not the nuclear brain ache. But what there's a premium on, is the dispatchable, the flexible power, that can fill in the times when there isn't any wind, right? But you're not offering that. So, instead of offering a flexible solution - which you can offer, by the way, by pairing it with industrial output, and so on - you're trying to offer an inflexible solution, and I think that's a problem.

TS No, no, I think I've explained that we do offer a flexible solution, by how you can combine multiple SMRs to produce a combination of grid electricity, produce hydrogen, produce ammonia, produce synthetic fuels, in a variety of forms.

ML But you need to be temporally flexible; not flexible in different uses, but flexible in the sense of when there's no wind, meet the need, but when there is wind, do something else. That's flexibility, that's what the grid needs.

TS Well, the grid still needs power to cover 8760 hours of the year. Now, I'm not saying that's the only use that we provide, but it's definitely a role that nuclear can play in a balanced energy mix. And it's maybe not a convenient argument from a renewable perspective, but again, we're not against renewable: I think there should be as much renewable built in this country as possible, and hopefully that 30% will get up to 50%; it will never get to 100%, even in a country like the UK.

ML Let me try this, in the interest of resolving this one. Let me propose how I see this working because I do see the role, but we're not quite there yet in this conversation. So, I see nuclear's role, if it's going to have one, would be largely to power big industrial processes. So, for instance, your electrolysis, your metals processing, desalination if you're in an area that needs desalination, and so on. And then the periods when the renewables - which are very cheap and lovely - but then they fall off the system, what happens is, the electricity price on the grid soars; it goes up into hundreds and hundreds of dollars per megawatt hour. And at that point, what you want to be doing, is shutting down - you and your customers - shutting down those industrial processes, and selling all your electricity to the grid. And if it's linked with hydrogen, then what you use to get through those windless or sunless times is some hydrogen or a derivative out of storage; and you serving the grid, which enormously reduces the storage requirements, and therefore is potentially highly economically attractive on the system's basis. But it only works if your customers can shut down. Because if all you can do is run 365, I think you become part of the problem, not part of the solution.

TS Well, I think we can see in those markets, where nuclear is operating today, a massive demand for that 24/7 clean energy; people are extending lives of those nuclear assets; people are bringing nuclear assets back online, to access that 24/7 clean energy. And I think there's a huge amount of opportunity and flexibility...

ML But Tom, what we also see is nuclear power stations, particularly in the US, having to be subsidized. Because quite frankly, it's cheaper now to add renewables, and to shut your nuclear unless it gets some kind of support mechanism.

TS That's not entirely fair, because [in] most of those markets, it's the marginal cost of gas that is determining the market conditions, that make it uneconomical for those nuclear plants to run. And it's the low-cost availability of shale gas in the US over the last decade that has made it harder for those nuclear assets to operate on a competitive basis. So, it's not fair to say it's purely a result of the large volumes of renewable energy in the US that's created that problem for the nuclear assets over there. My only point, Michael, is this: what you seem to be suggesting is, that many consumers of electricity need to change their consumption habits to reflect the large volumes of wind and solar energy that's going to come into this system. And I'm just challenging whether those consumers of energy are going to be prepared to do that, or whether they are indeed going to seek a 24/7 source of energy they can rely upon, that is clean, on a constant basis.

ML It's a good point. I mean, look, we're going to see enormous changes in consumption anyway, because we're going to see the electrification of transportation, with a lot of storage. And we're going to see the electrification of heat with, by the way, a lot of thermal storage, which is also hopefully going to be highly, highly useful.

TS And we should be building in more district heating schemes. So, I think the hypothesis that you build a nuclear asset that can operate 24/7, in this huge clean energy transition, but then ask it to shut down or come offline...

ML No, Tom, just to be clear, what I think it should be doing is... Frankly, it should be serving all of the different bits of, primarily industry, because that's large-scale demand, that can be subjected to demand response. So, it's not them, it's not the nuclear that shuts down, it's the industry that shuts down.

TS I think that is a harder change to implement. Because that changes the cost of their production, it changes the cost of their product, and so I think I think that is a harder sell...

ML Look, many of them shut down anyway; they shut down for maintenance, they do it when they want. What we're saying is, there's a lot more capacity to be flexible around industry. And lots of industries that have got no idea; they've never even looked at what time-shifting might entail, it just hasn't been important. Now, it becomes highly valuable. But the other thing is, long-term storage, potentially via hydrogen... That is - and you may know, I'm quite a skeptic on lots of use cases for hydrogen - that's one that I'm not skeptical about, because I do think that does allow larger volumes of time shifting in the system. And that, paired with nuclear - particularly also because your heat could be very useful in the production of hydrogen - that is potentially highly, highly promising.

TS I agree. I think the applications for nuclear to produce hydrogen, to produce synthetic fuels, to produce other forms of energy-intensive applications in this energy transition are huge. I really enjoyed your podcast with Francesco Starace: he was grossly affronted by the prospect of using this very precious hydrogen molecule for something as crude as generating power in a gas turbine. But the hydrogen production opportunity is there, in terms of how is it going to be produced, whether it's the colours of the rainbow discussion, or actually, how do you actually produce it cost competitively, I think is going to generate real opportunities. Because the demand is going to come for hydrogen.

ML And then both of us should probably defer to another Cleaning Up guest, Julio Friedman, who would say that what you're anyway going to be doing with it is using it for Direct Air Capture of carbon. I personally think that that's going to be excessively expensive, and not really going to happen that much. Similar to, by the way... I think we might disagree on the costs of synthetic fuels, which I think are, largely, not going to be a thing. But there's definitely going to be a need for a lot more clean electricity and a lot more clean heat. So, somehow this will figure itself out.

TS I agree: I think the electrification agenda over the next 20, 30 years is going to drive huge demand for clean energy. And I think that's the marketplace in which we expect our Rolls Royce SMR to flourish.

ML Tom, thank you so much for spending time with us here today, it's absolutely fascinating. Lots of questions answered, some questions not answered. But no doubt, you and I will revisit them probably face to face over the coming years. It's been a great pleasure.

ML Always an excellent and energetic debate, Michael, fantastic questions, and thank you so much for spending so much time on your podcast for another nuclear debate. Really, really worthwhile. Thank you.