Welcome to day two of Bank of America's Industrials and Transportation event. As I said, this is year two in the larger format. It's great to see you. I think today is the day of the greatest hits. So we're going to kick it off with the Vertiv CEO, Gio Albertazzi. And we also have Lynne Maxeiner, VP of Global Treasury and Investor Relations in the audience. Thank you for being here.
Thanks for having me.
We were joking. You can always be busy covering Vertiv, which is a blessing for us, keeps us in business. We really, really value the relationship with the company. Thank you for coming to our event. We are going to jump into Q&A right now. Thanks so much, Gio.
Perfect.
I think the idea is just maybe we're going to jump into liquid cooling right away. There have been a few entrants in liquid cooling, like CoolIT, Anand, Boyd, Nidec, STULZ, Delta, Lydon. It's a very, very long list, C hilled Iron, CoolCentric. And we didn't even mention Trane or Schneider, which both Motivair. So the question for you, how can Vertiv, which has 20%, 25% market share in air cooling, keep a similar share in liquid cooling?
First of all, we're not, of course, surprised with the fact that a relatively new space like liquid cooling and let's remember that the big acceleration in liquid cooling started to happen last year is attracting a lot of new entrants. The situation from a competitive standpoint is probably less different, if that makes sense, than it appears relative to the rest of the data center space. You mentioned the acquisition that Schneider Electric announced six months ago or so. In the list of the companies that you mentioned, some are very small, some are very new, some do not have the scale, some have the scale, some will be acquired, and there will be consolidation. In the end, some of the players in that list were a player and are a player in the rest of the thermal management for data centers.
What we see is, in the medium and long term, a consolidation and a competitive landscape that is not dramatically different from what we have experienced historically in the data center space. Another aspect is that we win on our competitive advantages. I want to mention innovation and technology being ahead of many in terms of knowing and working closely with the silicon manufacturers, a service capability that is superior to the majority of the players, if not all the players that you have mentioned, and the ability to scale. All reasons why we believe that not only will we be able to achieve a similar market share than historically in the thermal space or the cooling space, but we have an ambition to be above that because of the ability to be one of the first entrants and a large player in the liquid cooling space.
Pretty optimistic about that.
I'll have a follow-up to this after my NVIDIA question. As an NVIDIA partner, you're going to the market with reference designs for Blackwell today. You're partnering for solutions for Rubin tomorrow. You are clearly embedded in that ecosystem, not the only vendor, but one of three, four. It seems like people always say, oh, NVIDIA. It just seems they just pick that up to top three players, and that's how you become , which makes sense. How much of a competitive barrier is the NVIDIA partnership, both broadly and, again, more specifically on liquid cooling?
Yeah. I wouldn't like to call it a barrier.
Okay.
Simply because it has a defensive type of connotation to it, whereas we are much more aggressive in terms of going out and making sure that our strength is clear to everyone in the industry. It's the strength of our technology, our ability to scale. The relationship with NVIDIA is one of mutual benefit in terms of certainly the leader in their case, but us as a leader in the digital critical infrastructure. There are advantages in going to market in a joint manner. We have seen some announcement, for example, iGenius recently of large projects that we have won together.
Again, I'd like to think and to talk in terms of our mutual benefit and the fact that working on the future generation silicon that you mentioned creates really an opportunity to be ahead of competition, being ahead of the other players, and serve our customers with the technology that they will need in the future. Again, not defense, but if you will.
We will continue this conversation. It's probably a much broader topic. I remember late last year, I had a meeting with an NVIDIA executive at one of the industry events. He was basically positing that as power consumption and rack density goes up that w hat NVIDIA wants to see, they do want to see an integrated power and cooling solution, right? Because, one, with Rubin, once you're at 400 kW, or people are talking about 1 MW, you really want to have an integrated subsystem. I think it's more of a statement. It's a statement/. open-ended question. To me, Schneider by Motivair sort of tells you that what he said, what we expect of our partners is to sort of down the line be able to provide more and more integrated solution.
As I look at the business models, you versus other players, you are the only player that has scaled-up capabilities across cooling inside the cage, outside the cage, white space, gray space. Maybe can you comment, is that the right way of thinking about it, that the way forward is integrated subsystem, cooling, and power? Just what's the vision for the industry three, five years from now?
Sure. Clearly, the big driver is IT densification. And IT densification creates opportunities, creates challenges technically, but opportunity for the likes of us, of course, who have the portfolio, as you were saying, and the capabilities. We believe that integration will be an increasingly large part of the market and of the value proposition. I mean, that density cannot be easily addressed and solved just one component at a time. While the market for individual UPS c hiller, you name them, will continue to exist and to be good and strong, we see that integration and ability to do integration and densification will be a clear winner out there. We see it also in terms of the acceleration of our integration business, what we call infrastructure solutions.
Again, the market will choose different ways of building, but certainly the increased complexity of the solution from an IT standpoint will drive solutions that are thought through as a system and sometimes even delivered as an integrated system. We believe that having an entire portfolio, power, thermal, and now more and more also inside the white space is creating an opportunity in general because system level, you can optimize the solution for the customer relative to one product at a time, one point product at a time, but also the actual integration before delivery so d oing things in a factory alleviates the load on-site, and making our customers' life easier from a design and delivery standpoint is going to be a winner long term.
Excellent. Just talking about your cooling capacity, you literally grew by 45x in 2024. What's the capacity growth plan in 2025?
We will continue to grow. 45% actual is a pretty much, let's call it linear throughout the year, not exactly, but clearly the, let's say, capacity we had exiting 2024 was very substantial. We will continue to expand. There is a geographical aspect to that. We make sure that we have every region served regionally, continue that type of trajectory and add capacity. We want to be, and we have been deliberate and vocal about the fact that we want to have wiggle room and capacity to serve the industry in its wildest growth dynamics.
Is it fair to say we sort of track your gross asset additions? Is it fair to say that liquid cooling is going to grow in line with overall and maybe a little bit faster?
No. If we go back November or a few other occasions in which we talked about it, we believe that given the total cooling market for data center, medium long-term liquid cooling will represent about a third of that total, the rest being heat rejections, being air, et cetera. Clearly, going from zero to a third of the market, the growth rate is totally different than the rest. We have seen that growth rate in our numbers.
Thank you. Another thing that I think gets less attention, and I really sort of when I visited your facilities last summer, I just did not appreciate it, to be frank. Maybe can you give us an update on your chiller technology rollout in the U.S. and globally? How do your capabilities compare in air-cooled and water-cooled chillers by geography? Frankly, I did not appreciate the strength of the air-cooled portfolio.
No, absolutely. Strong air-cooled portfolio. We are developing and expanding our, let's say, water-cooled portfolio after organically and an acquisition that we made at the end of last year and a small technology acquisition that we made at the end of last year. Decades of chiller presence in the market in EMEA and in Asia. We decided about two years ago, a little bit more than two years ago, to go ahead and launch our chiller portfolio in North America for North America and satisfied about the growth trajectory. I mean, that was North America's zero market share for us. So it's all incremental, and we like the trajectory. We are with air-cooled chillers in North America, and we are launching, we are working on launching the water-cooled portfolio as well.
Is it going to start moving? Is it going to start moving the needle in 2025, 2026?
It is moving the needle as we speak, but the acceleration is happening.
Excellent. Thanks so much. I think we actually did talk about prefabricated components, but you clearly have a thesis about prefabricated components. You sort of touched on it as part of the broader long-term thesis, faster deployment, lower delivery. What percentage of revenue is tied to these integrated offerings today? Can you see that they're growing at above company sort of average?
Yeah. When you will see our numbers and mix, you will see prefabrication as a relatively small portion of the total. What we report is the pure kind of prefabrication value addition. We do not count into prefabrication all the technology that, all the Vertiv technology that goes in. There is a tremendous pull-through, obviously. If we have UPSs in a prefabrication or liquid cooling in a prefabrication, that all goes in the relevant thermal management, power management for obvious reasons. Otherwise, even market share pictures get distorted and everything else. Growth is certainly among the highest in the portfolio. I would say liquid cooling and prefab are certainly driving growth acceleration.
Excellent. Should we continue just sort of shifting to services? Should we continue to expect further acceleration in service revenues? Historically, services have been tied to UPS, three-phase UPS, but you're now offering service contracts on your cooling distribution units. I think there was an official launch of Vertiv liquid cooling services. How much does, why don't we focus on where we are with UPS, launch of Vertiv liquid cooling services, and then we can just sort of talk about total TAM services over time?
Sure. Historically, you are correct, especially true in the Americas, predominantly historically, but we started to change that years and years ago. It was UPS-centric. Already kind of a decade plus ago, we started to offer a much broader range of services across the entire portfolio of our technology across the entire install base. This is true globally. Do not think about kind of one product line as the service driver. It is the entire portfolio. Certainly thermal, certainly liquid cooling has a huge service potential, but it is not just maintenance services. I mean, we know that liquid cooling is extraordinarily critical for the performance of extremely expensive racks of GPUs, but also in terms of the ability to deploy and the ability to operate.
Both project services, as we call it, everything that is all the way to commissioning and getting the equipment running, and everything that we call life cycle services so the moment after commissioning throughout the life cycle. We see liquid cooling as a very important service driver. Again, we see it because we know that service is an incredibly important asset for our customers. The critical nature and the complexity of a cooling system, primary, secondary circuits, are such that require certainly a lot of capacity and capabilities and hence kind of being very public about our liquid cooling services.
Maybe a couple of things. I guess on project services, right? I think a couple of things have happened. A, data centers has rack down, density goes up, just architecture of data centers has become a lot more complex. The ratio between white space and gray space in the data center, I think, is changing dramatically. Some of the companies on the HVAC side actually have figured out that these project services is a real competitive advantage vis-à-vis competition that only sort of sells the product. How meaningful is this project services opportunity for Vertiv? Who would the typical customer be? Would it be a large colo? Would it be an enterprise customer? Do hyperscalers use these services?
Sure. Basically, everything that is medium to large projects for us is sold with project services. We absolutely support our customers throughout the installation and commissioning process across the entire spectrum of our technologies. That has been, let's say, part of our offering forever. It is a part of the offering that we have strengthened in terms of capacity and in terms of complexity of rendered services because the complexity of site has increased a lot.
Correct. That's what I'm saying. Yeah.
Absolutely. For us, it's not something new, but something that we've been investing in probably for the last two or three years.
Gotcha. Yeah.
We're very mature. We're very mature in terms of capabilities.
Just sort of going to.
One more thing, sorry. It's not just having the ability to serve the customer and the point project, but a lot of our customers are global in nature. So the same experience in Singapore, in Virginia, or in Ireland, that's what they expect and what we deliver.
Gotcha. Should we think, where could attachment rates on, I'll just call it HVAC, but on the cooling chillers, are the attachment rates comparable to UPS attachment rates? Could they even be higher? Are they lower? Just can you talk about what will the service component on the cooling side look like?
I think two dimensions. The rule of thumb is that attachment rate is a function of, first of all, how good you are at selling, how credible you are in terms of the services you render. The first kind of independent variable is how critical is the product, the technology. UPS is very critical. Liquid cooling is very critical. Chillers are very critical. Some types of air cooling are very critical. Certainly, some element of power distribution are very critical. That is typically the rule of thumb. We see more and more critical technology in a data center as, again, everything we discuss in terms of densification of IT is happening. That is a favorable environment for our services.
Also, what we see is that more and more the size of our data center, our customers' data centers is such that it's no more kind of a man and a van, and you have a truck roll and service it. It's more and more site-based services, more and more telemetry-based. The nature and the intensity of services changes, and so is the relationship. If you have with a customer an entire system, then you will have the entire system services to follow, and you probably have your Vertiv crew on site. The relationship changes dramatically, favorably for us.
Just to sort of finish with services, I mean, we've been reading, I think industry indicates that the rack architecture for Rubin implies rack density like, I think, 600 kW. How do you service a rack that's like 600 kW without electrocuting yourself and killing yourself?
We handle a lot. We handle a lot of power, power distribution. Our engineers are extraordinarily well-trained and professional, obviously, have been doing that forever in the gray space, if you will. Certainly, a question mark exists as the white space moves over time from kind of a low voltage, regular low voltage, sometimes low voltage DC to high voltage DC and further up. Could that offer a new opportunity for service, a service opportunity for Vertiv? Of course, we are looking into it very, very with a very, very keen eye, but I think it would be a little bit early to elaborate on that.
Right. Because the question I have, how do you send an IT professional inside the cage, right? It just completely changes the paradigm, right? Because you need a completely different kind of training. I mean, that's sort of the.
I do absolutely agree, man.
Okay. Okay. Fine. Okay, let's talk orders within the confines of what you're going to say. If a customer came to you with an order and down payment for delivery sometime in 2027, would that be booked as an order in the quarter?
In general, every time, an order for us is something that, has its binding legally, that more often than not has down payments and corresponds to a real deployment plan. The deployment may have started. The deployment may not have started. Again, it is the binding nature of the agreement so that there is no doubt about our right to some sort of compensation if the order is canceled. We do not call orders something that is hypothetical in the future. We have capacity agreements, et cetera. We do not call those orders. Order is something that is firm in the book and comes a certain date. It becomes material being ordered. It becomes manufacturing being kicked off.
Let me just clarify in this case. Do you need a specific delivery date, or could that be within that, if it is sometime in 2027, would that be booked as an order? It is up to the lawyers.
Typically, people do not say sometimes in 2027. Sometimes they give you a date because they know where they will deploy. Then maybe that project will be delayed or brought forward. Those dates, if they are somewhat out, can move. Let's go back to what we said at earnings call when we were saying, hey, guys, we do not see a dramatic or any meaningful material shift in the backlog phasing in terms of current year, future years relative to historical mix. We do not see kind of anybody going crazy and ordering just out there just in case. That is not something that we see happening. Plus, we want to keep our capacity available for real deployment, our stuff going on site.
Just to follow up on that, the pipeline in the first quarter grew even with the solid $2.8 billion in bookings in the quarter. That clearly implies there was no pull forward or pre-buy. If that's the case, what in your view was the reason for the sequential improvement in orders?
I think it's a combination of good market and a strong Vertiv in terms of portfolio, in terms of what we offer our customers and again it should not be surprising. I mean, we said, hey, we've been saying for the last multiple quarters, hey, our pipelines are going in the right direction. Hey, we see sequential increase of pipeline. We see a good level of stability in win rates. That translates into a solid trailing 12. We said, hey, don't be surprised, whichever direction the individual quarter orders may go because there is lumpiness in orders. I would say there's nothing different from the narrative that we have adopted probably for the last 24 months or so.
No, you guys have been consistent.
Twenty-four quarters. We always say, guys, the market is going in the right direction. Do not be spooked or overexcited by a single quarter performance, but look at the sequence. I think the sequence is pretty convincing.
Can I ask you, just looking once again, looking at the industry press, just the timing, there is a lot of speculation about the timing of rollout of GB300 versus GB200. I think the schedules sort of move around. What kind of impact does it have on customer conversations? Because part of what I have read is that, hey, because GB200 is being pushed out, now GB200 is closer to GB300, why not wait until GB300 and upgrade your design? As long as you are waiting, may as well upgrade the design and go with GB300. Another conversation, as we keep moving to the right on Blackwell, Rubin is getting closer. Once again, you would know the answer. I do not know the answer, but there is speculation about the timing of the initial Rubin rollout.
When you talk to the customer, right, h ave you seen real customers move their schedule around because of these moving pieces that clearly have profound implication on what the architecture of the data center looks like?
I think three elements to this. One, I think everyone is overanalyzing kind of a 200, 300, 350 , whatever. It is overanalyzed. I mean, there is hunger for GPUs in the market. And it is true in the U.S. It is true globally. The other element is I think this question is better asked to our NVIDIA friends. What we see in the market is demand for infrastructure. We were talking about pipelines and the growth of pipelines an instant, a minute ago. That stays. The other thing is that we should not think about an infrastructure that dramatically, dramatically changes between 200, 300, et cetera.
If someone is building today a data center tailored on 200, I think they're crazy because we know that at least, if not because of the dynamics we're talking about, but because of simply obsolescence of cycle of their IT, that very data center probably will need to upgrade to 300 or more in the future. That would be all those things start to be kind of factored in the way it's designed today. Anyway, the orders we receive today are for data centers that will go up 9, 12, 18 months from now. We are ready in that.
Right, right, right. Okay. No, that actually, that's a great answer. Thank you. Maybe how are you thinking about your supply chain post the tariffs, and are you making adjustments to your own manufacturing footprint, and have you considered new suppliers?
I think I want to go back to the post-COVID. I know that feels like kind of a millennia ago, if you will, a millennium ago. The fact that we have been working on the resilience of our supply chain and the way we design products and how many suppliers per each component we define upfront when a new product is launched and during the last cycle of product has been kind of a phenomenally helpful capability that we did not have and is serving us very well right now. To your specific questions, the answer is yes, but always. There are moments in which such like kind of a redefinition of the trade flows globally give us an opportunity to adjust further, but it is an ongoing exercise.
Having more region-for-region type of manufacturing and supply chain is something that we have kind of shaped, been shaping for the last two years. Again, we are adjusting that further for the current and fluid tariff situation. Yeah, absolutely, and v ery, very, very happy about the ability and capabilities and speed of implementation at Vertiv right now.
Do you think this will cause more of a supply chain sort of move, let's call North America? Do you think people will adjust, the industry will adjust, or do you think?
I think the industry has been adjusting for some, c ertainly, Vertiv has been adjusting for some time, but yeah, that's a natural trend.
I know you have a very different industry, but the budget, do you guys care about accelerated depreciation? Will that impact how much and where you spend, like the provisions of the tax bill?
We like things that are pro-business in general. We think the industry we are serving anyway has its dynamics. We serve a telecom and commercial industrial industry, those are a smaller portion of our mix. If that is favorable, fantastic. If there is favorability for us, that will corroborate some of the decisions we made in terms of what to manufacture and where to invest.
Excellent. Thank you. What are your conversations like with sort of traditional Colo companies these days, f rom the brokerage reports, rents are rising, occupancy levels are high. What do they want to talk to you about?
Certainly they want to talk about capacity. They want to make sure that we are there for them. Very importantly, very importantly, they talk a lot to us about what their infrastructure, what their design should look like two years, three, five years down the line. That is the best conversation for us because it is where our role as an industry thought leader, and like I like to say, the connecting tissue in the industry between silicon and owner of data center is really being fully exploited by the industry and something that we like a lot. We like to take the industry where it needs to be three, five years down the line to really accelerate everything, IT infrastructure and AI infrastructure.
Pre-COVID, your top 10 customers generated about 25% of your revenue. Has your customer concentration grown meaningfully with the hyperscalers' investment in AI?
We continue to have a very, very broad customer base. We are not disclosing things specifically, but directionally, not really materially.
Right. Can you just talk about what's happening in China because you do have strong presence in China? Is it a different ecosystem? How is ecosystem evolving in China? Is it different from the rest of the world?
China, of course, is a very competitive place. Certainly, we see good dynamics as we explained in the last couple of earnings calls. The competitive landscape is more different as well, but it is not unique to China. It is the same competitive landscape that we see in Asia, the rest of Asia. Not that technology is dramatically different. There are kind of some idiosyncratic type of system designs, but nothing that is not leverageable outside of China in terms of know-how, experience, product costs, focus on efficiency, all things that, let's say, overflow very well for us across the world.
We are right on time, Gio. Thank you so much. Thank you for opening day two. Clearly, we'll start tractions.
Andrew, you're always great. Thanks a lot.
Yeah.. Thanks so much.