Awesome. Hi, everybody. Mike Halloran here, and we are pleased to welcome Vertiv with us. Gio Albertazzi, CEO, is going to give some really brief, prepared remarks, and then we're going to dive into Q&A. As you all know, I don't cover the company. I've got a lot of good questions. However, if there's anything you want to talk about, raise your hand. I'll call on you. Most people don't do that anymore, so send an email to the card in front of you, and I will make sure to incorporate the questions into the conversation to make sure we cover what everybody wants to cover today. With that, Gio, the floor is yours for a couple of minutes.
Thank you very much. Good day, everyone. Very glad to be here. Most of you, I'm sure, know Vertiv, a global leader in digital critical infrastructure. 80% of what we do is data center, certainly a sector that is increasingly of interest. Very strong market growth, that growth that we believe will be strong and long-term, and a very strong position in this market with a truly complete portfolio that stretches from the entire power train, the thermal chain, white space, a very strong service organization that is one of our superpowers, and a lot of prefabrication. You might have seen, if I haven't, I recommend you do, we launched a fully prefabricated data center core, Vertiv One Core, basically reference design around GB300 and soon to come Rubin and future NVIDIA chip designs.
That is, let's say, the tip of the spear, the representation of all Vertiv capabilities, the entire data center infrastructure. This capability and the breadth of our range make us a privileged interlocutor for, on the one side, NVIDIA, and the other is the large colo, the hyperscalers, to really define what the long-term strategy deployment is. That, of course, gives us not only a seat at the design table, but certainly a preferential understanding of where the industry is going. That translates into quite some strong performance, as you have probably seen from our earnings call a couple of weeks back, in terms of actual growth and expectation for the rest of the year. I am very proud of what we're doing and certainly very happy about the visibility that comes from strength in an increasingly visible industry. With that, that's going to a Q&A.
I think that's a perfect segue. Why don't we talk about that visibility? I know in the recent call, you've talked about good visibility into 2026. Maybe why do you have that conviction short term? When you think a little bit longer term, three, five plus years, what gives you confidence that there's durability to this cycle? If we go back historically, this could be one, two years on, one, two years off. This feels like it's got more legs. Maybe take those two different time horizons and talk to the guy.
Yeah, absolutely. Far from, of course, talking about 2026 necessarily, but if you look at our position today with strong Q3 orders, both trailing twelve and very strong year on year, combine that with a 30% backlog growth, we've been vocal about the fact that the shape of that backlog very much resembles the shape that we had before in terms of the time horizon. So it's a backlog growth, not by virtue of a natural elongation, but by virtue of true and good growth of backlog. Combine that with strong pipelines that continue to be strong and growing.
That is certainly injecting a lot of short and medium term and long term confidence because our pipelines, or call them funnels, the way you want, but everything that is commercial activity, and we are pretty maniacal about managing it in a rigorous way so that it becomes a leading indicator for us. They are all pointing in the right direction for the immediate and medium future. If we go for kind of a longer term, then what corroborates everything we say and the fact that this is going to be kind of a long, durable growth in the industry comes very much from what we see happening in the conversations with our customers. We have a lot of visibility in the industry across hyperscalers and colocation and enterprise. There is a lot of investment coming. We see that coming from the silicon side, the hyperscale side.
We see a lot of growth on the neocloud side of the spectrum. Let's not forget that the bulk of the growth that we've seen today in the market in general, it's true also for Vertiv, is North America, U.S. based. There are a lot of other parts of the world that are still to be seriously developed in terms of AI capacity. That is something that is really not negotiable. I mean, there are a lot of sovereign data reasons for which this capacity needs to grow. If you want to do inference in jurisdictional territory, you have to have capacity locally. This growth will continue and expand beyond North America. One thing that we like always to keep a very sharp eye on is the trajectory of AI-driven monetization and business growth for our customers.
If you look at our, if you look at the hyperscalers' most recent earnings goals, you see that growth rate of their AI-driven business is really robust. All things that converge and that make us very, very optimistic and continue to believe in the market for a long, long term.
How does that play out then to revenue for you guys or even orders? You get a gigawatt scale announcement for a campus, right? When does that start mattering for you? How much visibility in advance do you get to what your win entitlement of that is and just kind of translate it?
Sure. The announcement that we all hear typically are done in one bulk and that the deployment is done in stages. Those stages may cover one, two or more years depending on the size of the announcement. That is all very, very noble. If you think about, say, 1 GW, which is, by the way, mind-bogglingly big if you look at it through the lens of five years ago. Now we are all kind of accustomed to kind of 1 GW is, of course, but it is a lot of power being deployed. What we see is really encouraging. That, for example, can be deployed in 200 MW chunks. Each 200 MW chunk probably goes through a cycle of order to deployment of between nine and 15 months for us, and the whole construction will probably be a little bit longer than that.
Of course, the sequence of, let's say, phases follow. That deployment can stretch over a couple of years, three years, a little bit less. It really depends on the project, but this is the order of magnitude.
Yeah. When you think about that opportunity and the competitive set there, one, maybe just talk about the competitive dynamics in the marketplace and maybe what your win rate looks like now versus history, however you want to phrase it. The secondary piece of it is how do you think about double ordering in the industry, risk profile associated with that, how you manage it?
Clearly the industry is of interest, and we hear a lot of people talk about kind of a play in the space. When you really look carefully, you'll see a number of players that are consolidating the industry. We all have heard about a few acquisitions lately, and I'm not talking necessarily last week or the last few weeks, but over the last few months. From our perspective, the industry is not dramatically different than it was before. Take liquid cooling. Liquid cooling, clearly a new technology a couple of years back, and certainly from a technology standpoint, still pretty early stage, something that has been deployed in the last couple of years. Again, we saw some startups, some new players, those new players being consolidated.
The industry is consolidating back to a number of rational players, you pick the names, rational players that are addressing the market. This is not a bad place to be. Actually, it's a place that we like. You see some players replicating or trying to play the same role that we play in the industry. The entire portfolio, two things. That is testament that our value proposition is winning. The second is we do not mind. An industry that is built by a number of players that have a good value proposition and are rational is something that certainly is good for everyone. It's good for customers. It's certainly good for us. What was the second part of the question you were saying?
Double ordering.
Double ordering. I do not see double ordering in the industry. Basically, everything that's being done is done against a project, a deployment. It's not just placing orders out in the vacuum and hoping that they will stick. Whenever there is an order, as now a case, we have terms and conditions that make it very painful to walk away from the order. We typically call an order a PO that's legally binding. We typically know the shipping address and everything else. It's pretty firm. The industry and the cancellation frequency is absolutely negligible and nothing different than we have seen historically in this industry.
How are you managing your own capacity in the context of this? I know you like to stay ahead of the curve, but this is a curve that has a lot of legs. How are you thinking about what your capacity needs are and how quickly you can or cannot ramp?
Sure. It is a constant expansion of capacity. We have been vocal two weeks ago when we had our earnings call about the fact that, if anything, we are accelerating the speed of capacity expansion, but we have been expanding capacity quite clearly and constantly over the last years. Capacity comes from net footprint increase. It comes from efficiency and productivity, something that also you see in our profit growth. Again, guiding 27% for 2025 growth, that needs capacity. There is capacity that is at play today and will continue to expand. It feels very good. Capacity expansion for us always leaves wiggle room in a sense that this is a project business, so you cannot expect everything to be exactly linear. We have capacity to absorb some fluctuations in demand, but think about long-term capacity constantly going up.
There's a handful of questions here kind of asking a similar thing, but what are the constraints from your perspective on data center growth? Are there choke points from a supply chain? Does zoning become a problem getting these things moving in? I mean, any constraints you think about in terms of that growth profile, assuming the demand curve that people are saying is actually there?
Sure. First of all, the demand is there. What we see is all the points are leading, indicating that demand is there. I do not think in terms of cliffs or ceilings that limit the ability for the industry to grow. Two years ago, a year ago when we had our investor days, we were saying, hey, we think about, for example, a hyperscale colo sector growing 15-17%. Now we certainly are in the upper part of this. Again, it is not kind of an infinitely steep growth as, if you will, the appetite may indicate. We have always been vocal about the fact that power is a pacing factor, power generation and power availability, that permitting is a pacing item, and that this is a construction work. Access to skilled labor are pacing items. All are being addressed.
I continue to see growth in the industry. Take power. A lot is foot in terms of behind the meter power generation. The industry is very vibrant. There is a lot of capital and a lot of ingenuity at play. We see that these pacing items are being addressed. Again, take the power behind the meter power generation. We see it in turbine guys that certainly see their backlog moving in the right direction a lot. If you look at it from our perspective, that makes our power train from power generation all to the chip more complex. We see that as an opportunity. Everything that we see in terms of ways to address the pacing items in the industry turn, we believe, into opportunities for us.
White space within the data center, how much of an opportunity is it for Vertiv and similar companies to address that? How do you capitalize on that?
Yeah. You know, I've been in the industry for more than two decades or so. It was up until not a long time ago, white space was not an attractive space for us. Yes, we were present with racks, with some in-rack power distribution, but basically the IT stack in the white space was not challenging. When you have a 3 kW, 5 kW, 10 kW rack, yes, there are businesses you can have in that space, but it's not phenomenally attractive. Now we see that with a high density, 100 GW, 200 GW, 600 GW rack, then the physical coexistence of IT power and thermal is such that the challenges from a power and cooling standpoint are such that that space is extremely interesting for us. Hence the acquisitions, for example, of Great Lakes. If you really think about liquid cooling, liquid cooling happens in the white space.
If you think about high voltage DC, we all heard about the gradual moving towards 800 VDC for power distribution, high-density racks. That makes power distribution within the white space way more complex and way more demanding also from a health and safety standpoint. All technology that goes into the white space, making the white space an area where we can thrive as opposed to an area where, yeah, we're present, we have racks, we have a couple of technologies. Some of you may remember from our investor call in July where we were talking about Smart Run as a white space fit-out, prefabricated solution. It's a technology that really incorporates all the Vertiv technologies and can reduce the fit-out time by an order of magnitude.
That means that if you are a data center owner at the moment, your data center is ready is when you start to fit out the white space. If that takes you three months, four months, six months, those six months are what separates you from starting to have an asset to the moment you start to generate revenues in that asset. The ability to shrink that while you deliver technology, phenomenal.
You referenced the prefabrication capability set. How important is that broadly? How differentiative of an approach can that bring to the table? What can that help solve for you?
Sure. Sure. I would like to, let's say, layers of things. First of all, we're very strong in individual technologies. We'll continue to be focused on making sure that at individual technology level, I don't know, liquid cooling, the UPS, the switch gear, the chiller, we have kind of a top of the range technology. We talk about systems, the power train, the thermal chain, the white space. It's not a layer where the system is optimized from a management standpoint, from an interoperability standpoint. The next level is the ability to deliver all that on a prefabricated fashion. That prefabrication really solves two problems. One is time to market for our customers because you can really shrink the time it takes to deploy capacity for a data center owner.
The other is you address one of the pacing items that we were saying that is on-site construction, skilled labor, of which there are certainly constraints. All things that are addressed. It is a multi-layered approach, but certainly technology is what drives it all.
How do you think about service in the context of all of this, both in terms of your capability set, how much you want to lean into it, how you handle it from capacity expansion? Just holistically go through that thought process.
I've sucked from the last point. Service and capacity expansion. When people talk about capacity, and we were saying the same, people naturally think about how many sq ft, how many factories do you have? For us, it is exactly the combination of factory, supply chain, and service. What we are seeing in terms of densification of the data center requires more and more service. The infrastructure is becoming more critical. We have north of 4,500 field engineers and growing and counting. We really know for a fact that services is a superpower for us. The ability to support our customers not only with technology, but also commissioning and life cycle, absolutely important. You see service as really central to our value proposition, to our growth strategy, but also technology evolution.
We made an acquisition that is small, but very critical in terms of predictive maintenance capabilities and system and data center performance optimization. Way late. We announced it in July, if I remember correctly, but anyway, they're out. Again, we see technology, capacity, all coming together at service level.
That makes sense. Maybe before we go back to some of the other hardware stuff, just the margin thought process, mid-20s by the end of the decade here. What gets you there? Is service part of the equation? Does that come in higher margin, lower margin? More importantly, what's that path? How do you see yourself evolving towards that range?
As you said, in a couple of occasions, our model is very much intact, and we continue to drive in that direction. I think that Q3 is a testament to that. Certainly, service is accretive to us. It is not just accretive. It is also underlying and supporting our entire value proposition. We like it a lot. Service, clearly, in a period of steep growth of everything product, let's say, as opposed to service, is lagging, but it is normal. We are building installed base, and the installed base has been harvested. It is absolutely normal, but we see an acceleration in services. More in general, I think two components. There certainly is price cost that continues to be positive going forward. There is an efficiency play. The biggest lever, the biggest contributor is operational leverage.
We have been delivering operational leverage, margin growth through operational leverage quite consistently in the last three years.
Couple of questions here. Balance sheet. How do you look at the balance sheet deployment from here? Ability internally to meet the demands? What the optionality looks like from an M&A perspective?
Oh, yeah. I think we have a strong balance sheet. Of course, that balance sheet is fundamental to continue to invest in capacity expansion. We are investing a lot in technology. I always like to say that the portfolio that we have today is very important, but it's worth nothing if the portfolio that we have tomorrow is not even stronger. In times of big dynamics in terms of technology, focusing on the future portfolio is absolutely of the essence. 20%+ growth in engineering and R&D, that will continue. You have seen us parallel path, organic and inorganic technology investment. Technology, portfolio augmentation, and being ready through M&A for future technology is a parallel path. When you have a new technology coming up, you can't just rely on your organic.
You have to spread, if you will, your bets and have partners that work for you with you in future technologies. That often, as we have demonstrated, can turn into an acquisition or not, depending on the path. Certainly, technology bolt-ons from an M&A standpoint will continue to be an important path for us. Again, we would not shy away from something that adds simply capacity, go-to-market, and market share on parts of the technology that we own already, as long as it's rational and it's not crazy from a valuation standpoint. We'll always be rational in how we allocate capital. We clearly have also still a lot of wiggle room from a repurchase standpoint, but with the dynamics, as we said, we will be opportunistic in our repurchase decisions. We like the space. We like to grow in the space.
Given how quickly the space is evolving, how do you think about collaborations with end users, customers, however you want to put it? The question that came through here is more NVIDIA, the 800 VDC. Maybe use that as an example of the collaboration piece and what that could do for you, what kind of adoption curve, however you want to go about it.
Clearly, we can look at the 800 V DC just as a, let's say, parallel to what has happened and still happening with liquid cooling on the thermal side of a data center. It's a future technology. A lot has yet to be written. We work with hyperscalers and certainly with NVIDIA and others in defining what that technology really will look like. The likes of NVIDIA will work with us and very few others at all because they want to make sure that there are people out there that can pioneer the technology that then will support their future deployment. If Rubin is out late 2026, early 2027, we got to be out at the beginning of the second half of 2026. That's what we have committed to with 800 V DC.
That means that the power train, as we call it, will evolve gradually to an 800 V DC, but you will see that it will not be kind of a clear cut. There will be a lot of the current technology continue to be there, the future technology, the 800 V DC, hybrid solutions. We are there to manage this transition just like we're managing the transition and the coexistence, the juxtaposition, dare I say, between air and liquid cooling. The same will happen here. Having a partnership like the one we have with NVIDIA gives us kind of a first mover, very early mover advantage that enables us to shape the way the industry thinks about high voltages.
Last one, the minute we have left here, hyperscale and colocation is driving the North America growth here. Do you think there's an opportunity for enterprise to start being additive? And how would you think about that going forward?
There is also new cloud that is a big player in this equation. I think when we talk about enterprise, two things about enterprise. You could have enterprise on our own proprietary data center. A lot of the enterprise is going to happen even for AI in leased and colocation space. We see that happening, gaining traction. Let's say on-prem enterprise is growing kind of a mid-high single digit, but do not take it as the old AI that is happening at enterprise level. There is a lot of AI at enterprise level that is happening in colocation. So enterprise proprietary, but on data center, leased data centers.
Great. Please join me in thanking Gio and Vertiv for their time today.
Thank you.
Thank.