PODCAST: Are small nuclear reactors the future of data center power?

James Cook: More people, more devices and more apps. They're all using the cloud for computing and data storage. And AI is expanding that too. And all of this growth is driving demand for energy. Projections show the data center power demand could double as soon as 2026.

Brian Gitt: In the summer of 2022, one of the largest utilities in the United States, notified the data center industry in Northern Virginia.

That they were going to be out of power and not able to connect new data center capacity. This sent an absolute shockwave through the industry and that sent all of the largest hyperscalers and co-location data center companies scrambling throughout the U S soaking up every available additional power asset they could find.

James Cook: That is Brian Gitt. Brian runs business development at the advanced Fision company Oklo.

Brian Gitt: We design, build, and operate small nuclear power plants and sell that electricity or heat to data centers, industrial companies, factories, and large campuses.

James Cook: Nuclear power has emerged as a potential solution to our data center, energy demands. It's high capacity. It's low carbon electricity. However the path to implementation is complex. There are regulatory hurdles. Public perception issues around nuclear power. And there are technical obstacles. So, on this episode of Trends and Insights, Brian Gitt is going to join Colm Shorten who manages corporate data solutions at JLL. And we're going to dig into the topic

This is Trends and Insights: The future of commercial real estate. My name is James Cook, and I am a researcher for JLL.

Brian Gitt: we think about a technology, we have to think about what problem is it? Actually solving. We're going to need a lot of more power. Everyone agrees about that, but what type of power and how quickly can that power come online and how is this all going to work?

When we look at the available options, most of all the excess power. There's a few pockets here and there, but generally in all the biggest data center hubs, that power is now gone are spoken for more or less. So we need a lot more transmission because that was ultimately the problem. Now when we look about the solutions of how are we going to deal with this in addition to all the new demand and the power constraints around transmission lines transmission lines take about 10 years to build.

They're really long term horizon to build that infrastructure. In the United States, most of the thermal power plants that are coal based and even many that are natural gas are getting. Forced offline prematurely.

There's a new EPA power plant rule that is going to force the premature closure of most of the coal plants by 2032, or if not by that date, they have to incorporate carbon capture and storage, which is an incredibly expensive technology. And many of these plants are not going to make that investment. So they're just going to set a planned retirement date sometime in the 2030s, wind and solar power, although great in terms of zero emissions, obviously are intermittent. They don't power a data center 24 seven, but they require the new transmission build out so you can build all the solar and when you want, but if you can't connect it to the grid, Because you don't have the transmission infrastructure. That's a problem.

Large nuclear plants are excellent power assets. 20 percent of the United States is powered by nuclear energy today, however, it takes at least a decade to build a new large plant. And here's the real kicker you are going to need 10 billion for each unit, at least. There are very few utility companies that are public companies where a CEO is going to bet their career, bet the entire company on one massive mega infrastructure project.

Natural gas plants. are amazing and clean burning fuel overall, but most fortune 500 companies don't count natural gas towards their emission reduction goals and their climate goals.

And there are new restrictions, for example, in the United States. You cannot build a new natural gas plant without incorporating carbon capture and storage according to the new EPA power plant rule. This is going to add a lot of costs. It also is parasitic on the load. It takes about 25 percent of the energy from the plant.

So Natural gas is about to get more expensive, and we don't have natural gas pipelines everywhere, and it's almost impossible to build natural gas pipelines. No one wants them built. So that leaves us with advanced nuclear, small modular nuclear reactors as being a really viable solution, because they provide the 24 7 clean power that the data centers need.

James Cook: So, Colm, are any of your clients, exploring these small modular reactors? Are they talking about them yet? Or is this too new?

Colm Shorten: No, I think we've seen a surge even in them in the public domain. Some of the key leaders would be the hyperscale. So we've seen from Amazon, they have a 52 billion expansion plan for their data centers. It's clear that the grid is not going to be able to support that.

People have used the phrase, bring your own beer. In this case, you're going to have to bring your own power, and they understand that. So they're now becoming involved in the solution and not just the problem. And beyond transmission, we have a generation problem and a distribution problem.

And the one thing that the small modular reactor gives is that you can bring that power to the site a lot quicker. It's scalable, it is modular, so we have seen Amazon get involved in a couple of investments supporting the development, and they're looking to bring on, we're seeing a convergence now of dates, both with Amazon, Microsoft, and Google.

Around the 2030 to 2032 timeframe. Now, what we find fascinating about that is that we know there's other challenges which are going to be around the social side and, getting people on board from a public perception because when people talk about nuclear sets light bulbs off in the head.

Whether it's Fukushima or Chernobyl, but it's a different type. So two things in parallel, I would say, and maybe we can get Brian to expand on it, is the type of nuclear that we're talking about is also different. Yes, our clients are getting involved in it, and they're starting to look at fission versus fusion.

One does have a waste product that's radioactive, whereas the other one doesn't. And what we're seeing is that because there is a drive and there is a recognition, That we need green power, we need it flexible, we need it modular, and we're going to need it pretty soon. We're now seeing billions of dollars being invested in this technology in order to make it happen.

 I think we've seen a surge even in the public domain. Some of the key leaders would be the hyperscale. So we've seen from Amazon, they have a 52 billion expansion plan for their data centers. , it's clear that the grid is not going to be able to support that.

People have used the phrase, bring your own beer. In this case, you're going to have to bring your own power, and they understand that. So they're now becoming involved in the solution and not just the problem. And beyond transmission, we have a generation problem and a distribution problem.

We know there's other challenges which are going to be around the social side and, getting people on board from a public perception because when people talk about nuclear sets light bulbs off in the head.

James Cook: This is interesting for me because I grew up in the eighties in central Pennsylvania, very close to three mile Island. I was there when all that stuff happened. I was a kid, but I remember it. Our perception was very scary about nuclear power.

Brian, I'm sure you run into this every day. What you're talking about is different from what people think of, when they think of that kind of nuclear power.

Brian Gitt: These advanced reactor designs are meltdown proof for practical purposes. They use natural forces to self cool and self regulate.

When someone thinks about a nuclear accident, let's say take Fukushima as an example, that's in people's minds, what happened? You had a tsunami. They created big wave of water that came in and they shut the reactors down. They shut the nuclear plant down. But there's a bunch of heat that's still in that reactor.

They call it decay heat. And you need to keep pumping in that case for that design water to keep it cool. When that tsunami came in, it knocked out the backup diesel generator. That was pumping the water in to keep it cool and when that diesel generator got knocked out, you couldn't pump the water and therefore it overheated and it melted down.

In these new designs, you don't need backup power. You don't need backup diesel gen sets. It self cools using natural forces, meaning thermal expansion, gravity, and air. You don't need any humans to do anything. So let's say this worst case catastrophic event, natural disaster a terrorist attack, you name it.

All the humans are incapacitated. They can't do anything. They're knocked out. Let's say you have no ability to pump anything. You don't have power. There's no power going to the power plant. You can't use motors or pumps or mechanical systems. Everything is disabled and knocked out. It will still self cool in minutes by itself without any intervention using gravity, thermal expansion, and error.

And this has been proven for decades by our national labs and stress tested. So there isn't any radiological release risk to the surrounding community, even in that catastrophic event. Now that doesn't mean that we don't have a lot of work to do to educate people and bring them up to speed because let's be honest.

No one likes anything built in their backyard. NIMBY ism, not in my backyard, is real. I don't care if it's batteries or solar or natural gas or whatever. People don't like anything built around them. So we do have challenges to overcome locally, and we got to educate people. In general, the public cares about two things.

They care about, is it safe? And what about the waste?

Nuclear waste is just used fuel. That's all it is. It is not a green goo Like you've seen on the Simpsons. It's not creating blinky the three eyed fish.

It is just very boring very stable Spent fuel that over 90 percent of the energy is still in it Which is amazing because we can recycle it and turn it into New fuel to put into our powerhouses to create 24 7 clean energy

James Cook: In the future when data centers have to BYOP bring their own power. Brian, you mentioned this business model as power as a service. Tell me a little bit more about that business model.

Brian Gitt: Think about it like this. When E. W. S. First launched cloud services. This was a complete innovation and business model up to that point.

If you were a fortune 500 company or startup company, and you wanted more server capacity, you had to go take a bunch of money and invest that to buy a bunch of servers, hire people to basically rack them and maintain them and operate those servers, ultimately decommissioned and replaced them. It was a huge undertaking of Capital and labor to do that.

Well, a new business model, all of a sudden you could just buy compute as a service and you could tap in to that service offering. And it completely changed an industry. It was not an innovation in technology. It was an innovation in a business model. Oklahoma is doing something similar with this 24, seven clean power resource known as nuclear power up to today.

Or very recently, if you wanted to access it, the utility was the only game in town and they would have had to already have a nuclear power plant built, right? But you couldn't just plunk down a data center and plug in and access 24, seven, clean power as a service. That wasn't an option now it is so now you can contract with oklo in a 20 year power purchase agreement very similar to how you would with a large renewable energy project or other agreement even with the utility and you have no upfront capital investment required to participate you're not shouldering any of the risk Of the energy infrastructure.

Now, of course you are shouldering risk because you're investing hundreds of millions of dollars in it infrastructure and counting on power to be there on a certain day. So I'm not pretending there's zero risk, but there's no risk on the power infrastructure side and no investment required. And this business model is truly unlocking, not only Access to this whole suite of customers that would never even have access to this before that it is also unlocking a completely different licensing strategy on how you get a license for this.

Because we design, we build and we operate and we sell the power. We can go through the nuclear regulatory commission in one step. Traditionally, The way a nuclear technology vendor would go through, it would be a two step process, which elongates the timeframe in terms of getting to market.

And so this gives us an innovation on how we actually get these types of power assets licensed.

James Cook: So let's say I want to build a data center where I want to build it. There's no power. So I'm going to go with the Oklo solution or something similar to it. What is the time horizon? Talk a little bit about the process.

Brian Gitt: Sure. We're looking at from a signed power purchase agreement, which obviously is a binding contractual obligation to commercial operation date of approximately 36 months after the caveat is after this first license is approved because all subsequent license applications to the nuclear regulatory commission, basically piggyback.

On the initial one on that safety case. So they only will look at what is new and different in those subsequent applications. Typically those are local environmental factors. Wetlands, floodplain, endangered species, historic preservation, those types of things, which presumably, if you're building a data center there, you've addressed those already.

Our first commercial units will be deployed in early 2028 and we'll be ramping from there. And so you're going to start seeing these units in the late 2020 starting to really ramp up and start powering these data center applications.

James Cook: And I don't know if we've mentioned it, but Oklo the chairman of the board is Sam Altman with with open AI. So clearly there's a sense that this is like an AI adjacent. Space. Do you feel like a lot of your customers are going to be in that, data centers in the AI world going forward?

Brian Gitt: I'll give you some numbers to back this up. So right now I'm engaged. With Oklo on approximately 5, 000 megawatts of project opportunities were engaged on what that means is we have a customer. We have a site. We have power requirements. We have a timeline. We're going through and negotiating on terms and conditions where that's what I mean by project opportunity and out of those 5, 000 megawatts.

95 percent of them are data center. So yes it, we serve industrial customers and military bases and other applications, but absolutely. And Sam Altman was one of the first investors in Oklahoma 10 years ago. This isn't just a recent kind of, Oh, trending thing. I'm just going to jump on the bandwagon. He saw a decade ago that.

artificial intelligence in power were inextricably linked and that he was doubling down on both.

So yes this is. It's a long game, and as it was seen a long time ago, it's really a critical issue to scaling AI applications.

James Cook: Colm, I want to get the global view from you. Obviously we've been talking about the U S but around the world, there's gotta be a variety of points of view, different countries, different governments. Where do you think adoption is going to be easiest for something like this?

Colm Shorten: It's a great question. I would say where it's going to be easiest is wherever the grid is going to, effectively be overtaken. That demand is going to outstrip capacity also combined with we're finding regulation and that this regulation isn't. Nuclear regulation or SMR related, but it's related to sustainability because we have another, serious problem with the planet with regard to getting to net zero.

And not only is the development and SMR is linked directly to data center, they're also leading the charge And leading the way in the technology roadmap and the path to net zero. we do recognize that. Nuclear energy is a clean technology, and we're focusing on the positives, and the regulation is coming for us out of Europe, and that regulation now is challenging, can, we do this cleaner and better? And we talked about coal fired stations being closed within the near future. We're also seeing that we are going to run out of fossil fuels. I'm in the Middle East, and there's a big focus here on renewable energy. We know that wind has an advantage. You can bring it to the table. Solar can, but sometimes the wind doesn't blow, and solar is not that good at nighttime.

So what we are seeing is the regulation is driving it in that side. And when we look at the environmental impact of nuclear, it's a positive story. So that's happening. Then outside of Europe, it's also clear that we've seen announcements even in the last 12 months, India, Singapore, even the Nordic countries like Sweden are starting to invest in these technologies also because we know how long it takes to build, as Brian alluded to earlier on, to build grit, to build generation or transmission or distribution, it's really difficult and if you're going to build it, you shouldn't be building brown or black power.

You've got to be building green power because otherwise the cure will be worse than the current symptom that we have because we have to be sustainable that we can't come up with a solution that's going to be really good now to solve just data center requirement at the expense of the planet.

And we're going to miss these goals, and that's why it's pretty positive and amazing that you do see Amazon, Google, Microsoft, as well as these other players like Oklo, starting to drive these technologies. We have to get away from the fear factor and we have to understand it. And pretty soon we're going to be talking about something even more fundamental like water.

And if we continue to generate power the way we are. We're going to have trouble with water. Now, the one thing about power is we can lean into nuclear and find a solution. Solving the water problem is very difficult. we can dig a very large hole or crater, but it's very hard to make a lake. I think in that sense, because of the water stress that we're seeing on the planet, nuclear, the acceleration of that, and the challenge I don't think will be the technology, it will be the regulation.

And opening the hearts and minds, winning over the hearts and minds, but we're now seeing a systemic move in that direction, countries and areas that wouldn't have looked at it in the past, whether it's in India or whether it's in Southeast Asia, they're all exploring now nuclear as a real option, whereas it was looked at in the past of it's out there, it's way down there, it's blue sky tanking.

James Cook: Before I wrap it up I want to ask both of you Colm or Brian, is there a question I haven't asked you that you think I should have, something interesting I haven't asked you about.

Colm Shorten: I think it's not just the demand from a data center standpoint, even though we're the biggest customer and maybe it's because they're cash rich. But if you look at other innovative technologies, whether it's 5G or electric autonomous cars, that's also bringing in additional pressure. And these are technologies that people want to use rather than, even though our data center audience are really valuable and dear to me, it's reaching the masses to figure out, are we solving a business problem or a need that they have?

Because unless they have it, they won't have an incompelling reason either to act, or to IE act positively or buy into this solution.

James Cook: Gotcha, Brian, anything I haven't asked you about that you think would be a value to the audience.

Brian Gitt: When we look through the lens of environmental sustainability, I think we need to look beyond just emissions reductions, and ultimately the fundamental metric that determines the environmental impact of any energy technology is the materials intensity. Of that energy. And what I mean by that is how much stuff do you got to dig out and mine and process to make something to basically produce a product, whether it's a solar panel or a wind turbine or a nuclear power plant or a coal plant, ultimately, it's the materials intensity of that technology that determines it.

And when we look through the lens of, Okay. Total life cycle, environmental costs, nuclear energy provides the lowest greenhouse gas emissions, the lowest materials use, and the lowest land use all energy technologies. And so it really shines through that energy environmental sustainability lens when it's being compared in that concept.

To illustrate this, imagine a golf ball. That golf ball of uranium metal could power your entire lifetime from when you were born to where if you lived a normal lifespan into your, let's say eighties that you would die your entire lifetime.

That's how energy dense uranium and nuclear fuel is and therefore that's why it produces the fewest amount of materials and requires the least amount of land and the least amount of greenhouse gas emissions.

Colm Shorten: Yeah, I think that's a great analogy. When you take the periodic table, which we studied when we were in college, and we look at the materials that we're using, either in data centers or digital technology, it's gone from. 30 years ago, where we used about six elements on the periodic table, it's now 42.

And then we look at batteries, whether we're going to lead or lithium, and it's wonderful for lithium. And then we do a comparison between a diesel car and electric vehicle, and we realize, hold on a second, to get 80 kilos of lithium to get an electric car battery, what have we got to do? So when we get to these SMRs, if I look at the technology roadmap within 20 years, we're going to be powering vehicles with this stuff because it's a clean technology once we embrace it and get over the fear factor.

James Cook: So Brian, how often do people ask you about Mr. Fusion on a back to the future on everybody's car?

Brian Gitt: I think fusion holds tremendous potential and I think it's absolutely worthwhile as a research and development project. And I'm really happy and excited that both the government and the private sector are investing lots of money into it, but it is not something that is yet proven or commercialized and we are focused on vision.

So obviously I'm biased. I work at a vision power company. We have a technology already. And I hope fusion comes online. Believe me I very much hope that comes to fruition, but we already have a technology that has 70 years track record of working, of scaling, of providing tremendous 24 seven clean energy.

And when we actually look at how much. Power could we provide and over what timeframe there's enough nuclear fuel right now, just in the earth or in through recycling, we could power it. Our current power consumption for 4 billion years. So we have 4 billion years’ worth of energy just using the amount of nuclear fuel that's available on the earth today. And at that point, the sun will destroy the earth. So basically, we have for practical purposes, limitless power today without fusion. So I think why not go all in on that?

Colm Shorten: Yeah, I, and I will respectively disagree because I'm agnostic and I'm just a dumb engineer. My view is we got to look at all the options and the beauty of it is. We know that the direction of travel is that way. It's nuclear. Do we know whether it's fission or fusion? Neither of them are commercially viable. We got to work on it. It's going to be top dollar before it becomes commercial, but there's

Brian Gitt: Fission is commercially viable. It's a whole industry for 70 years.

Colm Shorten: Okay, I'll take that point on board and again, respectively disagree, but when we bring the concept of Yeah, it's difficult to control versus it's easier to control vision. But when we look at the waste, one is radioactive and the other isn't. When we look at the occurrence, one is artificial, the other is natural. My view on it is we shouldn't rule these options out. And it's not turkeys voting for Christmas, it's trying to figure out which one fits. And there may even be something else. We're starting to see, ceramic pebbles and there's other technology. What is happening, which is beautiful, is we're opening our mind.

So when we don't have that tunnel vision, and the answer will probably be somewhere in between. I do agree that the Trident test, the 1950s, vision is definitely where we are starting point, but it may not be our end point. We're beginning on a journey, but we know where we're heading. And I think what we're doing is once we get rid of the fear factor, not having the waste or if the only waste is hydrogen, that could be good.

What is the answer? We don't know. And right now, as I said, yeah, it's commercial at the large scale. How many of them do we have operating today? None with those two exceptions in China and Russia. How many are we going to have in 2030? A lot. Why? Because we have no option.

James Cook: Thank you so much for joining me today. Both of you, this has been a fascinating conversation and that's why I host this show is because I learn about. Awesome topics like this from experts like you. So thank you so much for joining me today.

Colm Shorten: Welcome.

Brian Gitt: Thank you. Appreciate the opportunity.