Not Your Father’s Data Center - Revolutionizing Data Centers with Nuclear Solutions

Episode Date: April 7, 2026

In this episode, Raymond Hawkins, Chief Revenue Officer at Compass Datacenters, sits down with Matt Loszak, co-founder and CEO of Aolo Atomics. Matt Loszak shares his personal journey from growing up with severe asthma in Ontario—an experience that inspired his fascination with nuclear energy after coal plant closures led to both cleaner air and the end of his symptoms. His career spans from nuclear engineering studies to successful tech entrepreneurship, and eventually back to his roots in the energy sector with Aolo Atomics.The discussion explores the current and future challenges of powering rapidly scaling data centers, focusing on the unprecedented demand for clean, scalable electricity. Topics include the limitations of the current grid, the viability of modular and “extra modular” reactor technologies, and the promise of vertically integrated nuclear solutions. Matt Loszak offers deep insights into Aolo’s business model, the economics of on-site power generation, and the transformative potential of fast, factory-built nuclear power for the data center industry.Timestamped Overview00:00 Intro & Matt’s Background03:55 From software to nuclear impact07:16 Engineering Leader in Nuclear Innovation11:16 Nuclear Power The Future Answer16:28 Revolutionizing SMRs Lessons from SpaceX17:19 Vertical integration to solve delays21:15 Smaller plants over gigawatt nuclear24:50 Configurable real-time power solutions28:50 Pipeline vs power grid infrastructure31:37 Energy challenges in industry operations35:27 Nuclear the big picture solution39:56 Optimal small nuclear innovation40:35 Aolo’s customer-focused design approach

People are trying to get scrappy and take a lot of previously turned off infrastructure or existing infrastructure, you know, just get creative to soak up, you know, like a sponge, anything that's just low-hanging fruit. But I think that's kind of missing the point, which is there is not going to be 40 gigawatts of that. All right. Let's get rolling. Welcome to another edition of Not Your Father's Data Center.

I'm Raymond Hawkins calling in and joining you from the Compass Data Center Annex and Little Bitty Auburn, Alabama. And today, we are joined by someone who is going to solve all of our power problems and do it quickly. The co-founder and CEO of Allo Atomics, Matt, Luzac, Matt, thank you for joining us. Hey, Riemenda. Great to be here. Thanks for having me. No pressure. Just generate all the electrons for us. It's, you know, no one in our industry is really worried about where or how many electrons, but no, that man, our whole business. It's all all anyone talks about. Before we jump in and talk about how you're going to solve the limited electrons problem, Matt, if you're willing, do you mind telling us the winding road of what landed you in the world of modular reactors? And you go back as far as you want. You know, we'd love to hear where you're born, where you're raised, but you take us back as far as you'd like and we'll get to electrons in a minute. Absolutely. So, yeah, I grew up in Ontario and Canada. I first got interested in nuclear energy because I used to have asthma breathing attacks growing up.

We had around 62 smog days per year because we had a lot of coal power plants. But fortunately, the government decided to turn off those coal plants. We went all in on nuclear. And lo and behold, smog days went to zero. And my asthma went away. And I thought, my God, this technology is incredible. And so I studied nuclear engineering and physics in nuclear engineering and physics in university also kind of got the entrepreneurial itch. So did a few summer businesses. Ultimately,

you know, thought about starting a nuclear company when I graduated, but Fukushima had just happened. So I thought to myself, ah, maybe not the best time to start a nuclear company. So instead, tell myself to code, Googling, how do you build an app, how do you build a website, and did two software companies. The first one got a lot of users, but didn't make as much money. The second one, I thought, you know, has to have a much better business model. So started a HR payroll and benefits software suite in Canada, kind of like gusto or rippling for Canada, if you're familiar with those. I am, sure. So we scaled it to 150 employees. And we processed around $10 billion of payroll with maple syrup money, all the Canadian businesses. And then sold it for, you know, low nine figures. So it was a

decent outcome. But yeah, around 2021 or 2022, I started to think maybe now is the right time to start the nuclear company. And it's worth highlighting, you know, it still didn't really seem obvious because we were still shutting down gigawatt scale nuclear plants globally. And I was like, well, what are the odds that a startup can have success if we can't even keep multibillion dollar paid off assets alive? But it felt too meaningful not to do you. I think nuclear is the most misunderstood technology that I've ever come across. And so I dove in head first and the rest has been history. All right. I'm going backtrack a little bit for you. So Fukushima is 2011, right? Yeah. So 14 years ago. So how old are you in 2011? I was around 20.

Okay. So you're just getting out of school or you're wrapping up university and you're thinking, hey, this is what I'd like to do after I finish school. But with the bad press of Fukushima, you decide, no, no, no, I'm just going to learn software development. And that was a bit of a detour, hard to say a nine-figure sales startup is a detour. But you kind of took a 10, nine, 10-year detour in that route, but this deeply personal connection to nuclear, because it sounds like when you say, hey, I bad bad asthma. And when the coal plants went away, my asthma went away, this sounds like a very personal connection to, man, this is kind of a good idea. Is that a fair way to say it?

Yeah, I think so. And, you know, I always thought nuclear was incredible. It's like I also used to be interested in magic as a kid. You know, that's part of what drove me to study physics is because physics is like the closest thing to magic in real life. And nuclear is like, you know, literally alchemy. It's, it's crazy. You're turning one element into another element or multiple elements. But I just, I didn't think that starting a nuclear company would be possible. I just, I knew I wanted to start something in this space. And, maybe something with a physical product, but it just seemed too audacious at the time. And so I thought, hey, software is a better way to get started, you know, get my legs under me and maybe try to find a way to do this eventually. Gotcha. All right. So you detoured. You did a couple cool software things.

You made a little money doing that. And so now you're in your early 30s and you decide, you know what, let's go back to my first love. I still think the magic of nuclear is pretty fascinating and is, is, is, um, is, um, Is this the first? Is Allo the first attack at that for you? It is, yeah. I did, you know, I'm pretty methodical person. So I did my diligence.

And for six months, I was like, okay, I know I wanted to do nuclear, but let me just make sure. So I went out and I spoke to some people in other spaces like biotech, AI, crypto robotics. And I went down some rabbit holes there. And I was like, could I see myself doing that, you know, for the rest of my life? but nuclear kind of kept on pulling at my heartstrings. Yeah, yeah. It seemed like, keep in mind that this is before data centers were growing at this speed and before nuclear, you know, it still had a lot, even more stigma around it in 21, 22.

So it really still didn't feel obvious at that time. But sure. I was just like, you know, I have to give this a try. I'm not sure it feels obvious to everybody today, Matt, but I think you and I are going to do something in the next 20 or 30 minutes to change some of that. For sure. Yeah. Do you mind giving us one minute? So you've got a degree in physics. Your co-founder, what's what's his background or her background? I don't know. Yeah. So his name is Yasser. And he and I graduated around the same time. And, you know, similarly, he also loved nuclear for different reasons. So he grew up in Bangladesh. Whereas I grew up in the grid of the future with nuclear. He grew up in a world where sea levels were rising. Global warming was causing this, things like that. So he came out of from. a different angle, but with a very, you know, the same passion for nuclear and envision for, for nuclear, you know, should become. And so he decided to go straight into industry.

So he went into Westinghouse, worked there for 10 years. He worked on AP 1000 and did R&D. And he worked his way up. He is, you know, a gifted engineering leader. And so he ultimately led the Westinghouse Evinci program, which is their microreactor energy. And he is, um, he's, you know, a gifted engineering leader. And so he, um, he ultimately led the housing house evinci program, which is their microreactor entrant. And then he led the Marvel program, which was a government program at Idaho National Lab, which was the first reactor that the DOE had ever approved for design. So Yasser and some of our earliest employees came from that Marvel program. That was kind of one core tenant of our initial strategy was DOE authorization as this way to get a new technology built faster than might otherwise be possible through the NRC, the traditional

regulator nuclear. Yeah. So, and maybe I'm steering this in the wrong direction too quickly. So maybe correct me, you said your employees came over from the Marvel program. Are you, do you have end user customers yet? Or is that still, are we still too early for that? Is the business, for lack of better word, pre-revenue? Are you guys off selling stuff already?

It's been interesting. There's, there's nuclear companies that have been around for a few months and nuclear companies that have been around for 20 years and everything in between. We're definitely one of the faster moving nuclear companies. So two years ago, we were two people. Okay. And today we are 100 people. Okay. And we've raised around 140 million in that time. So we've been really expanding quickly in terms of the technology roadmap with the six million seed and the 30 million series A. We built out a full-scale non-nuclear prototype, as well as a 40,000 square foot pilot factory space. And we only spent around 13 million to achieve all that and scaling the team to around 50 people at that point. Gotcha. And so then we raised the 100 million this past summer. And with the residual, call it 20 million from the seed in Series A plus the 100. This is all paying for our first real nuclear power plant, which is the all-o-euvre. X. And that plant is turning on next year. So I hope when I say this, you're shocked.

Yeah. You know, because to go from founding to Fission in three years is really unprecedented. And obviously, you know, proving that tech out next year will be what really leads towards revenue and customers in the year or truth that follows. Well, yeah. So my father-in-law was a longtime NRC guy and 15 years to get a, you know, permitted and built. So, so, you know, in the traditional NRC world. So when you say concept to deployed in three years, that's breakneck speed. And by the way, in the nick of time, too, right? Because to the point you made, you know, so I've been in the data center business about 12 years now. You know, when I got here 12 years ago, it was commonplace to talk about kilowatt deals.

And rare that we talked about megawatt deals. not a dozen years ago. Then it became pretty commonplace within a few years to routinely talk about megawatt deals. We didn't talk about, you know, tens of megawatts until five or six years ago, meaning in a single transaction. And then we started talking about, you know,

tens and tens and now hundreds and gigawatts. All of that is in the last 36 months. So as you talk about your parallel path from C, money three years ago to, you know, your Series B and deploying next year, I mean, you are almost running a parallel path to the unique, you know, what we call around here, the rocket ship rise of the data center business or the demand for the data center business, right? And at the end of the day, we measure everything by kilowatts, how many kilowatts do you need? And it is just shocking how big the projects have gotten. And as we've talked in the industry,

I would say, you know, probably four or five years now, I think, you know, the people that think about these problems and worry about these problems have said the only answer is nuclear, right? Because we can't, we can't make the wind blow 24-7, 365, which is what my customers want. We can't make the sunshine 24-7-365 because that's what my customers want for availability. So if neither of those are the right answer, how do we get a sufficient amount of power that doesn't do, you know, a potential harm to our planet? And it is felt for years that the answer is nuclear. But to your point, it's shown up with a Chernobyl or if you're as old as me, a three-mile island or a Fukushima anchor around its neck.

And I think the sheer demand and the youthful nature of the decision makers in my customer base are transforming that. And you guys view, your industry, feel like an answer to that. So if that is a setup, no pressure, please tell us how you're going to give us lots of electrons. really fast all over. That's what we'd love to hear because, you know, we are doing data center projects in far-flung locations solely because it's where we can get sufficient generation. And it's okay for today, but it's probably we can't do that forever. And I think one of the only ways to do that at scale and clean is nuclear.

So you enter stage right, allotomics, please help. Yeah, yeah. Well, I think you really laid the groundwork nicely there, which is, you know, in the past 10, 20 years and before, it was pretty easy to just tap into the grid for tiny amounts of power. And, you know, there were other constraints that data centers might have looked at, right? But now when you're talking about gigawatt scale data centers or even, you know, 10 gigawatts, those are like cities, like, you know, medium to, you know, medium to, very large cities on par with how much, you know, energy New York or Toronto uses. So obviously, you know, this demands a new paradigm because while the grid can supply you on the kilowatt scale very easily, it's, it can't just wave a magic wand and suddenly be offering, you know,

one to 10 gigawatts like that, that requires entirely new generation infrastructure. And I think it's, it's worth, you know, laying the land a little bit on the way this is getting solved today. I'm assuming maybe most of the listeners are aware, but most of these new gigawatt scale data centers are being powered by natural gas. And, you know, there is some pipeline available that's kind of low-hanging fruit at the, you know, a couple of gigawatts scale. But if we're talking about 40 gigawatts in the U.S. in the next five years alone, and I've seen some estimates that are even higher, that natural gas infrastructure, you know, doesn't exist. So you'd have to do new fracking, new pipeline, et cetera. And the reality is there's no existing silver bullet solution, right? Like renewables use a lot of land and are more expensive when you firm them up.

Geothermal hydro, you know, you have to go find a magma pocket or fast moving water. That's not in many places. It's mostly tapped out. That's kind of why, you know, you see these hypers making agreements with nuclear companies. Because if there's no silver bullet, you might as well, you know, try to use. the solution that would be ultimately the best, which is the most scalable, the cleanest, base load, et cetera. Hey, Matt, you slid through that one real quick.

Can you talk about one of those deals? I mean, let's take the Susquehanna deal in Pennsylvania. And if you don't know the background in Susquehanna, it's okay. But will you talk through one or two nuclear deals that had been done out there with traditional, and I'm going to say, I don't know if that's the right word, traditional nuclear infrastructure, either mothballed infrastructure or, you know, infrastructure that's got some of its capacity offline. Well, there's a bunch. And we can speak to, you know, Three Mile Island getting turned back on. We can speak to Amazon and Tallin signing a deal. But I think that we can really summarize it as follows, which is, you know, people are trying to get scrappy and take a lot of previously turned off infrastructure or existing infrastructure, you know, just get creative to soak up, you know, like a sponsor. anything that's just low-hanging fruit.

But I think that's kind of missing the point, which is, you know, there is not going to be 40 gigawatts of that. And that's really what I think, you know, all I was focused on is, is how do we kind of go from the low-hanging fruit to the next 40 gigs and beyond? And so I just want to maybe spend a bit of time explaining what's so unique about our approach to this.

Yeah. And our stance. stance. So I'm going to go ahead and say, Matt, I've not heard the extra modular reactor term until I saw y'all's stuff. So we'd love to hear, you know, SMR gets thrown around in our space pretty regularly. I'd love to hear what's unique about that. So I think it's analogous to, you know, in rocketry, right? People would often think of landing rockets as improving the economics. And no one had done it properly until SpaceX. I think the term SMR is kind of similar in the sense that everyone's familiar. with the idea of, you know, make a bunch of modules in a factory, ship it on site. But the reality is nobody's done it properly yet. And I might sound, this might kind of sound like a BS marketing term, but I actually, I actually mean it. So let me, let me explain what I mean by this. If you look at basically all the publicly traded nuclear companies and many of the private ones, they don't have a single centralized factory. Some of the big names you might have heard of are

actually fully remote companies. And what they do is they design power plants. and then they send these designs out to dozens of suppliers who each have their own factory, and then they try to integrate it at site and problems ensue. There's some big publicly traded companies that are more young that fit this bill, and there are also older incumbents. For example, the last big plant that was built in the U.S. Vogel was very much like this. They had a bunch of different factories that made the vessels, the turbines, and everything in between. And when they tried to integrate it at site, there were slowdowns and issue.

with things not lining up properly. So we're like, okay, you know, we think the solution is taking vertical integration further and having a factory that takes raw materials in and outputs, modules that you know will fit together at the site. We could have called it an SMR, but we had to make a new term because no one's doing this. And it takes, you know, more capital. It takes more talent. It's a lot harder to do. But we think it's really what's necessary to solve nuclear's main problem, which is, you know, the utilities, they have a fear of cost and schedule overruns. What better way to resolve that than to have a single centralized factory that's outputting most of the modules? And there's a lot of subtlety there. Hold on. I want to make

sure I got what you're saying there. So the SMR term is somewhat generic and somewhat references, hey, a reactor that's smaller than what you've seen when you see a chilling tower off on the horizon. But most of that is done by cobbling together parts from multiple different facilities or manufacturers. Vendors. Yeah, vendors. Okay, companies. The XMR idea is, hey, I'm going to build. My company is going to build all of these parts. For lack of a better word, I'm going to build an integrated solution or a complete solution, and I'm going to deliver it from my factory to where it generates capacity. And because you felt that was a unique approach, you felt calling in an SMR, which general, I think people who aren't in the nuclear know, think that refers to scale,

size. Your reference is really about a business model or design and delivery approach. Is that, is that what your, is that a fairer summary? I'd say that's fair. Am I butchering it? Yeah. Yeah. And another way to frame that is the factory is also our product. It's not just the power plant, the reactor, the factory. And that's why we, you know, half. of all the money we've raised, he goes towards the factory. It's not just the reactor in this design. So we're putting our money and our talent where our mouth is on this. So your value proposition, Allo's value proposition to the market is, hey, just like you wouldn't like a Toyota that, you know, had parts, you know, assembled from 62 different manufacturers. You want one person that builds it.

You're trying to build this fully integrated, elegant solution all under one roof. Yeah. And also, just to be clear. So, you know, 80 to 90 percent of the modules initially will be coming from our factory. We're still buying, you know, the turbine, some heat exchangers, some pumps, and obviously the fuel. But otherwise, it's mostly raw materials that we buy. Okay. And metal comes into our factory and modules leave our factory. And I'm going to ask a really dorky question that, you know, some of our editors may edit out here, Matt, but I'm going to ask it anyway. When I understand the way a grid operates, it operates by the, the, you know, in the U.S., was it 50 hertz, and we're trying to maintain that balance in the grid.

Nuclear turbines doing the same thing, right? It's offering that same sort of grid stability as a regular turbine, right? Yeah. Okay. So I want to make sure I understood that, right? Because I do know that is one challenge when you have solar or wind. Your turbines don't have the same steady state, what maybe even called base load. I'm going to use the word harmonics, even though I know it's not the right word, to maintain that balance on the grid.

Nuclear doesn't face that problem, does it? Yeah, so nuclear is a thermal generator that does have, helps solve that. I think maybe another related point to make is, you know, if we're talking about traditional thermal generators, for example, a gigawatt scale nuclear plant, as compared to our approach, you know, let's say you're building a one to 10 megawatt data center. why not just build a couple gigawatt-scale nuclear plants, right? And there's a few reasons that we think our approach will make more sense. So one is, if you look at how data centers are built these days, it's in these kind of, you know, 100 to 200 megawatt data center shell buildings that are stamped out multiple times to reach gigawatt or more.

And those shells take less than a year to build and outfit with GPUs. So if it takes you 10 years in the best case to build a nuclear plant, that's just too slow. So, you know, that's why building smaller plants in a factory allows you to go faster and build more incrementally along with the way data centers get built. And it's kind of more aligned from a time value of money perspective with those data centers. The second big reason is because your gigawatt scale nuclear plant every two years goes offline for refueling. So you lose a gigawatt every two years and you need a backup, like a grid backup. That's a big ask of the grid.

Yeah, yeah. I can't, I mean, I get a zero downtime for my customers. So how long does a large scale nuclear plant go offline to refuel? One month. Oh, my goodness. I had no idea. So when you have a fleet of smaller reactors, not too small where the economics sucks, but, you know, a little bit smaller so that you refuel each.

smaller reactor one by one, and you're never that short of power, maybe 10 to 50 megawatts out of a one to 10 gigawatt fleet. So those are the two advantages. Yeah, by the way, I love your idea. Yeah, I mean, you nailed it. This is my business, right? I build buildings and, and let customers roll GPUs in. That's the business. And your understanding of it is accurate, right? If I have to wait 10 years, I mean, two thirds of my contract's already over. You're not even a legitimate interest to me. Where do you come down on either? being the backup generation or maybe even being a microgrid solution for me, or for me, meaning the industry, not me personally or compass person.

Have you guys thought about that? Because at the end of the day, all of my facilities have dual generation. You know, I'm getting a utility feed and then I'm getting some generation on site. Are you guys thinking of replacing the utility feed? Are you thinking of being a, are you considering yourself to be a microgrid or you can You send yourself to be a backup solution or all three. And are there pros and cons from your perspective on all three of those approaches? Yeah.

So we want to be your primary power. Okay. And the way we see this is we believe we will be the fastest energy source you can deploy, even including the grid interconnect at this scale. Okay. So the idea is, you know, let's say you're building a 1 to 10 gigawaw data center. You'd be deploying our Alopod products, which are 50 megawy, each electric.

Can you give me a footprint? Two and a half acres. Okay, two and a half acres. All right. And that's everything is going to be contained. I'm going to get 50 megs out of a two and a half acre footprint. That's to the fence boundary. Yeah.

Yeah, okay. Yeah, the idea is you'd be installing this in real time alongside your data center buildup. And then you'd have a choice. You could either install extra pods for an N plus one and plus two. and eventually get a grid interconnect and optimize your economics of your project by selling your N plus one, N plus two energy onto the grid to stabilize the grid,

be a good Samaritan. And if you ever, you know, when you're refueling or doing maintenance, use your N plus one or plus two as your backup. Or you could just install, you know, up to the number of pods you need and then use the grid as the backup when needed. But, you know, there will be options there,

but the whole idea is this can be primary power, and it should be, it'll be desirable to be, because it'll be much faster and very configurable. I will say this, the one, you know, this sounds good to be true. I want to say the tradeoff here is it'll be a little bit more expensive on an LCOE basis as compared to, for example, a gigawatt scale nuclear plant. But we think the customers will pay that premium because they value speed. you know, right now in the next five, ten years, we think this will be the fastest, cleanest,

uh, base load solution. Yeah. So, so conceptually where you're going, I like, can I, can I keep poking at it for a little bit? Please.

So, so if, if, if my gigawatt scale reactor has to go offline for a month to refuel, how, how long does my 50 megawatt, two and a half acre site have to go off? Is, is it a single generation or do you have some redundancy built into the 50 megs?

Yeah. So instead of losing 100% of your gigawatt, gigawatt for a month, you would lose 5% or 1% of your fleet for two weeks and that it would be a rolling. So basically, you would just change your nameplate at that point. So hold on. I want to make sure. So you're telling me you have a 50 megawatt pod. So I'm going to do a gigawatt site. I'm going to need 20 of these just to be at end. Right? So I'm going to be 20 at end. Does that whole 50 megs come down when I do a fuel regent or a fuel refresh or what do you guys call it? Do I take all 50 megs down at one time?

So for me to run it in, you know, I'd need N plus something to roll my maintenance window or my refueling. I guess it's refueling is the right term, right? Well, here's how the math shakes out. So it'll take a week or two to do refueling on each reactor. So to answer your question directly, you're only losing. So each pod is five reactors and one turbine. Okay. So do those five reactors all have to run to get the turbine to produce its maximum load? No. Okay. So that's where the secret sauce comes in. So let's say you have 20 pods for gigawatt. That's 100 reactors. And you have to refuel a reactor roughly every three years. So in three years, there's around 150 weeks. You can easily do refueling of one reactor out of 100 every week and still have time to spare.

a rolling maintenance schedule. You're just going to start on one end and roll to the other. Exactly. Yeah, yeah, exactly. Your vision is based primarily on the notion that, hey, I'm going to be the fastest to power at scale. So I should be your primary power source and let me use the grid as backup. That is largely because of speed to market.

Is that your theory, hypothesis? Yeah, speed and reliability, right, in terms of this, you know, you have this fleet where if there's maintenance on a turbine or refueling a reactor, you've got a fleet of these things. And so the high availability is so interesting, right? You look at this space and high availability is so hard to achieve with on-site energy generation. For example, you know, natural gas, sometimes the pipeline doesn't have natural gas supply. There's energy, right? The pipelines, and I don't mean to blow up your thesis, the pipeline is not a great

pair because of all the things you're already saying, right? I don't have sufficient pipeline infrastructure. I don't have, you know, six-nines level availability of the fuel. There's no question there are natural gas powered data center facilities today. But the biggest customers aren't ready to say that's my primary power source for all the things you're citing. Our biggest bottleneck is getting the grid built out to provide because my customers, they all. So at the high voltage level, they're looking at six-nines, seven-nines availability at the high-voltage level.

So the customers are trying to get as close to that HV line as they can get. Yeah. And are you comfortable with that level of six or seven-nines availability? Well, let's just debate this for a second. So, you know, I want to make sure I made my point clearly. I think we might be talking past each other a bit there because what I'm arguing and tell me if you disagree is that nuclear will somewhat uniquely be able to be that primary power because of the high degree of predictability of the fleet. And part of this, for example, is you buy the nuclear fuel and then you have it there.

You know it's going to work for three years. That's very predictable, both on, you know, cost and uptime. My understanding and tell me if this is wrong is with natural gas pipelines, sometimes the gas might not float and you might have downtime. And that's part of why. Well, that's 100% right. Yeah, you're 100% right there. Yeah, yeah, yeah.

Yeah, yeah, yeah, no, your thesis about. why the industry isn't gaga over natural gas is right. At the beginning, you said, hey, we think most large-scale data center campuses are powered by natural gas. And I'm not meaning to shoot holes in your thesis. That's not right. Most large-scale data center campuses are powered off the grid.

No question natural gas plays a role. It's just not number one. And so when I think about, and by the way, there's this term behind the meter. And when you say primary source, largely in our industry, that's behind the meter.

there's some in our industry just giving some insight into our space that view that as a four-letter word because the marketplace looks at a data center today and goes, wow, look at those evil data center developers, look at all the energy they're using, right? Nobody stops and says, oh, you know what comes out of that building over there is my Uber Eats, it's how I book my plane tickets, it's where my Netflix comes from, it's how I find my dates on Tinder. Nobody thinks about that. They just think, oh, that evil building over there is eating up a bunch of electricity. The public has yet to connect the dots that, no, all I'm doing is standing up a building that's providing you all of the things that you do on your cell phone, right?

Yeah. And so as an industry, we're very, very careful to think through what we're doing from an energy perspective and this behind the meter notion. And again, this is, I'm glad we're talking about this because this is not an assault on your thesis. This is a worthwhile thing for us to talk about. the idea that I can get quick, reliable, steady state access to power at scale is a game changer for us as an industry. We've got to get the marketplace, meaning the my next door neighbor,

my community, the place where I build my building, comfortable that me standing up behind me as an industry, not compass, standing up power behind the meter isn't a loss. And by the way, I'm not staking out a position, but there is a perception in the market that I'm, oh, if you just went off and built your own power, you're somehow hurting the rate payers in the state. And I just think all of that needs to be talked about. But for your business, your thesis of I'm going to be the fastest to deploy at scale, I think works great. I do think there's got to be some talk from a, hey, is it going to the grid first and then coming to the data center? That, I think, is still an open question and one you guys probably ought to wrestle with.

Yeah. No, I totally agree. And I'd love to run into area to buy you. But first of all, just to address your prior point, and I want to clarify, I think you're right. I was kind of more talking to some of the newer gigawatt-scale data centers that I've, you know, heard of or visited, for example, what Cruz was doing in Abilene or the Memphis. What's on the drawing board? Yeah, yeah.

Yeah, both of those are good examples. The drawing board says they're going to be Nat Gas. You're 100% right. Right, right. And the interesting thing there, and also with a lot of these, like you're saying, and this uptails into your second point, is the nimbiasm for data centers. And, you know, I had this tweet the other day, which was that, you know, it's the curveball no one saw coming, but nuclear could be the solution to nimbiasm for data centers.

Exactly right. Yes, yes. Because, you know, I know that water usage isn't as big of a problem as the public is making it out to be for data centers, but, you know, we don't require water for cooling. Yeah. Our reactor. And we can actually stabilize the grid. So there's a cool opportunity here for us to actually deliver some. Yimbism, some yes in my backyard, because, you know, you can stabilize the grid. You can

be a good Samaritan, not use water, and give them all those great services on their phones, as you mentioned, that they're so, you know, in love with. They're addicted to. We can go ahead and say that they're addicted to. Well, I mean, it's interesting. You know, I think what you guys are doing, you, Allo and you, the atomic industry, is clearly answering a problem we have because the demand for electricity is not going away, right? So we're not going to, we're not going to wake up tomorrow and go, you know what, we can live with 20% less electrification. So that's not happening. And we've got to figure out a way to do it that doesn't, you know, degrade our quality of life here on the planet, see also your asthma in Ontario, you know, 20 years ago. And then third,

which is the piece I think that is right in your wheelhouse is we got to do it faster. We got it, It's got to do it faster because the market is screaming for more power. And, you know, we are doing it, we're doing it at a scale that I think, you know, most people don't comprehend. I mean, you mentioned 10 megawatts as a city. Yeah, I mean, it's a thousands, tens upon tens of thousands of homes. And it's just hard for people to understand how much is a megawatt of power, how much is a gigawatt of power. We're talking about insane amounts of generation. and the way you're going to get that generation at scale is nuclear is the only answer.

So we got to figure it out. We got to figure it out and we can't wait 15 years. I agree. I agreed. So let's get off the phone and please go for me. Can you get back to work, Matt, please? Yeah. It is, it is, I think, you know, and I certainly don't speak for our customers, but I think, you know, lots of, if I could, you know, sort of cobble together lots of customer conversations.

I think there's lots of my customers that think this is the only answer from a big scale, big picture, long-term answer is that we're going to have to get comfortable with nuclear. And people like you guys doing this fast, doing it smart, doing it at scale is going to change things. I am interested, we don't need to do it on this podcast. I mean, I am interested in understanding the cost. Your thesis, you may mention you said, hey, it's going to come in a problem. premium, we think they'll be comfortable with the premium for speed. I think you're right today. But my customers, you know, they generally sign 15 or 20 year leases. And the cost of electricity is in longstanding ongoing operational costs that they're very sensitive to, right? Because

anything that you pay every month for 20 years ends up being a lot of money. And so I don't think you're wrong. I think there's early adopters who will be willing to pay a premium, but I think they're going to force your industry to get grid competitive from a price perspective, right? I think you can charge a premium for a while, but I think you're going to see that long term, as you guys develop a solution and end up being a viable answer, that you're going to have to get to competitive rates with the grid. Yeah. I know I agree. And I have like two or three somewhat disjointed points to make on this topic. Yeah, yeah. So maybe bear with me. Yeah. Yeah. Go disjoint it. By the way, I do disjoint it really well.

Yeah, and feel free to interrupt. But the first point is, I think people in your space probably realize, but it's worth stating clearly, when you do on-site power gen, you are saving half of the cost of delivered electricity because you're minimizing transmission, just a mutation, right? And all the loss along the way for those, too. Yeah, yeah, exactly. So that gives us a bit of a cost benefit there.

Secondly, I think you're right. You know, the onus is on us to lower the cost of nuclear and get grid competitive and get competitive with even natural gas and so on. And I think we'll get there. But it'll take a little, a few cycles of iteration in the factory to achieve that. But ultimately, that will be the moat, right? You kind of look at, for example, SpaceX, now they're responsible for 90% of launches globally. you know, they've come down the cost curve because they've made bigger rockets at a higher volume,

and now they're dollars per kilogram to orbit is, you know, very unprecedented in terms of how low-ins. Yeah, very reasonable. Yeah, yeah, yeah. So I think that's like the cool thing is this whole factory vision, it wasn't really possible in the past because there wasn't enough demand for a consistent enough product with a high enough initial willingness to pay to even have a chance achieving this. Yeah, why build a factory for something that doesn't have the demand that the benefit of the factory can feed? That's a great point. Yeah. Yeah, lots of the nuclear plants in the past were bespoke one-offs. There's a great

report that was done by the DOE. It's the liftoff report showing all the plants built in the U.S. ever. And there's like the most iterations of the same design that were ever built as like a a couple, not more than like five. So this is really interesting. And then the last thing I'd say on this is we think a lot about economics. And I can't give exact numbers. But I will say we will be in the range of 10 to 20 cents per kilowatt hour for the first of a kind plants. And eventually we think we can get down below 10.

On our website, we say three cents per kilowatt hour. We think that that's probably where this can go eventually with the right product. Like I said, SpaceX. Just heads up, Matt. You'll make friends at three cents. Just seen it. Yeah. I mean, you'll have lots of buddies.

I think that is a holy grail, you know, energy product, right? That's like that quickly deployable, clean baseload. And I think that from first principles, that's possible to nuclear. And what we're trying to do as a company is home in on the optimal balance of economy of numbers and economy of scale. Right. So, you know, we're not fooling ourselves into thinking, you know, when you go smaller, there are certain hits you take. But there's tradeoffs. There's also certain advantages.

And we're trying to find what is that optimum. And we think that is going to be the secret to, you know, achieving the kind of global optimum price for nuclear. We don't think it's necessarily a gigawaw scale nuclear plant with a thousand different suppliers. So that's what we're homing in on. And one, sorry, I lied. One more thing that I want to say. Yeah, yeah.

It's really interesting how a lot of other vendors are taking, you know, ideas from the past five, 10, 15 years and kind of forcing it on the customers of today. And I'm very proud of what Allo has done, which is I think what a startup is meant to do, which is go talk to customers and ask what they want and design around that. And what I've seen in the space is a lot of designs that were kind of frankly designed for different purposes being forced upon the customers of today. So some that are good for remote diesel or military, which cost 20 to 40 cents. a kilowatt hour, which is fine for remote diesel markets where it's that expensive, but not the right product for data centers. Or, you know, products that were designed in an anti-nuclear era,

when the only way to get social license to build your plant was if you could say you were using nuclear waste as your fuel, which again is a bit is more expensive initially to do that. So, you know, AOLO has been really intentional on just frankly focusing on economics that are right for data centers with a product that is right for data centers. with an approach that is scalable enough to actually stand a chance at servicing data centers, meaning design for manufactureability and assembly baked into the design process from day zero, instead of saying, hey, we're going to get our first plant built, and then we'll try to mass manufacture it. We think that's crazy because it's going to change your design if you try to loop

in manufacturing plans after you've got your first one built and regulated. So that's kind of a big part of what makes our thesis somewhat unique. Matt, this has been an awesome talk. We probably have to have you back because I think there's probably a lot more we can talk about. And also, we'd love to hear how things are going as you guys advance into, you know, building a prototype and then building one that's customer deliverable next year. We'd love to hear about it. So if you're willing, we'd love to ask you to come back and talk with us again.

Thank you for sharing with us your journey from asthma to solve on the data center PowerPoint. problem. That's quite the winding trail. And it's really, really cool what you're doing and excited for us as an industry because we're going to need the power for sure. So, man, thank you for joining us. They get mad

at me when I say when we're recording. We're recording Thursday, October 30th, with the World Series currently 3 to 2. I don't know if you care about your Blue Jays anymore now that you live in Austin. Let's go, let's go. We're pulling for the Jays to win it all tomorrow night in Toronto. But

I'm going to go ahead and predict that the Jay is win in six. We'll see when this gets published if I was right or wrong. So, Matt, thanks for joining us, man. Awesome. Thank you for having me. Appreciate it. Take care.