Vertiv Holdings Co ($VRT)

All right. We'll kick off day 3 here at the JPMorgan Industrials Conference. We're very happy to have Craig Chamberlin as well as Gio Albertazzi from Vertiv. I think Gio is fresh off a flight back from GTC.

So before getting into that, maybe just starting every fireside this conference with a bit of an overview on just maybe your -- the Middle Eastern situation and how we should look at that in the context of Vertiv.

Well, of course, the situation will unfold in the next few days or weeks for us. Certainly, an area in which we operate in terms of markets we serve with some manufacturing capacity there. We feel pretty good in terms of the supply chain resilience that we have built over the years. In terms of the market is a subpart of our EMEA market. So it would be probably premature to think about what the actual impact is, but we see good resilience across the spectrum. But time will tell. It's a little bit premature still.

Yes. But you haven't heard any of any of the -- I know there was some activity that was ramping there that's probably a bit deferred, but you haven't really heard any news about any of those projects that are not really moving through.

No news, no news at this stage.

And GTC, I mean, I've been at this conference, didn't have a chance to listen into what the CEO of NVIDIA had to say in depth. But maybe what are your couple of takeaways from GTC and how that would -- maybe has evolved for Vertiv?

Well, certainly, in our confirmation of the huge amount of demand that NVIDIA see and that's the demand that, of course, we also see. So our view of the market -- the views of the market align quite a lot. Two other things that are important is the continuous evolution of the silicon, but not just the silicon as a system because I'd say NVIDIA is past the silicon in and of itself, of course, the silicon GPUs are central, but it's the entire IT system -- the entire, let's say, part that have been developed. So continuous evolution more and more productive silicon and systems, which means that you drive more tokens per megawatt, which makes the whole AI more -- every token more cost efficient. And with that cost efficiency, of course, comes even more adoption of AI. So that was encouraging. And we were part and we also announced our partnership being part of the DSX, Omniverse type of infrastructure, again, aimed at optimizing the design of data centers or dare I say, should I say, AI factories, but also enabling a fast deployment and an optimization during the life cycle of the infrastructure at the IT factory. So we are participating in that effort. We are an important player in that effort with products like our Vertiv OneCore. So everything going in the direction of an acceleration in the industry and acceleration towards faster deployment and better utilization.

And the more efficient these chips get, the more tokens, I guess, come out, the price comes down of those tokens, but the net -- it's still net very positive for your customer from an ROI.

Very much so. Very much so. Yes, very much so. And we like this evolution of the technology because that technology becomes more dense, more demanding from an infrastructure standpoint. So that's exactly what creates value for us in terms of the market that we serve, but also leverages our unique capabilities. So very encouraged also by the incredible buzz around everything, GTC, also Vertiv in that perspective, very excited.

Right. I think there was more buzz here, though, at the JPMorgan Industrials Conference. That was the second most important event of the week for sure. From a content per megawatt perspective, you guys kind of bumped it up to 3, 3.5 -- as we move into more prefab and complex systems, where do you see that going over time? What's the influence there?

Well, yes, I think we will have an opportunity when in a couple of months plus at our Investor Day to further elaborate or elaborate in more details on this aspect of our business. But in general, we see fabrication, and we like to talk about this integrated and converged infrastructure. It's not just the integration, but it's a system that is designed as a converged infrastructure. But we see that, that helps us add more value and capture more value. So it's something that drives towards an increase in megawatt -- sorry, dollars per megawatt.

And when you think about those kind of prefabs, a lot of that -- how much of that -- of those prefabs is your equipment versus kind of a third-party content?

It is a lot of Vertiv equipment. That's a lot of Vertiv equipment. If you think about our portfolio as stretching all the entire range of power and thermal. So the majority is our equipment.

Yes. Okay. Unlike some others who may be packaging a lot of stuff that don't have...

One thing is being an integrator -- and there is value in integration. But another thing is to deliver a system that is thought of as one system that is tailored example on, say, Vera Rubin. And it's kind of -- how can I say, crazily optimized around that application. And that has a lot of value for our partners and for our customers.

Okay. Before getting into the technology, maybe just on the overall cycle, what inning do you think we're in here? It's always hard to tell. We're definitely out of pre-game warmups, but my view is we're still not quite middle innings. I'm not putting words in your mouth, but what do you think inning we're in, in this whole cycle?

Look, we are still at the beginning. We're really at the beginning. Now we see tremendous traction in terms of adoption of AI, but it's so the beginning. So we see a long, long trajectory. So I agree 100% with what you saying.

And when you think about any risks, what are you hearing from your customers as far as bottlenecks? We had a panel on Monday, and they basically talked about some of the power bottlenecks being relieved, but then something else pops up. But they seem much more, I guess, constructive on the ability to fulfill orders. So what do you think maybe bottlenecks and then risks that you watch out?

Yes, let's start with bottlenecks. If we go back a couple of years, maybe 1.5 years when everyone say, hey, the world will stop as we run out of power and say, no, power is a pacing item. Power, of course, with an industry that grows at this speed, there's always something that you have to work on and to remove as a bottleneck. That doesn't mean that power is not something that continues to be worked on and optimized. But I say, hey, the demand is there. Demand is real. Capital ingenuity are being thrown at the power problem. The power problem will be, I wouldn't say eliminated, but managed for the growth that the industry needs. The same is true for other possible challenges, permitting. But of course, it depends on the country in which you operate. So I'd say, yes, this is an industry that is growing very fast. There will always be something, but there will always be an answer to the challenge at hand.

So when you look at it today and you kind of rank those, what are the top 2 bottlenecks you think the industry is facing?

I would say continue to be power as a pacing item. The other is I go back to what I always say, this is a construction industry in many respects. So construction ability -- sorry, labor and complexity on site is something that is being addressed. And we were talking about prefabrication. It's exactly the way in which this has been alleviated. So there will be other things in the future. There will be other solutions in the future.

And any risks you see out there today that make you nervous that we could have an air pocket or hiccup or anything like that?

We don't see that. We don't see that. We see strong demand continuing. And yes, no air pockets.

Yes. I'm not going to ask about orders, but you said you would talk about pipeline. Pipeline is still doing pretty well.

Well, we go back to what was it mid of February -- so I'll go back to what we're saying there.

Okay. And then EMEA, it seems like there's a bit more optimism from your tone on EMEA. What are you -- what's going on there? And I know it was -- it's going to be down here in the near term, but how do we see the trajectory in EMEA?

Sorry, we go back to the...

Putting Middle East aside, of course.

Yes, of course. I go back to our earnings call in February. We see the spring uncoiling as we coined, I think you coined that way of putting it 6 months ago, so when we were to get. That is definitely happening and pretty optimistic about that.

From a customer perspective, you -- there's a lot of different players now in the market. How do you guys make sure that you're -- you can't -- I mean, you do serve everyone, but how do you make sure you kind of choose the right projects, the guys that are going to be there 5, 10 years from now? Or is that not really a concern right now just because of all the funding and the capital that's coming in? So go ahead.

No, I think -- one, we are selective in certain regions and understanding what we have to do from a long-term perspective of winning. But for us, it is a quick cycle to cash. So we do want to deploy and get our installed base out there. I think where you're going, Steve, what I think about a lot is who's going to be operating that data center 5, 10 years from now so that I can get intimacy with that customer and drive what I call the services model on the back end. So I want to ensure they have stability there so that I can ensure that they're continuing to go and I can wrap my arms around it. I think on the OE side, the quick to cash helps us out a lot because we end up getting our -- I'd say, our items deployed, our equipment deployed and we get it out there and we get our cash in the door. The back end of the cycle is what I always want to make sure that I understand what the customer is going to be and how long they're going to be there so I can continue to drive revenue on that piece of product that I put out in the market.

And then when it comes to the different types of technologies, whether it's ASICs or GPUs, do you -- I mean, is one more content-rich than the other? How do you think about that split?

Well, in general, when we with that part of the market, the ASICs, anyway, advanced silicon, pretty logic kind of all live in the right side of the spectrum that we gave you and that we will further elaborate on in our Investor Day. But clearly, if you take the Vera Rubin or even future, then it's the kind of right most end of that spectrum. As we said a few times already, NVIDIA defines the envelope, and the rest falls inside the envelope. But again, when we talk about AI silicon, it all sits on the right side of the dollars per megawatt spectrum.

Got it. Got it. So it's not -- obviously, NVIDIA is over here, you're doing their kind of cutting edge, but the ASIC is still right within that.

Absolutely. Absolutely. Take TPUs and you take an example of very, very high densities -- very high densities. That are good for us.

From a -- just getting into the technology a little bit here, the 800-volt architectures that are coming out, where are we in this journey? And how do you see it playing out from an architecture and then also a penetration perspective as we move into the next few years?

Well, we are progressing well with the plans that we shared with you with investors being kind of second half of this week -- sorry, this week. Second half of this year. Still jet lag -- a little bit jet lag. It was a short kind of red-eye.

Weeks feel like years...

Exactly. So that plan continues to be executed upon, and we are happy about the pace. When it comes to the architectures, we shouldn't think about it as a black and white. One day, someone will flip a switch and everything will become an 800-volt ECO plus/minus 400 volt depending on the type of silicon and let's say, application. But we will see multiple powertrain configurations from, let's say, full medium voltage to more traditional -- from more traditional type of powertrains. And when there is an 800-volt DC silicon at the end of it, you will have conversion, an 800-volt DC done at row level to fully, let's say, medium-volt DC native powertrains. It's a big spectrum. And I -- we do not see a day when everything will flip. We see a coexistence at least for the foreseeable future of multiple architectures, which is something that we like. We're very confident on our medium voltage or 800-voltage DC technology, and we see the fact that all the technologies will coexist in the market.

What are some of the moving parts within that? Is there anything that's pretty clear right now that will happen within the architecture that will change?

You will see -- if I understand the question, you will see...

Just thinking about the traditional architecture where it's transformer to switch gear to UPS to what a power shelf like just thinking about that train. I've not seen the NVIDIA white paper. Maybe that's it. I don't know, but I'm sure there's like 1,000 other variations.

There are a lot of variations...

And I'm curious where we are today, like what's the most -- what's seemingly kind of the most definitive thing you can say?

Look, today, we still see the majority in the traditional powertrain designs. We are working a lot on everything, battery energy storage system plus medium voltage UPS as an area that is very, very important for the future.

And as far as the content there, I know -- I think you have a partnership with LS Electric, but I'm not -- is that related to this 800 volt? Or do you have what you need in-house to manufacture...

It's pretty much Vertiv thing. Of course we have a good relationship with LS. But this 800-volt and anyway, DC portfolio expansion is very much a Vertiv in-house thing.

And solid-state transformer, is that something that you guys will be -- or is that too far outside the Vertiv envelope?

Well, I think it would be a little bit premature to say. But as usual, when technologies become important for the industry, we are very early in that space.

Okay. And just lastly, this adoption trend, you're going to start to sprinkle it in second half '26. How does that ramp look? Like what percentage of installs could it be in '28 or '29?

No, I think it's still premature. I think it's still premature also for the market in general and just for Vertiv. And you'll see different type of operators, different type of data centers adopting different types of technologies. So I think talking about mix and volume speed of acceleration is a little bit premature. But we have a good understanding of, let's say, of range of outcomes. And of course, preparing not just from a technology, but also from a capacity standpoint.

Right. And there are still, I would assume, some hurdles from a safety perspective, and this is a pretty robust change?

This is an important change. It's an important change that goes in the direction of making our services and our services capabilities also in the medium to high voltage more important from a customer relations standpoint and ability to serve a complicated infrastructure.

Shifting to the cooling infrastructure. How do you see that evolving? Obviously, liquid cooling did go from 0 to 80% of capacity or whatever now. Anything on that front? And how do you see the -- I guess, when I also ask kind of the role of the chiller going forward because you guys are a systems provider, so you're may be in a good spot to kind of understand.

No, absolutely. So first of all, of course, liquid cooling is ubiquitous. So -- and will continue to be. There is no turning back. There are some customer players that continue to be focused on air, but the majority. The majority is going to liquid with, of course, a portion of air as well because they are in the data center loads that continue to be air cooled. But let's go to the system. When we think about the thermal chain is really as a system, the heat rejection, so where a chiller sits is becoming a more and more important element of the system. As liquid temperature are increasing or can increase with some of the newer chips. And going back to what Jensen Huang mentioned, what was it, a month -- a couple of months back. Then you see that the ability to use a smaller amount of mechanical cooling, that becomes a real opportunity. Is an opportunity, of course, it depends on the geography. It depends on the climate in which you operate. But it means that your heat rejection system or system of heat rejection equipment becomes from one size fits all to you can modulate chillers, dry coolers, trim coolers in our case, our portfolio. So a product that is in between that can depending on the condition of the locality, depending on the type of climate, depending on where -- when you are in the year, you can modulate the amount of mechanical versus free cooling, generating a lot of efficiency, but also needing much less electricity and power that can be converted to generating tokens. So the whole system becomes more efficient, effective, but also more complex to deliver and to orchestrate. That's something I would like.

So if we look at the kind of growth rates of liquid versus air, I think the liquid cooling TAM has to be significantly bigger than anybody would have thought 1.5 years, 2 years ago, obviously, just because of the megawattage, nobody predicted that. How do we -- is the air cooling market still like a really solid double-digit market? Or is there -- is liquid just taking over that TAM?

Yes, there is not a taking over, if you will. Clearly, the speed of growth of air cooling is much smaller than liquid. I would say that probably in -- I don't know, but soon, we will see those 2 TAM growth started to align because the mix shift will probably be completed in a year or 2.

Right, right, right, because you're currently penetrating dramatically and...

We're still in a phase of a partial substitution and that will stabilize with both, of course, but at the -- let's say, steady-state mix between air and water.

And from a growth rate perspective and just looking at your current long-term growth rate. The cooling market and the power market are those pretty -- is one growing faster than the other in that the 12% to 14% that I think you still have out there as your...

I would say that we should look at the market as synchronous -- pretty much synchronous in terms of growth rates.

Okay. Got it. How are you just pivoting back to the customers here. How are you seeing the funding environment evolve there? There was a super interesting deal you guys did with Hut 8, I believe. That seemed like it was a pretty -- not creative, but like a pretty innovative deal. Are you seeing more of those out there from a financing perspective and funding perspective for the customers?

I think we go back to the comments where they say, what are the pacing items for the industry? Well, when you have a pacing item and the demand is very, very strong, than the market becomes a very good at finding solutions to problems. So that's what we see, very, very strong solutions and very convincing solutions, and that's what we're seeing.

From an end market perspective, obviously, the colo is growing really fast. What are you seeing in the enterprise?

Still behind, of course, everything colo, hyperscale in self-build or new cloud. But we start to see an increasing interest in enterprise for, let's say, enterprise sovereign or private data and private AI. So we're still early stages. We're seeing, again, going back to GTC, NVIDIA starting to have systems that are silicon and systems that are designed specifically for Edge enterprise applications. And the market is there. And the market is there. And I think it is early stage still, but encouraged by the dynamics and by the direction of travel.

Is it growing? Is enterprise growing today?

It is growing. It is growing. I believe there will be an acceleration. And I think we will have an opportunity in 1.5 months when we have -- 2 months when we have our Investor Day.

And then telecom, any signs of light there?

Well, that sector is still growing relatively slow. I think we go back to what we're saying low mid-single digits.

Okay. And then just lastly on the business and the top line, talk about services and where you are today on that initiative. Is -- I would assume with all these liquid cooling installations and the risk-averse customer around that new architecture that there is a real opportunity for you guys here?

There is, and we are capturing that opportunity very, very, very well. And it's not just a risk aversion and certainly, the risk aversion is a fundamentally important survival trait in our customer base, but it's really the sheer complexity of the infrastructure of the building. When you have a liquid cool infrastructure, I go back to talking about thermal, the same is true for power with more complicated heat rejection systems, you have to have a very, very skilled, very, very reliable and very, very knowledgeable partners with you such as we are. But again, this increased level of complexity creates new opportunities. If you go back to our PurgeRite acquisition, 2 years ago, nobody would have thought that the fluid management space would be so important as it is today. So the complexity of the system, the sensitivity to perfect calibration at commissioning and during life cycle of the primary and the secondary fluid networks are creating beautiful opportunities for services.

I know it's a relatively smaller part of your business. You guys include installation and commissioning and things like that in the service number. But are you getting any visibility on maybe the multiple or the content per megawatt installed that you could get over the lifetime of the product. I mean these things haven't -- these haven't been installed for more than like a year. Any visibility on that kind of business model, how we should think about that?

Yes. I would say that we would not be specific here. Of course, we have our models and our -- I say good understanding, of course, of the trajectory. When we talk about complexity being favorable in terms of wear in the 3 to 3.5 megawatt -- million dollars per megawatt, well, then service is clearly also pushing that to the right. And the services on complex system is certainly very favorable in terms of content.

How fast can that kind of core service business, so not the install-related stuff?

Yeah, yeah, what we call life cycle.

How fast can that grow? Is that -- if your revenue grows 20% this year, whatever, high 20s, how fast is that business growing? Or should we think about that?

We're -- it's growing right now in double digits. You just don't see it because it's not growing as fast as the OE. And I think that's the -- what ends up happening is a lot of that OE equipment stays on warranty for a couple of years. And then as it comes off, we attach and we try to attach for the, let's say, the extended services and extended warranty period. So in my mind, it's more of a steady growth, but I like it to be a steady growth as -- and what we've talked about, Steve, is at some point, the OE party is going to slow. And as it slows, I want to keep that backbone growing at the low double-digit kind of growth because it's going to continue to give me the -- I'd say, the mix that I want and the outcome that I want and the revenue that I want. So Gio and I are pretty maniacal about ensuring that we wrap our arms around every piece of equipment we have out there and ensuring that we have the installed base service because as you said, it's early innings. We don't even know what that revenue model looks like yet because some of the stuff has just been deployed for year.

Another comment here. When you think about, if you will, almost a chasm between commissioning and life cycle services kicking in contract. Well, that is closing very quickly in, for example, liquid cooling or systems that are complex because it's not about being covered against things that you are cover anyway during warranty. It's about you need to balance the system. You need to have reliability built in and very often a crew of Vertiv service people on site day 1. It's not about, oh, yes, I'll see you in 2 years. It's day 1. It's critical day 1. So maybe the content is a little bit smaller because there will be part of the activity that is covered through warranty, but it's really kind of the monitoring, the present, the service level guarantee on site that play a very big role. And we like a lot the direction of travel.

And that's what Next Predict is really all about, right? Next Predict is getting in there and ensuring the stickiness. Like -- again, Steve, you've been around industrial is kind of where I've been around. The ultimate end game at any industrial build-out is to be there when the services start paying off. I think we're lucky enough to be part of the build-out and you can grab that OE, but really the end market and the future staying is going to be how much you can grab of that services market and staying in it for the long term because that's the real 10-, 20-year kind of revenue stream that you want.

Right. And that incremental cost, obviously, to protect what the $1 trillion of CapEx you just spent is relatively small in the event that something goes wrong. Yes, yes. Just on the margin front, the margins last year impacted by tariffs a little bit this year. We had always thought about you guys as a 35% incremental margin company. You're a little bit below that this year, below gross margin. So maybe just talk about how you're kind of metering your productivity efforts against demanding customers and the investments you have to make. What is the -- what's the biggest walk down from your gross margin to kind of the low 30s incremental? And is low 30s incremental the right run rate to think about longer term?

And I think longer term, we'll come back at Investor Day and kind of give the same framework. I mean we've always talked about 30% to 35% is kind of the space we want to be in, and we want to grow towards that. So I think we'll get more of a framework around it as we talk in Investor Day. But to walk down to this year, it's one, the bounce back, right? We're bouncing back from the tariffs. So we're ensuring that we get that back. But then you -- we did have a heavy investment. As you know, we're going from 2% to 3% to 3% to 4%. So bringing on that additional capacity, you do have a little bit of a headwind from that, that we'll work our way through here in the first half and into the second half. And I think we'll exit the year. We'll begin the year like probably high 20s, 28%, 29%, and we'll work our way up into the 30s and then we'll exit at that rate. So I think that's kind of what we're expecting, and we'll put a more bow around it as we get into the Investor Day, but that's really the big walk down for us is bringing that capacity on effectively, not rushing through it.

Any movement in recent movement in raw materials that we have to keep an eye on? Or are you guys pretty -- managing it pretty well with price.

I think we keep an eye on it regularly. We've increased our pace in the terms of how we look at it and how we go back to the market in terms of pricing, we regularly understand what that input cost looks like and try to ensure that we're pricing for it and pricing for it -- and pricing for end margin. So for us, it's a pretty regular drumbeat. We look at it almost on a daily basis and understand what it looks like, what does our pricing look like, ensure our quotes are updated for it and update not just for the one that's going to be delivered, but look at when it's going to be delivered and how we expect that input cost to change.

And do you target margin -- just remind us, do you target margin neutral, margin accretive, dollar neutral? How do you think about?

We look at it at a minimum.

Price cost -- on a price cost basis.

At a minimum margin neutral.

Got it. And then just lastly, on the margins, restructuring. Are we at the tail end of any restructuring? Or is there always stuff to do? And how do we think about the targeted efforts there?

I mean I think we look at it by region, by what we see from in terms of what the market looks like and these markets change dynamically. So you'll know that we've done some in the past, and I think it's always based on what we see the end market looking like. We don't want to sit and wait if there is an action that needs to take. We want to be out in front of it and understand what that means. So I don't say it's completely gone. We're still working through a couple of them. If we see the market change on this, of course, we're going to go out and take the cost out because we don't want to be sitting with unused cost. So that's kind of the way we think about it.

Exactly. I know it's a state of mind of permanent optimization -- permanent optimization of cost. It's not that something that happens every now and again. You constantly think about you optimize the company.

Yes. Any color on the segment margins and how we should think about which ones have the most opportunity over the intermediate term?

I mean I think when you're looking at segment margins, again, we feel like Americas always has the tailwind of volume behind it right now. EMEA is bouncing back. And I would say it's bouncing back because you felt this first half of the year, it was the dip on sales. So they get a little bit of headwind from that, but they'll bounce back in the second half. And I think we see a steady trajectory in APAC of continuing to grow and optimize there. So that's kind of the way I look at it is ride the volume in Americas, bounce back in EMEA and the steady growth in the APAC region.

And then just one more for you. What the heck are you guys going to do with all this cash? I mean it's pretty abundant, and you guys are at, I don't know, 0.5x...

I mean, we like our balance sheet. So I'm not going to apologize for having a strong balance sheet. It's what got us investment grade. So I'll stand behind it 100% of the time. But that doesn't mean we're not going to look at things that will strategically fit in the portfolio. For us, it don't spend money to spend money, spend money smartly and spend money where we can actually get a large return on that money. So it's -- we've done it in the past in terms of the add-ons and the bolt-ons that we've gotten. We always look at them at what is the growth potential behind them and what's it going to return to us. And for us, it's not a, hey, we got to go play in the inorganic market. We really like the portfolio. Now if something is missing or we think there's a spot where we can really jump in and grow it, we'll go do that. We're a little lighter when it comes to stock buyback or dividend. And we like to keep the powder dry in case there is something that pops up and we want to go. But the best place for us to spend our money is R&D and CapEx because that's an immediate return. I know what it is. It's tangible. We spend the money that's going to be spent there, and then we start looking at how do we deploy. I don't know if you have anything else.

No, I think it's absolutely as it is.

And I know there have been a couple of deals, but you're still -- are you still cold plate agnostic, if you will? Or is that an evolving area enough so that, that may be something you guys would be interested in either developing or.

In our industry, you have to be server agnostic. You do not know what server or what cold plate your thermal infrastructure will serve. And indeed, the 2 things are disconnected. So you know what type of loads, but you do not know what the cold plate that you will indeed serve is. So clearly, agnosticism versus relative to the server is a very important part of our strategy.

Okay. Sorry, one more point. I just wanted to clarify this, which I should have before -- because I'm getting this question all the time. On 800 volt, as you see it today, that is within the -- your content window that you've talked about, whatever the architecture that you've seen, that is within the content window.

Yes. Not only that, but we believe, again, it's on the right side...

On the right side...

On the right side of the window.

Well, thanks a lot. We had Dave Cote here on Monday, and he was singing your praises as he reflected on his career. So congrats on a great run and -- keep it up. And congrats.

Thanks.