Vertiv Holdings Co (VRT) Earnings Call Transcript & Summary

Okay. Thank you, everybody. I'm Chris Snyder, U.S. multi-industry analyst. Very excited to have Vertiv up here. Obviously, a very topical stock. We have CEO, Gio Albertazzi; CFO, David Fallon. Thank you, guys, for joining us. Clearly...

Thank you for having us.

Absolutely.

Very much so.

Clearly, very good demand in the market. Maybe not a surprise. I think you -- 35% organic for you guys last quarter, 1.2 book-to-bill. Oracle with some eye-popping numbers last night. So I guess from I think everyone knows that demand is strong, but for the investors and kind of on the outside looking in, what's the best way to gauge if things are accelerating, decelerating, getting better? What should we look for?

Well, I think there is -- talking specifically Vertiv or in general?

Vertiv.

Well, I think it is a combination of all the things we share with investors. And certainly, what we say in terms of sales growth, book-to-bill, all elements that are adding a lot of information. But a very important parameter for us, and we've been vocal about that quite a few times, indeed, every time is a strength of our pipeline. And again, strength of the pipeline tells what is ahead for us. I always like to remind everyone or to reiterate what pipeline means for us. We're very rigorous in what we define as pipeline is only active commercial opportunities which there is a quote. So when we say pipeline is growing, that is a very, very good proxy for what is ahead. So that is a way in which we signal the direction which the business is going. And that's beyond the individual quarter order intake, et cetera, has been quite strong, as we know, on both the year-on-year or a trailing 12 perspective, but listen to what we're saying in terms of pipeline and pipeline direction. I think that's always a good indicator for us.

And I guess, Obviously, the company has a lot of backlog, and then you have a pipeline that's even -- I guess, in addition to the backlog, how much visibility does that give you? Like do you guys -- I feel like '26, you have good visibility, does that pipeline even extend into '27 at this point?

Well, yes. The answer is, yes, it extends into '26 and '27. Of course, further out in terms of active quotations that are in the hands of our customers. Of course, it diminishes, but it doesn't mean that we do not have visibility beyond that. We have a lot of visibility in the industry. We have all the right relationships to continue to lead in the industry. And our conversations with our customers are years out. So a lot of visibility for us. But typically, our pipeline covers 12, 24 months. And again, I'm talking about pipeline. I'm not talking about backlog. Pipeline is opportunities that are yet to turn into orders.

Appreciate that. Then maybe on liquid cooling, which is obviously a big change that's going on in the market, Vertiv clearly does very well there. I guess how should us on the outside looking in kind of track the progress of liquid cooling, think about all the gigawatts that are being added, how much of that could go to liquid cooling? Are there any parameters out there we should look at maybe Blackwell shipments?

Yes. I think clearly, we see that the industry is shifting to liquid cooling. And this trajectory is clear. When we think Blackwell and further future generations, they're all high-density liquid cooled pretty much. So that is a good proxy to find what part of the industry is needing, demanding liquid cooling. But I would like to remind everyone that even in a liquid cooled data center, you will have air cooling in parallel to that simply because there are loads, and there will be loads that will continue to be air cooled. And then there is, let's say, downstream if you think about the flow of heat. There is a big chunk of the data center cooling market TAM that is in heat rejection. So never think about the liquid cooling is the only element of a thermal chain in the data center. It is part of a bigger chain. So when we think long-term, let's say, steady state, we think about liquid cooling as representing about 1/3 of the total TAM of data center cooling. It doesn't mean that only 1/3 of the data centers will have service that are liquid cooled, that percent will be much higher, but there is much more to the TAM than just a very important liquid cooling part, a part that we are very happy about in terms of our growth and the part that we love in which we excel...

Another change that's going on in the data center market and something that's driving a lot of questions for us, is the shift in spend between gray space and white space or infrastructure to things like servers and chips. I guess, as a company that competes across and sells into the whole ecosystem, what's your view on that? And what does it mean for Vertiv?

Yes. I think this is an important question in terms of explaining what happens in the CapEx structure. Clearly, the total CapEx numbers that we all hear about, total CapEx is infrastructure -- physical infrastructure, power, cooling, the building, et cetera, et cetera, and the sheer IT. In this total CapEx, percent-wise, the IT is growing a little bit. But the net share growth of total CapEx makes the pure infrastructure, our TAM, grow very, very, very strongly. So yes, the mix is heavy and heavier in the IT part, but the net growth that we see on physical infrastructure is very, very encouraging. And we've been vocal about that in our Investor Day. And yes, we see things moving certainly in that direction. If anything, on the upper range of the growth rates that we indicated specifically for colo and hyperscale market. But one thing that we're doing -- I'm going to be a little bit long winded on this one. One thing that we are doing is really increase our presence in the white space. Go back 3, 4 years -- 4 years. The white space would have very little challenges and content in terms of power and cooling and infrastructure. No longer the case. As density increases, think about the extreme 1-megawatt rack instead of 10-kilowatt racks, so multiple orders of magnitude, it's mind boggling. And the complexity in terms of how distribute power, how distribute cooling inside the rack and then the row and then the whole white space poses challenges that we are very well equipped to address. So the white space is becoming a very fertile ground for us in terms of growth. You heard about our acquisition of Great Lakes exactly to further strengthen system level offering. So it will be -- it is liquid cooling, power distribution. It will be a high-voltage DC as that type of kind of, I would say, liquid cooling-like type of transition happens also with power. One thing that we are super excited about is prefabrication inside the white space. We were explicitly talking about our SmartRun prefabricated white space fit-out solution in July when we had our last earnings call. And that can cut the deployment actually, fit out, using the exact terminology, the fit-out time by order of magnitude. Think about, you have built this very expensive massive data center and then you spend months fitting out the white space. And that month can become just weeks. So you -- as an investor, your time to revenue is shortened dramatically. So all things that we like a lot, and in the space of 3, 4 years, the space -- the white space that was a little bit, yes, we were in [ also ] play. It becomes an absolutely playground for our strength in power and cooling and prefabrication. So very excited about that.

I appreciate all of that. When we think about the infrastructure side, are these -- is the new greenfield AI data center being future-proofed? And is it bringing the thermal and power requirements to support where chips are going 5 or 10 years from now? Or do you think that even within that physical infrastructure, there's an upgrade cycle because of where it's going?

We can certainly talk about the next 5 years. The next 10 years is a little bit [ early ]. We don't mind that. But certainly, we are a partner to all the important players from silicon side, the NVIDIAs of the world of course, but also hyperscalers and colo. We are a partner in defining what their future infrastructure should look like exactly to be future proof. The -- an example, we shared with all investors at our Investor Day, a product that is quite cool in that respect, that is CoolPhase Flex, a product that is fully fabricated in terms of thermal management infrastructure, and that can flex liquid and air depending on how the evolution of the loads inside a given data center will happen during the life cycle of the data center. So there are a number of ways in which this flexibility can be built in. Whether that will be sufficient 10 years out? Well, it remains to be seen. Of course, if that turns into a 10-year out replacement and upgrade opportunity, then we will like it a lot. It's early to say. What we can say though that if we instead of looking at the traditional gray space, the power and mechanical -- the electrical and mechanical infrastructure traditionally outside the white space, we go into the white space with that electrical and mechanical infrastructure, power and cooling, call it the way you want. And you're designing now a data center for 150 kilowatts per rack. And maybe you have stretched that 200, 300 because you're forward-thinking and forward-looking as a data center developer. But what happens 6 years from now? So the likelihood that the white space replacement cycle that did not exist before will have probably a 5-, 7-year rhythm. It's quite high. Time will tell, but it's very logical to think that, that's what will happen.

I appreciate that. You guys help customers with front-end design work, you're part of the NVIDIA preferred supplier network. Can you talk about your closeness with the key customers or key industry players? And how that is a moat versus new entrants that are trying to come into the market?

We believe a very important one, a very important competitive advantage in twofold. So the -- being involved in road mapping, the future of silicon is absolutely important. We have to provide the technology that will enable Rubin, Rubin Plus, et cetera. And we need to have those technologies in our road map today because our products and our systems need to be out probably at least 6 months or a year ahead, so that the infrastructure can be built so that those new -- that new silicon can be than deployed. So very important. And being there at the drawing table as an enabler as opposed to say, "Oh, okay, now we see where the CDU technology goes. Let's build that too." So being a me-too type of player in the liquid cooling, as an example, but the same is true with high-voltage direct current power distribution. All big advantages because those conversations, that knowledge then enables us to sit at the table with a large customer, be them hyperscalers or large colocators everywhere in the world and say, "Hey, this is what we see happening." Sometimes it's a three-way table with a silicon provider, with NVIDIA, the customer and ourselves, and say, "Hey, this is what is happening. This is how you should design your data center." It's never [ us telling. ] That would be, how can I say, not the right characterization of the thing, it's really engineers and engineers cooperating and saying, how do we figure out the future challenges. And our role is bringing our knowledge about the future technology of silicon, bringing our knowledge of what is the art of the possible because of our knowledge of our road map in terms of electrical, mechanical white space infrastructure, et cetera. But also at that table, we define and understand what their needs are in terms of scaling. So it's very mutually helpful. And our conversation, our customer knows that the conversation with us is the entire infrastructure, the entire data center. They don't have to sit with 10 different customers, one that knows UPS very well and other that knows CDU very well, et cetera. We know the whole infrastructure.

Maybe outside of competitors, how do you feel -- or what do you think about hyperscalers vertically integrating into the infrastructure? I know Amazon has been doing it for a while, but there is an announcement a couple of months ago that got a lot of headlines. I guess what are your thoughts on that? What does it mean for Vertiv? Is anything changing there?

I would say not really. The -- no two hyperscalers behave the same way and have the same philosophy when it comes to how they build and what level of involvement they have in the definition of the technology. They all have a very strong opinion, as they should, as to what the data center needs to look like, but how exactly a CDU or a power distribution unit looks like, some going very much into the details. When you say vertical integration, we do not see vertical integration even in the examples that you have given, if anything, it's ownership of IP. So the -- if we -- specific case, you have an hyperscaler who has traditionally wanted to own the IP. And that's true across everything. So when people say, "Hey, AWS, were you surprised about that?" No, we are neither surprised nor is it a novelty. And generally speaking, the hyperscalers, even those who have a strong grip on the deep IP of the technology that goes into their data centers then collaborate with the likes of us or us directly in developing that technology or certainly in deploying that technology as a supply chain player. So yes.

I guess last quarter, 35% organic growth, you build backlog on top of that. I guess what is the -- what's the constraint on growth here? Is it customers time lines and all the upstream power issues? Is it Vertiv's capacity? Is it the supply -- your suppliers' capacity, customer time lines?

Let's start with the industry. As what are the -- I don't like to talk about constraints, but I call them pacing factor or moderating factors. Clearly, there would be an almost infinite appetite for capacity increase in the data center world. Well, you need power as we know. Power is -- especially in some parts of the world, is a limiting factor. Not in the sense that there will be a cliff. I do personally don't believe there will be a cliff. But certainly, the speed at which power is being made available is slower than the theoretical speed at which the industry would like to travel. But what we see is an enormous amount of capital and ingenuity applied and entrepreneurship applied to the power side of the equation with a more and more battery energy storage systems being deployed, more and more, let's say, off-grid data center being -- or partially off-grid data centers being designed. So things will happen. That will continue to be a moderating factor, but I don't see a cliff. The other fact is that the data center industry is a construction industry. So there is a limit to how fast you can go. I was talking about the -- I was mentioning our SmartRun technology, but prefabrication is a way to liberate speed if you will, in the construction business. When it comes to us, the growth that we have demonstrated in the first part of the year and anyway last year is certainly a testament to the fact that we can accelerate. Now we are very rigorous in the way we look at forecast and capacitize for that forecast. Forecasts are never an exact science as we know. So that's why we make sure we keep wiggle room in our manufacturing capacity. But again, it's not just manufacturing capacity, and we are adding -- we are increasing our CapEx. We are adding square feet of manufacturing space. So we are certainly ahead of the growth curve. But a very important element that is overlooked is you have to have services capacity. And the services -- the speed at which you're growing your services needs to be at least equal to the speed at which you're growing your manufacturing capacity, that is more obviously what people think of when -- and we've been doing that. And I think that continues to expand that moat that we're talking about earlier.

Yes. I mean maybe just kind of a follow-up on Vertiv's capacity. Like I guess, is it reasonable to think that your capacity is growing kind of in line or quite similar to the revenue we're seeing? And then a secondary question on capacity is that the view in the market was that as companies add capacity, orders would go down and backlog would go down because there's less concern around lead times. And we saw kind of the opposite. Do you feel like this capacity is actually maybe a driver of share gain in that you can now more so accommodate customers?

So let's start with the first part, is your capacity proportional to revenue? Yes, almost by definition. So clearly, as I said, there is a wiggle room where you can accelerate, use more of an asset or less, but you always have to have in the long run, in the medium run, absolutely, it is -- there is a proportionality between capacity and -- capacity and output. But again, capacity doesn't necessarily mean a square -- additional square feet. There is certainly an efficiency and productivity improvement that you drive in your infrastructure. That needs always to be the case. Otherwise, you forgo an important element of, let's say, the operational leverage that we promised all of you guys. But -- what was the second part of your question?

Driver of share gain.

Yes. And the -- how that correlates to orders and why the industry was expecting orders. I think that the -- as we've been explicit about the fact that we have not seen our requested lead time. The customers need a certain product, system at a certain date. And the time between when they place an order and when they need the stuff, it's typically driven by their construction plan. So of course, we are shrinking our lead times, but it doesn't mean that all the time we get an order, the customer is using the shorter lead time. The shorter lead time is to gain market share when the customer finds themselves, the supplier is creating problem, Vertiv can you come in? Or we have an opportunity to accelerate the project, can you serve us at a higher speed? So we continue to think that, that 12 to 18 months lead time -- average lead time that we see in large projects will stay. And that doesn't have kind of a consequences in terms of order intake reduction. Now we are very rigorous in terms of definition of backlog. We only define backlog, what is a legally binding PO. So that is not going to change dramatically, maybe 2 months more, 2 months less. If we were, but we're not, if we were to say, hey -- every time we have kind of a long-term commitment and people say, "Hey, do you have capacity for the 3, 4 years out?" Well, maybe the answer could be slightly different. But those 3, 4 years out type of conversation continue and are very active because people are aware of the importance of securing capacity in the data center industry that continues to accelerate. But yes, I don't see a kind of a dramatic shift from an order intake standpoint.

Appreciate that. Maybe moving over to margins. Margins down in the quarter. We saw gross margin pressure despite 35% top line growth. Can you just talk about the headwinds incurred in the second quarter? And kind of what's the pathway for recovery there?

Yes, sure. So I would pocket the headwinds into two different categories. One, of course, are tariffs. And it took us 25 minutes to get to a tariff question, that's pretty good. It's tariffs. And everyone has been impacted by tariffs and we're no exception. And the tariffs certainly hit us in Q2 without the opportunity to put in effect countermeasures. So both pricing and then supply chain. In addition, as it relates to some of the supply chain actions that we have put in place, those have created some operational challenges. And to the extent we changed the footprint, we probably underestimated the impact and the disruption that has on the actual operations. So that's what you saw in Q2, but we also laid out a path as we go through the year, addressing these headwinds, both the -- related to tariffs and also the operational execution. So if you look at our Q2 quarter, operating margin around 18.5%. We project that going to 20% in Q3 on lower volumes because of timing. So that's absolutely reflective of some of the work that we're doing related to the tariff offset and also addressing some of the operational challenges. And then when you get to Q4, we're guiding to an operating margin in excess of 23%, and that would translate into an incremental margin between 30% and 35%, which is absolutely in line with what our long-term ambition is for incremental margins. So I don't want to say we were ever off the path to hit our long-term margin goals, but certainly looking at Q4, it provides pretty good confidence heading into 2026 that we're back in line with what we would expect in the long term for margins.

Yes. I mean you guys obviously have a lot of backlog. Is -- does that make it harder? Something that we've heard in the channel is that it could be hard to reprice the backlog against data center customers. But I think maybe the more important question is, do you feel confident in the ability to get price and margin on the next order, on new orders that are coming through?

Yes. I would say incrementally, yes, it's probably a little bit more of a challenge on backlog than going forward. But I would say every single case is different, every single customer, everything single order has a nuance that makes it a little bit different from case to case. But as I mentioned, one of the responses to tariffs has been pricing. We're absolutely doing everything we can first from a supply chain perspective to mitigate the impact of the tariff. And we're very transparent with our customers as far as what we're doing. But we've been pleased with the results we've gotten with our pricing discussions. And I would say it's not unique to tariff so -- or it's to tariffs. So as we laid out in November '24 at the Investor Day, we see commercial execution and in particular, price/cost as a lever for margin expansion. So we continue to be price/cost positive in '25 with and without tariffs. And we would expect to continue to be price/cost positive going forward.

No, I appreciate that. And you kind of talked about Q4, that normal 30%, 35% incremental. I think you guys talked about still some headwinds in Q4 and maybe like exiting the year kind of more at a neutral. So -- and then we, I guess, have easier margin comps into the first half. Like is it reasonable to assume that next year could have above normalized incrementals?

Probably a bit premature to talk about 2026. But I would say there is nothing that we imply with the margin that we expect in 2025 that would suggest 2026 would be any different than what we would expect any year heading out to 2029. So in general, we expect that 30% to 35% incremental. And there may be a little bit of a tailwind as we get kind of the year-over-year overlap with some of these tariff actions. But probably a little premature for '26, but I don't see '26 as being anything different from an expectation perspective versus '27, '28, '29.

I appreciate that. And then kind of staying on the tariff, is the policy having or bringing any sort of competitive advantages to the company? There's European competitors, there's Asia competitors, Americas, where we're seeing most of the activity?

I can take this. I would not comment for our competitors. What I would comment is we are pretty happy about our footprint. Well before tariffs, we were pursuing a region-for-region type of supply chain architecture within reasons, of course. And so what we have done as tariffs hit us is to accelerate this. So the resiliency of our supply chain is still -- is there and it's even stronger. So we think that we are very well positioned. And I go back to the question we're saying about some of the challenges that we have experienced in Q2 and everything else that David explained. In many respect, we're an industrial company, you call us tech industrial, industrial anyway. So you see a quarter that is growing 40%. For example, in North America, 40%, we had physical stuff being shipped out of the door. And oh, by the way, while we're doing that, we were just reconfiguring our supply chains to make it happen. That means a lot is going on. So you in operations and supply chain have your hands full. So I'm pretty proud of what we have done. I mean we know that we -- as always, we can do much better, and that's our commitment to all our investors. But well, there was a lot going on that...

40% sounds like more of a decade's worth of growth than a year. I think the strength in Americas is -- I think it's pretty well appreciated and known by the audience here. Can you talk a little bit about what you're seeing in the international markets? Europe seems like it's disappointed from a growth perspective in the industry. APAC seems stronger. I guess, what are you seeing there?

Absolutely. Just like you said, Europe and -- Europe is slow and it's something that everyone in the industry is saying. There are multiple reasons for that. Certainly, a stringent regulatory environment, a little bit of a lag that has always existed that relative to what happens in North America and the U.S. That lag is a little bit exaggerated relative to what we have seen historically. But one thing that we know is pipelines are very strong. Pipelines are strong. The second thing that is encouraging is that if you look at the vacancy rates in existing data centers are really low. The other aspect that I always like to think about is, whereas from an AI standpoint, you can train everywhere in the world, when you go to inference, you have to influence within the, let's say, sovereignty jurisdiction in which you want to operate. This is particularly true for the EU. So we like to think of Europe as a coil spring, not sure when that will start to uncoil, but there are very, very compelling reasons to believe that, that will not be too far out. Asia is certainly a good story. And we should talk India and China and the rest of Asia or Australia and New Zealand individually, but generally speaking, it's quite encouraging [ market. ]

Well, we're up on time. I can ask questions all day, but I appreciate you guys coming up here with me.

Well, thank you very much.

Thank you. Thank you.

Thank you. That was great.