XNRGY Founder and CEO Wais Jalali on Data Centers and Power Innovation

I'm really, really excited. Very excited. It's morning, little fall weather in New York City, my favorite time. I'm in New York today and I'm here with my dear friend Weiss Jalali, founder and CEO of X Energy. Weiss, welcome.

Thank you, Amram. Glad to be here.

I want to talk about X Energy, the data center. All this AI infrastructure has been built. Everybody's building data center. But let's start with you. Tell us about yourself. You are an immigrant. You came to Canada at first? I think so. So tell us about your story. How did you get started?

Well, I mean I started as an engineering student, but I've been in thermal management all my life, since 1992. So engineering school and then I started into this business in different fashions, sales as well as designing data centers and healthcare facilities for about 30 years.

Thirty years ago, data center was not hot. Why did you get into thermal management?

Well, the data center was still. It wasn't hot, but there were still AOL days since I lived in the D.C. area. There was a lot of national security projects for data centers. But at that time, 1 megawatt data center was a pretty big deal. And now as you can see, we're into the gigawatt era, so.

So world changed a lot in last 30 years. So tell us about how did you come up with X Energy? When did you start the X Energy?

2017. I was at the World Economic Forum in Davos and then I learned that 53 gigawatts of deployment was going to happen. So the density was going out through a roof. And since my background was in fluid dynamics and if I could change the friction loss, which basically impacts the motor, impacts the power, it would be very beneficial not only from a power usage effectiveness of a data center, but also for the environment as well. I called a lot of people that I knew in the past from the engineering and manufacturing community. I told them if they want to put their skates on one more time to make an impact. And it started really in Montreal because Montreal is sort of a H Vac hub. There's a lot of companies there. And then McGill and Concordia also teach H Vac classes on thermal management. So it's always been in North America, the center for where all the manufacturing goes. I gathered a team of who's who. Then I went to Oregon, met with Larry Hopkins, who's the guru on fan impeller technology. That's how it started. We call ourselves Oceans 18 because there's 18 of us and all 18. Actually the average experience of all 18 of us is about 30, 35 years of experience.

So how did you get the first initial capital? Is it a capital intensive business?

It's a very capital intensive business. But when I started this I thought, you know, as long as I build 100,000 square feet, it'll be okay. It was my money at first. And then I, one of my friends, they used to be number three at Lone Star, came in as well, Mike Thompson. But then they just exploded and we had to raise private equity money which we did in 2022 I would say. And then in a span of from 2022 to now. We're at a million square feet now. So more people just came in. We just had a press release last week that D Car Partners, which is a JV between BlackRock and Tomasa came in as well. As well as Activate capital and climate investment. We've raised a lot of money for, for the expansion for growth.

So it's a very capital intensive business.

It is, but I think we're doing really well. I think our revenue versus investment is a three to one ratio, which is great.

I want to dig into more on X Energy and what problem and what impact. Obviously you already talked about some of the impacts, but let's talk about all these data centers.

Right.

OpenAI is building data center hyperscalers building data centers. So what are the biggest problem that you saw that you thought that X Synergy can solve for these companies?

Well, there, I mean there are multitude of issues. My chief technology officer and I, we've been on the other side, we've been on the owner side. So we were sort of not being a vendor, we're being more of a partner where we de risking the client from all sorts of issues. And reliability is the biggest thing in the data center for the GPUs and the CPUs. You might have read some articles three years ago where data centers went offline in certain climates where you have high ambient temperatures. So we, when we tackle this, there was multitude of issues for data centers. One is to make a very efficient data center. The powers everything. So the mechanical electrical of a data center is literally 75 to 80% of a data center. So you got to look into vertical integration because right now if you look at our products, which is comprised of really three, four different elements on the mechanical side, which is the chiller that produces the water and then basically it flows to all of the air units that basically you would have for the transfer rooms, the fire closets and then obviously what's called the cdu. Which is a cooling distribution unit. Because there's a huge transformation happening right now going from air to liquid cooling. And liquid cooling is really taken over, but you're still going to have 25% air and 75%. The biggest problem, and that's the DNA of the company really, is that we don't use any water. So because water is very scarce, probably.

Good for environment also very.

And also we are, we're solving the ambient problem. So what happens with these data centers, if you look at some of them in Texas, for example, they would have 10 buildings and each building would have like 100 of these machines. And in Texas, the ambient temperature is called a plume effect. So what happens is high plume effect. So the temperature of some of these things could range from 130 to 140 degrees on the rough. Okay. And what happens with existing technology is some, some legacy incumbent. The competitors that we have, they're utilizing technologies that were made for 104 degrees. So we basically joined forces with Copeland, which is owned by Blackstone. We co develop on the control side. The algorithm was a machine that can withstand very high ambience. We just tested it for one of our biggest clients in Toronto. We pushed it to 140 degrees and it did not fail. Wow. So that's a huge problem that we're solving and getting traction with some of the clients. So it's a game changer really in the industry, especially for AI because the density is going, even going to go higher because I think Nvidia is going to come out with more chips that the racks will become higher density as well.

Let's say you're talking to OpenAI or X or XAI or doesn't matter, anybody. And you are having one minute to convince that they have to take X energy. What do you tell them?

I give them reliable cooling at the least amount of power.

And why can you do that that other people can do it?

It's just our innovation site on the air side where we have developed gadgets, or if you want to call it, if you want to dive deep into it. It's an oval tube coil as well as our impeller design on the fans. So it's got the least amount of followed amps on the motor. What happens is every time you change the geometry of air, you lose pressure. We utilize the same concept. What do you call boundary layer theory or Bernoulli's equation of fluids? This is what Bombardier, Boeing, Airbus do when they design their planes. It's the same thing. It's how do you get the air from point A to point B at the least amount of power for the motor. And as you know, motor is half of the world's energy. And some of these data centers, by the way, if they're 100 megawatts or higher, they will have thousands of motors on one campus. Now on the liquid side we also are going to have a basically an oil free compressor that basically is going to have the least amount of power to develop. So I mean power as you know, you've been reading all these articles before everybody was staying close to the fiber, close to the cities, but now you're getting data centers in Nebraska and Iowa Dakotas. Wherever there's power, that's where everybody's going. And obviously the, the grid is over ramped and some of the data center manufacturers are building micro grids based on clean energy as well as battery storage. So you're seeing a lot of that. One is being developed by Digital Bridge in Nevada right now if somebody asked me, I don't use any water and I will give them at least amount of power. Now what does that do? That saves the money on electrical infrastructure, saves the money on generators, motor control centers, transformers, wiring, installation. We don't only have an operating impact as your total cost of ownership, but we have an impact on the capex immediately.

So yeah, that makes sense. You're not using water and you're also reducing power consumption. So you kind of touched on it. You like to talk about Brunello principle, a key concept for fluid dynamics. So talk about it. What does that equation mean and how that fits into x energy?

If you look at the logo of Zenergy, it's multicolors and the logo is basically velocity ISO vectors in the airstream. So what we do is that we have the least amount of contraction and expansion of air to make it in layman's terms, we deliver the cooling at the least amount of power because we're reducing the friction if we go point A to point B. So that's what whether you call it boundary layer theory, you call it Bernoulli's equation of fluids. On my staff I think we probably have over 70 patents on this principle on air and liquid.

How did you build all these things so quickly? The company started in 2017, I guess eight years.

Well actually to be honest with you, we've gone to 10 digits in a span of two years. So I mean really the first, first systems because you know, Covid came right. So when we started the company and we built a greenfield in Montreal on what I call Zenergy 1. Now we're up to Zenergy 4. We're going to Synergy 5. And the plan side, basically it came together because a lot of people joined because the company's motto was to build smart solutions for what the environment was giving us. And by the way, the data center world is changing every day so we have to stay ahead on the innovation side. We started in life science and healthcare because you don't just go to one of these hyperscalers. I'm not going to mention their names but you know, everybody knows there's only four or five. But you don't go into them, say, look, I'm a startup and I want to provide all the cooling solutions you need for your data center. They're going to ask you, okay, let me see your balance sheet, let me see how many installations you have. Let me talk to some of your customers. So based on my experience in the past, I knew a lot of people in the healthcare and life science and semiconductor space. So they gave us the chance because they knew that what we did before with them and then we got into three years later, which is really 2024. Last year is really when we made our wins with two of the hyperscalers because they went through a three year audit in getting us certified as one of three vendors worldwide. For them, the other two guys, I'm not going to mention their names but you know their market cap is 100 billion so. And where the David versus the Goliath right now coming up so. But what they saw that the other guys could not do is our innovation side and the problems were solving. In order to succeed in life, you have to solve somebody's problem, you have to have a solution. You can't just be a me too company saying I'm going to compete against a hundred billion dollar company but you know, you should trust me, I'm going to do the job. So. But if you demonstrate that you've done it before for the last 30 years on multiple levels, that gets you over the top. Because everybody asked me the question, it's a great question of why you. Right, right. These guys do this.

Yeah, they've been doing it and they have a lot of money, they have a lot of more people and then.

They have a lot of R D.

Departments, they have a lot of relationships.

Relationships, exactly.

So how are you winning?

We're winning because we're solving the problem. That's it. And, and they, our products are completely more innovative than anybody else's. And just what I said on the, on the Development of the chiller that we have right now, which is really the heart of any data center because it's the first thing everybody orders the order the transformers, the motor control center, the switch gear to get the power in and they order the chiller to get the cooling in. So I think that that development is a game changer for industry and I think all of our competitors are going to go to that machine. But I have a four year hit start on them. So because it's been four years and that we've been toying with this thing.

So what happened after four years?

What happens? We have to come up with new, new innovation. We just unveiled our state of the art Mesa 1 facility in Arizona. We're also going to go to Mesa 2, which will be online here in January. In a few months, they're building a state of the art thermal lab where we're going to push the envelope again on new technologies to stay with the density demand on the, on the data centers.

And when you're talking about these facilities, these are facilities of the. Your customers or these are your facilities?

No, these are our facilities. These are facilities, yeah. So we're going to put a state of the art R and D plant.

Your customers are putting GPUs there or.

We'Re going to put GPUs there just to test. But we're also going to test our technologies on how it performs.

Got it.

Whether it's airside or liquid side. Both. And CDUs. Because right now you saw the two acquisitions by Schneider and KKR into the CDU field. At that point, even Last year, the CDUs were the biggest everybody was asking for was like a 1.4 and 2 megawatt CDUs. Now customers are asking us for 8 megawatt CDUs. So it's just changing, evolving rapidly.

So the industry is changing dramatically, right?

Absolutely.

Like, how do you keep up with it? Again, you're a small company. How many employees do you have now?

Last time I checked, we have over 500 employees. Okay.

Grew from 80 to 500.

Yeah, exactly. God, think next year we're probably going to be over a thousand and then it's going to keep going up because I mean, our backlog right now is significant. We're sold out for 2026.

You're sold out for 2026.

So we're taking, we're taking 27. Yeah, exactly.

Obviously one of the biggest thing that people like me are trying to figure out, you know, everybody's figuring out, right? Media. I was like, look at New York Times. This morning they're talking about, you know, is it an AI bubble? The amount of build out going on? Do you think it slows down anytime soon?

Absolutely not. I mean if you read all the reports, although I go with, I'm in the middle of it. So I'm seeing my customers tell me I just signed on 15 gigawatts of leases with all the clients. So 15 gigawatt equates to about 15 billion for energy. Just so you kind of know the numbers. That's just one client by the way and we're working with like 50 clients.

So it's not only hyperscalers, there's a.

Lot of COLA providers and there's two developers that are doing a lot right now in the industry core, we've been one of them. Most of the hyperscalers are hiring certain developers and they're also building their own data centers too. But it's like an arms race right now who gets there first. So the speed deployment and that's why our waterless, I mean we don't use any water. So the speed for our air cooled machines, you build a data center a lot quicker with air cooled versus the other way to do it would be what's called a central utility plant where you would have large centrifugal water cooled chillers. Then you have to build a building, then you have more piping. So speed is really the name of the game right now for some of these guys and they want to get to building a 200 megawatt data center in six months. So that's how quickly it's moving. I don't think it's going anywhere anytime soon in my opinion. I think Sam Altman just stated at a meeting that just for his company, I think it was 30 gigawatts just for, for, for you to deploy 30 gigawatts right now will take you five years in in my humble opinion where AI is going, I don't see anything happening until 2032 at least. Wow. That's my humble.

That's your humble opinion?

Bullish opinion.

Where are all these companies going to get the money from? I know that's not to do with X Energy, but do you worry about the credit? Like if you look at. One of the most interesting thing is that Oracle for example, their CDS spread is going up quite a bit and a lot of people are concerned that companies like OpenAI will not be able to pay them because they don't have the money. Right. You know they say 30 gigawatts but it costs a lot of money, where they're going to get the money from.

That's a lot.

Yeah. So again, you cannot comment about your customers and I'm not going to ask you, but do you worry about the credit quality of your counterparts as you're doing it?

We do, but I mean, if you look at the balance sheets of some of these companies, with respect to synergies concerned, we have lots of strategies where we get some payments upfront. You know, when we ship it, we get. We get the rest of the payments. So our customers have been pretty good on the payment cycle. But yeah, I mean, look, 30 gigawatts, I mean, just us being 1 billion per gigawatt. So if you build the whole data center, you're talking 8 to $10 billion per data center. So it's massive.

8, 10 per gigawatt?

Yes.

And that's so 30 would be like 240. Correct. Which makes sense. That's the kind of contract they're signing with Oracle.

That's right. That does not include the GPUs.

Right.

That's just the construction. It is incredible. Yeah, because we're building one right now in Texas and I think each building is like 8 to 10 billion or something like that. But you know, SoftBank has come in, Microsoft. I think Oracle just invested all of these companies. Oh, yeah, look, I mean, they're. But you also have to look at construction people too. Right. Getting to more into modularity to save labor and. Because labor is also a challenge. Yeah, labor. You're talking about long lead items. So yeah, those are all challenges. Power, I mean, everything. It's a bullish thought process.

But look, so what does that mean for the economy? So there'll be a lot of new.

Jobs will be created, a lot of plumbers and electricians. We don't have. That's not enough. And also tax too. Because look, I mean, when you're deploying this massive amount of equipment, you need text to service it to.

Right.

We have one project in Texas that we had to like deploy a dedicated team to live there, as well as to build a complete parts warehouse near this site, which is in the middle of nowhere. Because just for them to service the equipment, you're talking 800 to a thousand machines per data center. These things are not small. Right. So they're like 50ft long and 10ft wide and. And they have lots of moving parts. In a data center, you can't be down. Then you have what's called buffer tanks. Because when we're designing data centers, if there's a Power failure or something goes wrong, you need like a 15 minute ride through on the cooling cycle because those GPUs, if you don't cool them immediately and some of them spike between 20 to 80% in a second, you know how much those Nvidia chips cost. So I mean they just fries the whole rack. So there's a lot that goes into reliability and maintenance of these facilities as well. One thing that keeps me up at night is, is how many texts do we need to have per site if something, you know, happens? Basically making sure that all those machines stay online. So the reliability is number 1, 2, 3, 4 and 5. In a data center they go to all sorts of commissioning, different levels of commissioning before the white space is handed off to the local people. Now, data centers does not have a lot of people that work in it, right? Few. And plus the temperatures are 89 degrees, but you need an army of people to build them.

Let's talk about leaderships, right? Started with the ocean, 18, 18 people, now 500, going to thousand. And all these problems. Very, very, you know, sophisticated customers who have, I'm assuming they're very, very demanding and they're paying a big dollar.

Oh yeah.

How do you manage the team?

You have the C suite, then you have the middle management. We have a lot of engineers. I mean, I think one thing that we invested in is engineering. But you also, as a, as a CEO of the company, you also have to have the ability to take some risk, calculated risk, but you have to be nimble enough to take some risks. We have a lot of tools that a lot of our competition doesn't have. Like we can do internal computational fluid dynamic modeling to tell them where the issue is going to occur. We try to partner with all our customers, but I think one thing, one thing that they're all seeing is that we're de risking them by being a partner and not a vendor. We're not selling just some equipment for them, but we're selling our 30 years of knowledge of what not to do and in a data center. And the reason is because we've had some pain in 30 years, most of us, of seeing issues that occur. I'm not going to mention the customer, SEMI Large Semiconductor 2003, they lost thousands of motors and all the manufacturing production was down. And then we found the issue and there was basically a ceramic bearing that you had to put in the motor because what was happening is the frequency controls were arcing, going to the bearing and burning the batteries out. Now if you look at a facility that lost 3,000 motors not once, but twice. It's not an easy thing just to take a motor out and put another motor in. So now fast forward 22 years. Every one of our motor that goes out has that ceramic bearing in it just because we don't want issues on projects. So that's just one example.

So one example. So you have learned a lot from your mistakes.

What not to do.

The amount of growth is happening from some of these companies building data centers. You know, the exponential growth, right. You look at their capex numbers, do you think these guys know what they're doing? Like, do they have the know how of building these things?

Everybody's outstretched. Look at all the engineering firms, all the architectural firms, they're outstretched. Everybody's going at warp speed. And, and once app and goes warp speed, quality control becomes an issue.

Yeah, I'm really worried about that.

Gotcha. Everybody asked me that question when they, when I get interviewed is what keeps you up at night? It's quality control. Because when the demand is 30x the supply, right. You're basically trying to capitalize or you're trying to get so much revenue for the square footage you have in your facility that you might skip some steps. And it is happening in our industry. There's a lot, I'm sure it's all issues because I run a company.

I know when you move fast, people.

Just shipping stuff to the project sites and, and then they have to fix it afterwards. And also engineers are so outstretched that there's, you know, sometimes you might not think of a certain thing that because you're so busy to try to deploy so many projects at one time. And I think that's one advantage to us is that we have another set of eyes not only on the equipment side, but also on the design side. We want to, you know, as far as calculating loads, what happens, you know, if you have transient issues, if you have fluid issues. I think we're getting the market's attention because we're also another set of eyes here that basically said, nope, don't do this. You're going to have an issue here and not worry about just me selling them 10,000 more systems. So I think the biggest worry right now is to have a quality control job on time and on budget.

Yeah, and quality control for your own company as well.

You just shoot for us for quality control. Our CEO comes from Bombardier. When I interviewed him, the two, the one thing that was won me over with them was quality control. He said if I had, if he has to be late, he wants to be late, but have the quality unit out of the box. And I think that's our DNA too.

So how do you talk to your customers? Because I think one of Silicon Valley has this culture that, hey, run fast and break things. But you know, if you build a 8 billion dollar data centers that, you know, fixing that, it's much different than fixing an algorithm.

Yes, it is. Because I mean, the deployment, especially the call guys, because they're leasing the space to the hyperscalers, it's on the fly. I mean, basically everybody's running on the fly to try to fix things. But yeah, look, you're gonna have to take some risks and you might break things in the process. But at the end of the day, I think if you're 9 out of 10, we like to be 10 out of 10 for, for all systems, because then you have an entire team on a project site trying to fix stuff and then basically it creates a whole domino effect. Everything else, we have those issues. Our competitors have those issues. I'm sure the contractors have those issues. The suppliers of electrical equipment have those issues, the engineers have the issues. Because we're all being pushed to limits that we've never, I've never seen anything like this in my entire career.

Right.

This is like a new industrial revolution. And 50% of it's being built in the United States on a very massive scale. And if you walked one of these data centers, it's like almost a half mile long. It's not like a, a common building that you see out there. You just keep walking and walking. You need a golf cart almost so massive. And these pipes are like, you know, two and a half feet diameter. I mean, that the water is going through. It's something to see. Everybody should walk a data center just to kind of see that how big one of it is. So. And currently we're working on 58 of them.

58 of data centers?

Yes.

Wow. Different 500 people.

Different sizes, though.

Different.

Different sizes.

Different sizes.

We have, some of them are 20 megawatt, some of them are 200 megawatt, and some of them are 1 gigawatt, which is comprised of a whole family of buildings on a project site.

As we finish up, let's talk about this leadership thing a little bit. If you have to tell, you know, wise eight years ago, what are the two or three key advice you would give yourself that you learned over the last seven, eight years from your mistakes or from your experience?

One is, I can't do everything by myself. That's number one. I have to Surround myself with a lot of talented people in different spaces. Because, you know, when you're in your 20s, you think you can do everything. And then when you get 30s and 40s, I'm over 50 now. You started. You started saying, okay, you know, you need all these guys and different departments to basically make a whole team. And that whole team will make certain decisions all the time. But the vision has to be with. With, you know, there's a. There's a vision to execute. The customer is everything to us. We're very conscientious on the customer side. And you live and die with problems with customers, too. So we don't alienate the customers. There's three philosophies I've always had in life. Friends do business with friends. You work hard, and you also have passion and fun with. With working hard. Because I tell all my employees that if they want to get out of the bed in the morning and they don't want to go to work, this is not a place for them at all. And three is you solve problems and then you discuss whose fault it was later. Is a big differentiation in this company. Some of our competitors would lose a client over $4,000, and basically they're building hundreds of millions of dollars worth of contracts, and they will lose them over four.

That's the problem. The big companies like an accountant sitting there.

That's right. Well, the secret sauce of this company is that an engineer is running it. When an engineer runs it, it's very different, I think, then when, you know, when everybody's looking at the bottom line. I just thought philosophy is that if you treat clients in a great way, you will win in the long run. And you might not win in the short run, but you're going to win in the long run. In long run, thinking is really where everybody should be at. See what's going to happen two years from now or three years from now and plan for it.

So here three things that you said. One was, you cannot do it all yourself, so you need a team. Second, you need to set the vision as a founder. And third, you need to focus on the customers.

I think Jeff Bezos said it best. You have to have customer obsession, right? The customer is everything, and we need to adapt or come out with designs that solves the customer's problems instead of reacting the other way and just saying, oh, this guy's doing this, so we'll do the same thing. But if you think outside of a box, which we do, it's intellectual curiosity, really, in terms of. Because look, all the information is out there for everybody to use is how you compile the information into some meaningful purpose and solve the client's issue at the same time.

Speaking of customers, right? Customers can be sometimes unreasonable. Customers sometimes could be wrong. You know, because you have the 30 years experience. They don't necessarily all the time. So how do you deal with those kind of difficult customers?

Well, we don't. Look, I think the luxury everybody has these days is that you can pick and choose who you want to work with because there's so much demand is very high. Demand is very high. It's getting better. 30 years ago if you had a facility director on a project and the guy was just basically said nope, I've done it this way for 20 years, I'm not going to do it any other way. And then you just have to walk away from that, that's it, there's nothing you can do. I think there's a lot of young engineers now that are working for these big companies that have more of an open minded frame into collaborating with the manufacturer to, to solve issues. I think from a delivery standpoint, I think everybody's being unreasonable because everybody wants everything yesterday arriving on the data center site. The two things for manufacturer is people and infrastructure. In terms of how many square feet do you have? You cannot build 3,000 machines from 250,000 square feet. You need 750,000 square feet. You need more. The only unreasonableness that comes from customers right now is just how soon they can get their product. Because we are, I mean without us there's no data center. I mean a lot of people ask well why is thermal management so important in a data center? Well, your GPU scan run, there's no information. So it's, we're an essential part of the data center. Also on the electrical side, I mean those are the two. Other than that it's just a tilt concrete building with 75 to 85% of it is electrical and mechanical.

This is amazing. You are in the middle of one of the biggest, you know, revolution that's happening. So thank you for taking the time in middle of that and talk to us. I appreciate it.

Well, thank you Imran, thanks for being here.

All right, bye.

Thank you.