The Exchanges

Q we hadn't thought enough about before we started chatting with you folks and, and getting involved with elemental. It's different, um, and, and distinct from project development of, of a wind or a solar project, for example, or a battery project, or even a gas or coal project. So you just give me a quick snapshot of like end to end. What does project development look like for new nuclear?

A Yeah. So we start off with, and this is the first thing we did was we went out, find, uh, A customer. Somebody who's willing to say, I want 20 to 30 years of power at this price range, and willing to put in a portion of the development capital. That was probably the first 12 months of our existence, was doing that, ah, with all the hyperscalers, just Google happens to be the first one across the line, um, on it. And then working with them saying, where do you need these assets, and working cooperatively with them and their host utility, because many times they're building data centers Where they have existing data centers, and they know the existing utilities. Um, you know, we're coming in saying, we don't want to get between you and your customer, we just want to give you what your customer wants in a way that works for you as a utility. That's a very different model from what you saw, you know, in the late stages of fossil, which I participated in, which was basically, if you build it, they will come. Um, also with renewables, where go out and get a plot of land, connect it, and then go hunt for a PPA from somebody. We've reversed that process, uh, by putting the customer and the demand first, then finding a site, working with utility, go through the red flags analysis of it, um, do it in a technology agnostic manner, uh, which you can do in the NRC process under an early sit…

AI assessment note: “So we start off with, and this is the first thing we did was”

Q some kind of announcement or another On nuclear. So, it feels like one big change there is, of course, there is load growth in general, but also you've got this class of corporate buyers that are generally, universally leaning into nuclear and interested in doing something creative with their balance sheets, right? Going above and beyond just saying, okay, I'll sign a PPA when you get the price down sufficiently.

A Yeah, you know, absolutely. The first time around, it was all based upon regulated utilities getting the rate payers to underwrite it. And, you know, at the time, they were trying to underwrite Underwrite, you know, 17,000,000,020 billion dollar projects, and their balance sheet or their market capitalization may be 20 to thirty billion dollars. And the growth didn't show up, and their expectation for high gas prices at the time didn't come to fruition. We had cheap gas, and so those two things really conspired against the demand side of it, where we have kind of the opposite, where we have, you know, pretty much all-time low gas prices, but still a huge demand. And a requirement from the hyperscalers out there as corporate buyers that want, you know, reliable baseload power and, you know, increasingly want to see it be clean and built for, you know, long duration so that they're going to have it available to them for a long period of time. So that has really changed from the first time around, both in terms of what's driving the demand and the rest of the environment around it from a market perspective. Um, really much better, uh, for nuclear this time than the first time.

AI assessment note: “And a requirement from the hyperscalers out there as corporate buyers”

Q in project development. And then there's also this question of how do we get the cost down in nuclear, which is at least in part a function of building multiple projects, not a single project, and then driving down a cost curve. So I, I know the hyperscalers have been thinking about, okay, what can we do beyond the obvious PPA. How do you think about that suite of possibilities?

A Yeah, that's been the biggest change from the hyperscalers probably in the last two or three years, Shale, is that, you know, when I was, um, before I came to Elemental, I was at New Scale, and we talked a lot to the hyperscalers, and really said in order for projects to go on, they needed to have some, you know, skin in the game, or some, some, um, you know, exposure to the development cost, because, you know, unlike renewables, where the development cost might be in the tens of millions of dollars at most, um, It's hundreds of millions of dollars for nuclear, and you can't just have a power purchase agreement to make that kind of investment as a developer. And, you know, it took a while for people to realize that that was what needed to happen. And, you know, necessity is the mother of invention, is that once people realize that I need clean and large amounts of base load megawatts, nuclear needs to be part of that, um, how I was doing before isn't going to work, that kind of moment really solidified over the last You know, year, year and a half with the, the hyperscalers are saying, okay, I need to do more than just say, I'll write you a power purchase agreement. We're going to make some investment. They made investment, you know, mostly in the technology, but increasingly as this deal we're doing now with Google is in the upfront development, which is really the key part, u…

AI assessment note: “increasingly as this deal we're doing now with Google is in the upfront development”

Q To fulfill the capacity of a data center, or both?

A It's really both, um, because we're solving for two things, right? One is, um, they have huge needs, but they, they want to see us get down the cost curve, which means you need to have multiples. And when you look at the sort of intersection of doing multiples, the size of the need they have, the capital they can deploy, and what it costs for various, whether it's micro, small, or large reactors, The small light water reactors fit more generally into that category, um, as well as some of the smaller gen four, um, reactors as well. So this is kind of anything from 75 megawatts to 300 megawatts electric. Um, that fits pretty well because, again, you get more swings at bat, so if you have one not work out, the other two will carry you through, which is how you want to see a portfolio work. And when you have multiples at a site and multiples in a, in a deployment, you You can get down those cost curves both in the factory and in the field more readily than you could with a large reactor. Um, so that's really how we, we saw it playing out, uh, in real time. Um, you know, and the hyperscalers looked at all of them. And, you know, it was our goal as Elemental was to find the right solution that kind of threads the needle between the OEMs, utilities, and the customers in terms of what makes the most sense. And that's where we landed up, uh, in our conversations with Google.

AI assessment note: “It's really both, um, because we're solving for two things, right?”

Q one company that has actually built a new nuclear generating facility in the past, I don't know, three decades, four decades, whatever it's been. Um, and now they, many of them are also exploring new nuclear, but I think, you know, coming into it maybe a little bit, a little bit more caution, at least than the hyperscalers seem to be. What's your perspective on what you're hearing from utilities?

A Yeah, so, you know, what I hear from utilities now is what I heard from utilities back in the, in the past, right, is that, you know, they fundamentally can't take development risk. Um, and, you know, that's a, having done fossil plant development, I did that for 15 years of coal and gas plants, that's a business where you expect to make two or three times return on your, on your investment, uh, on the development at FID, and that doesn't happen, uh, for utilities. They're stuck at 10%. Right? So they're taking three X risk for a 10% return. It just doesn't work out for them. And, you know, we're seeing with a, with many of them, not all of them, but many of them is a desire to have a developer like Elemental come in, align with the hyperscaler, take care of the commercial risk, take care of the development and the construction risk, and then transfer to Uh, the asset at commercial operation when, you know, you can just look at what Constellation trades at in terms of what their cost of capital is. Pretty easy for them to do that. Very huge markups and value to them to do so. Um, or they can operate the plant, um, as the operator for it in their territory. And we're looking at all those models with them, but, you know, what we really see is that the, the recognition that somebody needs to come in to take that, uh, development, uh, And construction risk. Uh, the government's tak…

AI assessment note: “what I hear from utilities now is... they fundamentally can't take development risk.”

Q it's a different answer if you're Google or Meta or, or whatever versus, uh, versus somebody else. But like, clearly we need to drive costs down. It's not entirely clear. The first project anybody builds is going to be the cheapest. So what does it take? Is it just repeatability? Do the same reactor over and over again? Is it at a site level? Like what's your thesis on cost?

A Yeah, there's, there's, Probably two levels to it. One is, is repeatability, right? So just doing a one-off design isn't going to work, and so that's why it led us to the three projects with Google. You know, is three a good number? Yes. Would five be better? Absolutely. Does it need to be five? It would be a bigger check for people to write to do that, so you kind of have to be constrained by what that part is, and you know, to put it mildly, you can't do enough large reactors in order, you know, and raise the capital to do it. It's just too much of a check to write. Uh, to make it happen. So you need that repeatability because, you know, Google and everybody we've talked to recognizes that, okay, the first one is going to be more expensive than the second one than the third one. And, you know, if I have a project that has two units on a site, the first unit on that site is going to be more expensive than the second unit on the site. That was true for Vogel, that was true for Baraka, that's true for any single, you know, any time you've looked at a construction that had multiple of the same units on the same site, units two, three, and four, you're amazed at how much less they are. Um, but it shouldn't be surprising, because if you ever went to Ikea and bought, like, four pieces of furniture of the same type and constructed them yourself, The first one, you probably break it a…

AI assessment note: “One is, is repeatability, right? So just doing a one-off design isn't going to work”

Q offtake side. Um, let's talk about the supplier universe. Obviously it used to be at new scale, which was, is one of the, Suppliers, OEMs in the space, but it's an interesting landscape there of like lots of activity, but also very few proven and or approved from a regulatory perspective, uh, products, I suppose, that you can buy as a developer. So how do you think about that landscape?

A So, you know, we approached it in that there's been billions of dollars invested by these technology, uh, developers and either, you know, traditional light water reactors, which you call Gen three or Gen four, uh, Uh, reactors, which might be high temperature gas or sodium cooled. Um, and they're all kind of nearing the end of their development process where they're not quite ready yet. They're all kind of trying to solve for something, whether it's in supply chain or licensing or whatnot. So we're technology agnostic in the belief that, or the knowledge that we can go out and find sites that are amenable to the hyperscale is what they need for offtake and location, um, that we can develop in a way that in a year from now or, A year and a half from now, we can pick the technologies that have made it through their sort of crucible of licensing and development and, and cost, um, estimation that we can pick with confidence that we know what we're going to get, when we're going to get it, and what it's going to cost. They're, they're almost there, and, you know, I can't tell you which one of them is going to be successful, but there's so many of them out there that have invested so much. I'm very comfortable that, There's going to be, you know, probably three or four at least that are going to make it through that part of it that will be successful in the long term. Um, just hard …

AI assessment note: “So we're technology agnostic in the belief that, or the knowledge that we can”

Q Do you see any evidence? There's obviously been a lot of talk about having the NRC reformed in some fashion or another to speed up permitting. Um, do you see any momentum There from the NRC, has anything changed? Do you need anything to change there?

A So, you know, there's been a lot of legislation over the last several years really driving the NRC to revise its policies, become more, um, balanced in its approach for, you know, making sure to maintain the safety to the public and the environment, but, you know, recognizing that there's a public benefit for nuclear power given its clean baseload attributes, um, That has been demonstrated by reduced timelines for them to review various applications I've observed just in the 15 years I've been in the nuclear space where, you know, what once took four years is now being done in two years. Uh, the hope is that we'll get down to maybe a year, year and a half. But, you know, what you'd like to see is that for this baseload technology, which is what it is, uh, to be sort of in the same time frame of development of a traditional combined cycle gas plant, which might be You know, two years of development and licensing, and then two years construction, and then, you know, four years or five years, you're operational. Um, we're not there yet, but I do see a path by which you get down to, from the seven to eight years, we're looking at the first ones, to getting down to five or six years, which to me really makes a game changer when people have their planning horizons. That brings it all forward to be a real equal, um, discussion. And, you know, I think we're going to get there, um, and …

AI assessment note: “what once took four years is now being done in two years.”

Q How do I take advantage of the fact that I do actually have experience in nuclear to, to get that capacity back online or online in the first place? And that seems kind of universal, right? Like, I feel like every utility in their boardroom, if there's nuclear somewhere in their fleet, or there was supposed to be nuclear somewhere in their fleet, they're taking a fresh look at it.

A Yeah, I mean, I think that most of these utilities that are operating, uh, utilities of nuclear plants, if they had the opportunity to pick up an existing operating asset or one that was, you know, largely complete and de-risked, um, they'd be very inclined to do it, uh, because you just look at what, you know, Constellation stock price has done. I mean, that's the example I look at since it floated. Um, you know, it's just been phenomenal returns because of the preponderance of nuclear in its, in its base. And that's a, Purely merchant generator. It's not even a regulated utility constellation. It's, you know, phenomenal in terms of what it's done. But that just proves the value of nuclear, um, to all these folks. It's how do you get the assets on there? And, you know, interestingly, through the investment that, that EIP made in, in, in Elemental, we're exposed to a large number of those operating nuclear utilities. Um, so we have that capability of, uh, of working with them, and they're familiar with what we're doing. And, you know, in part, Validated what we are claiming to do, which is to really, you know, span the void between the billions of dollars that have been invested by OEMs doing new designs and the gigawatts of demand required by hyperscalers. How do you meet those two things and us coming in the middle and making that happen is really what Elemental is about.

AI assessment note: “if they had the opportunity to pick up an existing operating asset... they'd be very inclined”