Thank you all for joining us for this presentation at the Southwest Ideas Conference hosted by Three Part Advisors. Our next company is Solaris Energy Infrastructure, trading on the New York Stock Exchange under the symbol SEI. And here presenting on behalf of the company is Kyle Ramachandran, who's President and CFO. And we have Chairman and CEO Bill Zartler. Bill?
Good morning, everyone. Thank you for being here. Kyle's here. I don't know how we'll figure out how to talk in the mic or we'll talk real loud and get through this. So we have about an hour and a half presentation that we're going to get through in 15 minutes and answer questions. So with that, Solaris, we took public in 2017. Our original business was sand handling equipment at well sites that was really growing as U.S. frac took off . We quickly gained about 35% market share in 2018, and it's been there ever since. We've added additional equipment into that offering to double our margin per frac fleet. So as we've seen frac fleets dwindle down over the last couple of years, we've actually been able to hold flat and actually see our numbers and hold true on the return on capital.
With that, we've been looking around in the company about what's the next leg of the business stool for Solaris for a number of years. Our equipment started out all electric, so we were running 30 MW of our own diesel generation, saw the trends, worked with all the electric frac providers as that came on, and recognized that there's an opportunity in that part of the business to really add capital and transform our business into the mobile power generation business.
With that, in this summer, we announced and we closed in September the acquisition of a company called Mobile Energy Rentals, who had a fleet of 100-ish MW, plus or minus, of mobile turbines who we had actually rented from before, knew the company well, saw the opportunity both in the oil field and more specifically, really to rapidly grow it to support a couple of data centers, AI generative data centers where the power loads are pretty tremendous, and placed an order with Solar Turbines Division at Caterpillar for another 350 MW of power generation, of which we've started to take delivery of and put all that equipment to work, so kind of the key highlights of the business, we've found a new market for the equipment in the business, and that continues to grow.
We have lots of opportunity to acquire additional equipment both in the generation side and the electrical side of that business. So not only are we supplying a turbine, that turbine is useless until you have the right kind of transformers, switchgear, breakers, cabling, all that ancillary equipment, balance of plant equipment that goes with making that turbine deliver power as a service, and we've got a very high level, high touch level of providing power as a service to our customers. We've been doing this a long time, solving bottlenecks for the industry successfully and turning ideas and businesses into companies is kind of what we've done and been in our DNA for a long time now.
The business hasn't caught up yet on a valuation perspective to where we think it will trade and should trade, not being a solely U.S. completions-based service company into the power leg of the stool. We don't believe it has been fairly valued yet. And so a lot of opportunity relative to where others are trading in that space. And as owners, so the management team and our original shareholders still own over 50% of the company, pretty motivated to create shareholder value and maintain a discipline about what we do from an equity perspective.
Sure. Yeah. So as Bill alluded to, on a historical basis, our business is 100% focused on Lower 48 completions. Through this acquisition, we are significantly transforming the business into actually being more predominantly driven by power generation demand. And so at closing, we were roughly 60/40 with the legacy business. Going forward, today it's roughly 50/50. And as we look at sort of the exit rate, as we build out the new fleet, Bill alluded to it, but today we've got roughly 230 MW of owned capacity, and that's going to 535 MW by the third quarter of next year. So as that fleet gets rolled out and deployed with customers, we'll see the continued transformation as far as the earnings profile. As we look at sort of the market story, I don't think we need to beat a dead horse here, but power demand is growing.
Broadly speaking, in the U.S., the grid is aging both from a generation standpoint as well as an infrastructure transmission standpoint. We've seen generation capacity come offline. We've seen projects delayed, and that's being mixed in with demand, particularly in these large Generative AI data center installations that are growing at just incredibly rapid rates, and we're talking about adding small to medium-sized cities every time a new Generative AI data center is built, so today we're working for a customer where they've got roughly 1 million sq ft of a building, and within that building, they've got a demand of between 400-500 MW. And so that can be a medium to large-sized combined cycle plant supporting the grid, well, the grid really doesn't have that spare capacity.
So we've gone in and installed a series of mobile turbines in parallel to provide that demand within the span of, call it, 30 days. So it's incredibly rapid deployments. You need natural gas supply, but we've got plenty of natural gas in the U.S. We can produce it quite cheaply. And when we look at the all-in economics, we're actually quite efficient relative to the grid on a $ per kW basis. We're not saddled with all of the transmission, all the overhead, et cetera, all of the renewables that have gone into the grid that are being absorbed across all of the base load capacity. So at closing of the transaction with Mobile Energy Rentals business, we had a set of contracts that were, I would call it, short-term in nature, on average roughly six to nine months. We closed the transaction in September.
We announced it in July. So between July and today, we've effectively contracted the vast majority of our pro forma fleet at anywhere between two and four years. It's a mix of data center applications as well as the energy business. So in the Permian Basin, there's a huge call on power demand as we have a continued focus on electrification of the Permian Basin to support primarily production, gas processing type of facilities. Today and going forward, we don't have any of our power demand supporting the frac market, but rather more on the production side. I don't know that there's a ton to hit here. This is just a visualization of the fleet at announcement. It was 120 MW . Again, between the fleet we've added to date as well as firm deliveries with OEMs, we've got a capacity of 535 MW by July of next year.
And beyond that, just want to go back one. Just beyond that, that's where we are today. The supply chain is incredibly tight. We've got our eyes on what does the next leg of growth look like for this business. We look at the supply chain. It's maybe on the inside date called 12 months out before you can secure additional capacity. There's some inorganic opportunities to grow the business through mergers to gain additional capacity. However, most of that capacity is likely working. So there's really not a ton of capacity available. So we've got our eyes squarely on how do we handle the next sort of leg of growth here through additional orders to support the market's demand.
This is a visual representation really of what it is to build a behind-the-meter power plant effectively for the data centers. And so if you think about the generators, that's one piece of the puzzle, adding the transformers of voltage, pairing that up with the grid potentially for certain levels of power to supply the voltage. And generally, depending on the type of data center, you're fulfilling two different loads. One is the load to run the chips themselves at a unique voltage. And the other is to run a massive cooling system because these generate a ton of heat. So you're sitting there with effectively a massive exothermic type building that if you were to turn that cooling out, it would melt the chips down within hours. And so it's very, very heat intensive.
The chips refer to NVIDIA's name on them, but they're built in with water flow lines built in and around the chips and the data centers. And so you've got to keep that flowing or it doesn't work and it effectively ruins your billions of dollars of investment. So this is critical power for these data centers to load to build. And one key point to that is if they were going to rely on grid power, most of the time they would have to have some sort of backup generation capacity on site anyway to ensure that they could run if the grid went down enough to keep the cooling load going.
So when you have a behind-the-meter plant and you're looking at your only critical item, your gas supply. A little gas storage buffer there with some CNG provides you the same units can be your backup as others because you've got enough capacity on there and enough flexibility to rely on. Have enough gas there to rely on that equipment as your backup and your primary source. And so as we look at the fleet, I think we sort of covered this, which is that we're used to building a service organization. Our company has always been there about reliability. The increases in efficiency and frac activity and frac fleet have been driven by companies like ours, ensuring that the raw material supply chain is there 24/7, the buffer's there, and your critical equipment has redundancy built into it or ways to ensure that our customers see uptime.
One of our daily statistics with our entire management team is reliability, and we're running 99.97% reliability today with the sand equipment on site in terms of frac activity, so our goals and the cultural fit between the power business and what we do on a day-to-day basis really is a really good fit, rapidly deploying, rapidly hooking up where needed, and ensuring that our customers don't see downtime driven by us. Just a little history. I mentioned earlier, we grew the business throughout the team. We've had 30%-35% market share ever since with great uptime performance and a small piece of the cost, as in the data center's power is a small piece relative to the cost of the chips.
This is our equipment, the Top Fill system that we mentioned adding new capital to it that's expanded our touch with our customers and given them more flexibility on the trucking supply options of sand. So if we look at, we mentioned sort of where we're trading and what our historical return on capital has been. We've had a fantastic little niche business for the last, we started in 2014 with great returns on capital. We've been able to return significant cash. I think it's in the $185+ million range plus or minus over the last several years, both with stock buybacks and dividends and continue to see a focus on finding and adding this mobile power business into here with similar looking return on capital.
And you couple that focus with that, which is where we're trading today, is sort of the combination of really high return on capital and relatively low trading liquidity, partially because we've been U.S. land focused and it hadn't been the most sexiest thing to invest in with power business combined into that business that gives us a whole different look. This is on our website, so this presentation is available publicly. We back up to that and then kind of talk about what we see and how this bridges ourselves going forward with the new plant, with the new power business. We see it being similar kind of return on capital. We've got a great early mover advantage. And as we roll our fleet out, that will be coming in. We're actually sub-renting some equipment today.
So as our new fleet goes in to replace stuff that we're paying third parties for, that begins to continue to impact our margins and drive that up. In terms of capital, the old business has been a cash flow business. So we've reinvested in the new equipment, the Top Fill system. That ended last year. We started buying back stock. That's going to continue to deliver us a large cash supply to help us grow the power business. Cover the details.
Yeah. So just to back up one slide, so this is just the capital outlay. So we've made this big couple of big orders with solar. They've got a series of installment payments along the way before you actually receive the equipment. So it's a bit front and middle- ended weighted. And so Q4 of this year, this quarter is sort of the big quarter of capital. And it starts to phase down as we receive the orders into next year based on our current order book. And that translates obviously into growth going forward. So at the Q3 run rate, EBITDA on the business was roughly $125 million. And that represented roughly $50 million of contribution from the power business.
Going forward with the incremental capacity that we've ordered and contracted, we see roughly $170 million of EBITDA from that segment on top of the roughly $95 million out of the legacy core business, so we've got guidance of roughly $250 million as a run rate exit number going into the end of next year, and that's all based on firm orders with the OEM and firm contracts with customers. I think that's, yeah, one thing on the shareholder returns we didn't hit on, if you're new to the story. In the fourth quarter of 2018, we instituted a dividend. We have increased it three times since then. We've never missed a quarter. Bill alluded to a buyback. We have roughly $15 million left on our buyback authorization. We bought back a fair amount of stock in the first quarter at roughly $7 per share.
And we're trading somewhere at $19-$20 today a share. So we've been prudent allocators of capital both on shareholder returns as well as obviously today creating this new growth wedge for the business. We put in place a $325 million term loan to help finance both the acquisition as well as the new orders for capacity. As we look at exiting that full capacity, we're significantly below two times on a leverage basis. So we feel very good about where the balance sheet is. And then as Bill alluded to the pie chart below, another unique piece is management and legacy owners, founders of the business own roughly 50% of it. So significant alignment here going forward. Key takeaways.
Key takeaways. Obviously, we've hit most of these points. I mean, it's a critical piece of equipment. It's very tight supply. We've got an edge with our team's ability to run the equipment, manage equipment, and deliver power, not just a turbine, and a great relationship with Solar and deliveries committed through next year and hopefully beyond. We've been able to contract up the fleet. When we closed the business, we were not in a position to announce that. We've got those signed contracts now, which is great. Two-to-four-year contracts, we think those actually extend even longer over time as the grid connectivity continues to get delayed out and the size of these facilities dictates that you're going to need some of this power there anyway. We're all in . We started this business in my living room in 2013.
We know every piece of this, every team member. We've had a say in putting that together and a real focus on a culture of ensuring that we provide top-notch service to our customers and our shareholders. Our core business is one that's not sexy, but it's a great cash business. It continues to generate cash flow. We love it. We like to hold on to it. We've got a great customer base there. We've been able to cross-pollinate already some of our high-end field technicians into the turbine business. We're both running off Allen-Bradley control systems and state-of-the-art equipment. We've been able to now cross-pollinate. Four of those folks are now actually running a data center turbine business with some oversight yet. But we've got, they're very capable and we'll continue to do that.
And if we look at the valuation of the company today and the returns that we see on this capital, we think there's lots of continued value to be created here. And we're aligned to shareholders in the story. So we made it in that one-hour presentation and we got it done in 17 minutes. So we got 12 for questions.
Yes, sir.
You're buying the equipment from manufacturers. Is that correct?
Correct.
You don't manufacture.
We are not manufacturing turbines. No, sir.
Do you sell it to the data centers or do you lease it to them?
We're renting that to them along with the power supply piece of that so we're taking purchased turbines, matching it with our own power equipment, so transformers to get to the right power, switchgear, breakers, and cabling, and delivering that to them on a rental basis.
Dumb question. Why don't the manufacturers of the turbines cut you out of the middle and do this themselves?
Because they don't, I mean, Caterpillar could build boats and they make marine engines, but they'd rather just make the engine and sell it to somebody else that puts it to work. So I mean, the manufacturers know that they want to manufacture. I will tell you that Caterpillar dealers tried to do this a couple of years ago and started a company called DPS, and it was a failure. So a third of our fleet was purchased from them later when it didn't succeed. Probably had a discount.
Yeah, had a discount. So this is a service business. So having the mentality of providing that service isn't what you do as a manufacturer. So it's a balance of what the customers want in that service level. And we have a great relationship with Solar, and I think they recognize their strength is making really good equipment. They're building a new 35 MW unit with their Titan package that we're going to work with them on how do we roll that out in a way that makes some sense too. So it really is a who's good at what and how do you make that work?
See, it seems like the applications of this beyond the AI centers?
Everyone needs power, right, so this can go anywhere and our fleet is running in the oil field for oil guys running ESPs on it, running batteries on it. We're going into two gas processing plants that can't get grid power for multi-year contracts, so they're running the electricity. The Solar turbine was built as an industrial turbine, so one of our customers is Enterprise, which Jim Teague is on our board of directors. They have 220 Solar Turbines in their business today, all running compression, running horsepower. This is not a unique piece of equipment to the oil field in the hands, but it's an application that, of course, you can run this anywhere you need power. It could be backup after a hurricane. It could be anything you need it for.
Is there any movement within the chip industry out of computer guys to develop computers that use less power?
Of course there is. I mean, their job is to try to use less compound. Unfortunately, what's happening is every chip they're building has more activity level, more computing power, and that's an exothermic reaction, which just needs more power to it. And now what you're seeing is the more power they can deliver, all it says, if I'm the user of that chip, I'm just going to put more in the same area, so I need more power or the same amount of power.
Generate more heat. So it's a little circular, but everyone would like to do it more efficiently, and they're all trying to do it. But in fact, the reaction that these things generate is just going to create heat. And they're trying to put micro channels through the chips and things to make the cooling more efficient and all that kind of stuff. And it will, but the computing power is growing faster than the ability to cool it.
Where are you based?
We're in Houston.
You touched a little bit on the competitive landscape and what that looks like in securing equipment, long-term contracts, things like that. I assume it's similar in other industries where you're kind of the big player. You get the equipment versus chips on it. Can you talk to that?
We've been able to secure a large portion of Solar supply. I would say we're not a big player. And 500 MW of units is small in the kind of power demand. We're growing it, growing it rapidly, but we have gotten ahead of this from a demand perspective and have a very tight working relationship with Solar. One of our principals that we bought with MER was instrumental in helping Solar develop the mobile package for the frac industry. So the U.S. energy industry and the frac in particular has created a whole engineering push on efficiency. And how do you do stuff? How do you make it more mobile? How do you set up a power plant and turn it down in six hours? And all that drive to make things more efficient actually was created some of this equipment out of that.
There are a few others that generate power that will do this. I mean, there's been the VoltaG rid and the Evolution and the private guys that have started really to supplying the frac market. We had some equipment in the frac market sub-leased to a few of those. We've pulled all that back for the data center market, forcing them to backfill their own equipment and capital needs into the oil field as frac grows, and so I think the electric component of fracking continues to be the green shoot in that industry. That's going to keep growing. Even if frac fleets stay flat, there's a pretty rapid switch over from the old diesel fleets.
Do you have also equipment that takes because it's gas, right? A lot of times in the field, they generate gas and then basically take that gas to generate the electricity.
Yeah.
To do that. So I just came across something I would love to hear your thoughts on, that is, where we're deploying all these data centers and all that is to use those in the fields, the gas and the energy that you produce there.
We have one unit being rented to an operator today who's using his own gas that effectively is free, and they're sub-leasing some of that power, selling some of that power to a Bitcoin miner, so it's happening today in those kind of small scales. But yes, the need for power for anything does that. The gas itself, the turbines are flexible into what kind of gas they can use. The rental business really needs a specific heat quantity of gas, and so most of that business today, VoltaGrid especially, is reliant on CNG, so they're having to take a pipeline quality gas, compress it, truck it to the site and use that. The turbines could run on field gas if they wanted to.
Right now, in today's economics, you'd rather take some of that natural gasoline and propane and butane and sell that to what it's really worth and keep maybe you could keep the ethane in the gas and burn a little higher BTU, and some of our customers, the gas processing guys, actually view the turbines as a flexible option for them to reject their ethane and burn it themselves and not run into pipeline spec issues, so there's an arbitrage around the gas side of it as well.
Yeah. What would be the expected wait time for adding additional capacity beyond the 500 MW? How does that compare to maybe other people who are also waiting in the queue?
The order book, if you talk to GE, and I think they've publicly stated it's kind of 28 now, 29. We have some solar orders in discussion for late next year, early 2026. So it's still a year out at a minimum for most of this. And you got to pair it with the right electrical equipment. So it's not just the turbines. You got to have the right transformers. You got to have the right power equipment and all that stuff. And we have a supply chain of that in process as well.
I was going to ask about site location. Do you get involved in that decision process with your customers at all? Or they say, "I want to put a site here," and then you work with them and then all?
Most of this has been them telling us that most of it at this point, if you go back in history, they picked a site. The grid would tell them, "Yeah, I'll get you your 100 MW, 150 MW." They said, "We need you to be a bridge power because, oh no, we didn't think. We thought the grid was going to be here in May. And you know what? They're telling us it's next May. So we need a temporary solution." And that begins to roll into, "Uh-oh. We didn't know it was going to be three years for the power. And by the way, these chips are going to use three times the quantity. So they're never going to get me the 400 MW I want.
They only have 100 available in a year from now," and so it's evolving very rapidly as chip design and data design does it. We're not with the developers. There's a lot of developers running around. There's probably at this point 10 or 12 developers developing sites for every two data centers that get built.
The old not in my backyard issue is not your concern particularly at all.
No. That's their problem, and we just show up and try to hide it as much as possible, and we've got some right now. We have a 250 MW plant in a parking lot. I mean, this is a scale of how this stuff goes, but we are working through developing, adding. So most of these things designed for the oil field didn't have scrubbing. The solar units have a SoLoNOx technology, which gets them down to about 9 ppm of NOx and SOx. The GE units are about 25 ppm of that. We're putting catalytic reformers, electric catalytic reformers on the back end of these things. So basically a giant catalytic converter for your car. We're putting it on the back end of the turbines to be there longer.
So we will have to file air permits for the customer or the customer's filing air permits, which will fit underneath it with additional equipment, which we'll then re-rent to them. Good question.
How about your labor supply? Keep all this going.
Labor supply is about culture, and we have really good culture. We've got several tracks into the military. I have a West Point and an 82nd Airborne guys in our management group, and we recruit pretty heavily from there. We have been able to cross-pollinate now, and some of our traditional legacy business folks are now on the turbine business, and we're cross-pollinating. It's always about finding the right people to run this thing. We've got a really good operations management team. We've been scaling up inside of this. It's also a lot easier to put 300 MW i n one parking lot than it is to put 10 MW- 30 MW sites spread all over the Delaware Basin, so this does give us a lot of efficiencies in some of those customer bases for labor as well.
But we've scaled our business, core business up a couple of times since 2014 and think we have the ability to do that and continue to do that here and are ahead of it. Back.
Thinking through just some of the additional contracts you've added, data center customers, I would assume that they're waiting in line to be connected to the grid once that contract ends. That's safe to assume?
It was safe to assume that was their original assumption. It's safe to assume that was their original assumption. We had a six-month contract when we closed. It's now a two-year contract, and they're expanding to the point where the power demand at that site is more than the utility will get them in 10 years, so it all sort of started that way until they started using these chips, rolling out the actual load and the demand. And the demand is real and it's big, and most big, if you're near a municipal, these are hard chunks of power to add to a city to add a gigawatt of power in a couple of new buildings.
So in a way, it's just ensuring the ability of the U.S. grid to keep the proper demand for the next five to 10 years.
Yeah. And there'll always be you're going to build it in chunks around the area. And so there's always going to be a bridge need for one to three to two years to sometimes it's more reliable when you think about the factoring in of the grid, all the extra costs that have to be absorbed in that rate base, plus the ability to have to have backup on your site at the same time. You roll that in, you may be better off just building your own power plant attached to your facility behind the meter.
In terms of price competitiveness, in the last earnings call, I mentioned that they were potentially more price competitive than the grid. And so I was just wondering, is that foreseeable that that could continue going forward?
Yeah. To the extent the grid has continued to have to meet new standards, absorbing renewables. Look at the losses in the line depending on the location. You start factoring all that in. And then from a, if I'm the data center and I have to build a backup power plant also to make sure my cooling load doesn't go down, and I start weighing those two options against each other, it's very competitive.
So, you're just trying to. I'm pretty good at this.
Yeah.
Visualizing the parking lot scenario, you had to pipe natural gas into the turbines, or you call them turbines, in the parking lot, and there's enough natural gas capacity in that area to do it.
Yeah. We're tapped off of a large MLP's main line and run a 12-inch line to there, and we can supply the power we need. Yeah. Actually, two of them into there now.
How about in the Northeast where I understand they don't use natural gas?
Yeah. They're not going to run a data center on a wind farm. But so they're not going to be located there. I think that's the issue with the further you get up Northeast, the less likely you're going to see a bunch of data centers built.
Okay.
Yeah. Yeah.
On the buildouts themselves being 12 months in terms of say, put an order today, build out in 12 months, what did that kind of look like? Is that expanded from kind of six months to 12 months, or is it just probably 12 months here?
So we put the order in in June. The first delivery of that was December. So it's expanded. Yeah. We got three minutes and 45 seconds to go. More questions?
Do you have this online?
This presentation is all on our website. Yes, sir. Yeah. Yeah. Yvonne Fletcher, our investor relations, has a hard copy if you want that, but there's one online.
One online because I'll be carrying them all day.
Yeah.
You're the president?
I'm the Chairman of the board, CEO. He's the President, CFO.
What are the initial IRRs again on the what end here?
What was the question?
Initial IRR on a two- to four-year contract?
IRR.
We think about it as a three- to four-year.
Probably a three-year payback, so it's that 20-ish% IRR depending on how long.
The life of this asset once we build it?
The turbines will last 20- 30 years. Effectively, it's about 10% the replacement cost. Maintenance would be accruing about 10% a year. About every four years, 30,000 hours, it's about $4 million on a 16.5 MW unit, $4.5 million. So it's about 10%, it's about 25% every four years.
Do you happen to have a geographic map of your locations in your package?
There's not.
Is it temporary skewed to the south and the southwest?
For the power, yes. Yeah.
Power is focused in the south-central.
Southeast, if you will, south-central, and the Permian Basin, and then the legacy oil field business is across, I think, 13 different states across the U.S., anywhere where there's drilling in.
Yeah.
And everything has wheels today, so it kind of moves.
It all moves. Any questions?
You made inroads in California where they don't want to put more power lines up and all that?
They're going to eventually have to have these out there, but no, that's not been a focus of our effort. There's other easier places to do business.
I understand that, but they may have.
They're going to be in a world of hurt, and they're going to need these things. And we may get a call. We may get a call. And yeah. So summertime.
Do you anticipate these pocket nuclear plants, these little leaks, competing with you?
I think they will be competitors of this solution at the point that they're ready to do that in three to five, seven years. If there's not a not in my backyard issue, it's a lot harder to hide a little nuclear turbine than a nuclear generator than it is a turbine that looks like a Walmart truck.
And to your point, everything I read about it is sunshine and roses, but what about the people who are not in my backyard? I don't want to make a new one.
Yeah. BANANA, which is build almost nothing anywhere near anyone. And it's great technology. There's no nuclear power. It's fantastic technology.
But being a skeptic with my grandchildren and everything, do I really want what can be perceived as amateurs running nuclear plants all over the place?
Yeah. It'll be perceived as that. It probably isn't that. And they know what they're doing, and they're very safe. But it's not going to be easy. No questions.
Awesome.
Yeah. On that note, thank you all.