We're together last time. Great. Well, thanks, everyone, for being here. It's my pleasure to have next Vertiv, and I think this is sort of the first time really at this event. So thanks very much for being here on stage with us. It's my pleasure to have Gio Albertazzi, CEO, David Fallon, CFO. Maybe just start off with a sort of very live topic around kind of orders. They're naturally lumpy given the business, the nature of sort of shipping greenfield equipment. But people are always trying to understand sort of where do orders move from here. Have they been at a high rate, a normal rate versus how you see customer CapEx? What should we expect orders to do from here?
Certainly, we are very satisfied about the order trajectory that we have demonstrated during the course of 2024, as represented by our trailing 12-month of at 30%. That certainly is in the direction we want to go. Even if you take that in absolute terms, that certainly is robust orders, robust backlog creation, and certainly reinforcing the model that we have shared with the entire investor community. When it comes to looking out in the future, I think we should reiterate a couple of the messages that we have shared with the community Wednesday last week. As we think about 2025, of course, we think about a year in which there will be our book-to-bill will be greater than one. Backlog creation.
On a trailing 12-month basis, I mean, first off, we come from very high comps, even in absolute numbers. As I said, don't be shocked if trailing 12 normalizes. We feel very good about our pipeline. The news across the business and across the industry, that's what I meant, across the industry are strong as the example CapEx public statements of the hyperscalers demonstrate. The industry is alive and well and heading in the right direction. We have certainly a very, very prominent role, and we are even taking market share as we look at our numbers and orders and revenues.
That's great. And when you think about the sort of breadth of your business within Vertiv, you're supplying all types of customers. How much of your growth, say, the last 12 hyperscalers specifically versus, say, colos or the traditional on-prem corporates?
Yeah, we do not talk about hyperscalers in isolation for a simple reason that it's very, I think, not a reliable picture. The hyperscaler business, any type of hyperscaler business really turns into, from an infrastructure standpoint, both hyperscaler and colo. A lot of the build-out, the infrastructure build-out happens on a leased basis. So that's why we consolidate the two. And when we look at 2024, think about hyperscale and colo. So colo, what we call colo cloud, represented about 50% of our revenue. So we don't necessarily talk in terms of revenue. So certainly an important portion of our total business. It's a broad base of customers because it's a large base of customers. And certainly, that demonstrates that we participate in that space very, very well.
Great. And if we think sort of geographically for a second, I think there were some questions around the EMEA outlook. Last year, you had exceptional growth, I think 20% plus in the region. I think most companies at this conference are probably minus 2% in the region last year. So you've got a very different base. Maybe help us understand sort of how you see EMEA the year ahead.
In general, we see certainly not an extremely favorable regulatory environment and some power headwinds in the region. That is what we believe has caused some of the lumpiness in our EMEA order numbers that we have been very, very explicit about in the call last week, but when we look at the pipeline, when we look at the meaningfulness of the customer conversations we're having in the region, in EMEA as a whole, then we will feel very good. We know that there is always a lag between what happens in North America, and historically, there have always been a lag between what happens in North America and Europe or EMEA more broadly defined. This is still the case. Hopefully, Europe is waking up to the need to accelerate AI. That awakening is in the pipelines, and we look forward to that turning into a real acceleration.
Got it. And maybe in Asia, clarify within China, kind of what's the exposure there? What types of customers are you selling into? Any sort of thoughts around that?
China for us is shy of 10% of our revenue, more or less. We are a Chinese competitor in China in the sense that we are truly, from a technology, from a supply chain, from a presence, we're Chinese in China. We participate in the broader business there, the broader spectrum. We have certainly a good enterprise and colo and cloud presence, but also a very good commercial, industrial, and telecom presence. We believe we have quite some meaningful strength. China's not only interesting because China as a market, of course, is interesting, but for two reasons. One, China is a phenomenal base to, let's say, springboard for Southeast Asia. Very, very important. The second is that competing in a very competitive market like the one in China keeps you very fit globally. We're very proud of what we're doing there.
And what do you see there in the way of sort of local competition? I think people worry that with all the data center growth globally for equipment the last two to three years, it's pulled in new and incumbent companies to add capacity, try and take market share. How do you see sort of that competitive landscape right now? And maybe start off by describing what you see in China on that front.
Well, again, we tend to think about the data center industry as an industry. All of a sudden people, oh, we have competition in the industry. Oh, things are going to change dramatically. Well, guys, we've been operating in this market for decades, multiple decades. And this has always been a competitive space. China has always been competitive. So very respectable local competitors. We all know their names. And that competitive intensity has always been there. If we go outside of China, some of the players that people think, oh, they are new, they're not new. If you take some of the HVAC players, for example, in the North American markets, they've been playing in the space for decades. And some will look at Vertiv and say, oh, Vertiv did not have a chiller offering in North America. Now we have. So it goes both ways.
So we are very accustomed to a very competitive landscape. What we see in North America and in Europe is some elements of consolidation. We heard about acquisitions. We see rational competitors consolidating the market. So when we have rational competitors as competitors, it's not a bad place to be. But again, don't think for a moment that these industries have not been competitive. When you think about hyperscaler, when you think about colos, you always think about very commercially and technically savvy players. And they have been so for quite some time. So there is no kind of a, oh my God, the rude awakening. Not at all. Been very awake for the last couple of decades.
That's good to hear. And if we think about sort of the capacity additions, they're trying to keep up with strong demand growth. How do you describe sort of the outputs from that supply-demand balance in terms of, say, industry pricing for some of the main products you compete in? And then also the sort of lead times for some of the major.
Sure. I think, again, when it comes to supply and demand equation and how that translates into price, well, it's only natural that the industry that is growing quite fast, and if you go back to our investor day November 18th, we were talking about if you isolate the colo and cloud space, then think about a 15%-17% over the next few years type of growth, which is a lot of growth for an industry that is fundamentally a construction industry. That is clearly attracting attention and attracting capacity. But that capacity and attention and attraction is justified by the fact that there is growth. We continue to think that the demand-supply equation is favorable. Having said that, the only value that long-term pays off is the one that you can build on the value that you create for your customer.
The value created to your customer is based on technology, innovation. It's on quality and reliability and service level. Put all things together, and it's a perennial race being ahead of competition here. It's that what drives long-term the value to the customer and the value that can turn into value for us. These are pretty much the dynamics we are in.
It seemed like sort of a year or two ago, there was a lot of capacity constraints, tough to supply the customers. I think in areas like chillers, we've seen those lead times maybe normalize for the industry. How do you see those lead times for kind of some of the other big busway, switchgear, UPS?
Yeah, the second part of your question that I didn't address. Thank you for going back to that. So certainly, there is some of the lead times are coming down. There are some categories that could be outside of our portfolio. As gensets , you have transformers that are still quite long lead time. So sometimes they dictate the large projects, overall project plans, durations, and critical paths. I can speak for Vertiv, and we have been deliberately taking lead times down. I think the industry needs shorter lead times, and again, I was talking about service level as one of the value levels. And we strongly believe that that is essential. Short lead times for us means a more finely tuned supply chain, so a more effective and productive supply chain for us to our customers, but also means that we can build competitive advantage.
So if you look across our portfolio, we think that actually we can certainly say that we have been gradually reducing lead times. That doesn't mean that our customers order on shorter lead times. Again, very often, and I was talking during the calls about a requested lead time between eight and 18, sorry, nine and 18 months in general. That's what the customer requires when projects are big because that's their project plan. But the ability to be fast, I think, is essential from a competitiveness standpoint. And we're driving in the right direction.
When you think about some of the larger customers that you probably interact with more than others just because of their wallet share and so forth, how much visibility do you have from them on their spending plans? How has that changed? Not so much a formal lead time, but the discussion and framework.
Sure. Of course, it depends on the relationships you have with customers. But the strongest relation typically gives us a sense two, three years out of what their plan is. Of course, the further out, the more fluid the situation is. And that's very much normal. And that goes back probably a comment that needs to be made about our backlog and our orders. Everything we call backlog and orders is something that is corroborated. And it's based, not corroborated, it's based solely on an actual legally binding PO. But there are a lot of conversations beyond that extended much longer, let's say, window.
Got it. And so you've got next 12 months, it's very good visibility. And then you sort of think about the comps. And that's why you have this dynamic where the year as a whole will have good backlog growth, but probably faster in the second half than the first just because of the comp base.
I didn't say that. I didn't say that. Because if you take the first, I was talking about TTMs.
Yes.
But if you move the conversation in absolute dollar numbers, that could be different. But I don't want to over-elaborate. And I go back to the comments. If that's okay with you.
One broader question, I suppose, is on electrical intensity of data centers. Because we have seen that cycle up and down depending on generations of type of data center. Went up 15 years ago, down a decade ago, been going up more recently. And you have been one of the few companies who's given that sort of $275,000-$350,000 per megawatt number. So when we think about that number, any kind of historical context, maybe where was that number X years ago? Where could that number go later this decade?
Yeah. That $275-350 that we gave as a range pretty much defines what we think the evolution of design current and evolution in the next few years. Going back to the past, it's probably a little bit hard to say. It may not have been five years ago as accurate in estimated spend per megawatt, but certainly one thing is sure. As the density, call it compute density, rack density increases, the infrastructure becomes more demanding and more complicated, and hence the TAM for the specific spend per megawatt installed power has a tendency to expand. That said, we believe that that range that we indicated is a good representation of that. It corresponds very well to the value model that we have shared with everyone a couple of months, three months back.
And one area, I suppose, where you've added to the TAM perhaps is organic and inorganic investment on liquid cooling. You clearly have a very strong pedigree of precision air cooling. You've added chillers. You've added liquid cooling. So your sort of cooling capability has gone up substantially within the data center. How do we think about sort of the growth rates, you think, across those parts of the cooling landscape? Liquid precision air, and then chillers.
Sure. And let me add one thing to the previous question. And I think that was important. Because if we go back a few years ago, our TAM per megawatt would have been lower because our, let's say, portfolio would have been narrower in the power side of the equation. So the E&I acquisition, of course, expanded that TAM because we were missing. But in general, back to your liquid cooling and cooling in general question. I think what we were saying a year ago or a bit more is all the more reasons very true right now. That time will tell. Technical evolution will tell. But we see liquid cooling pretty much as a complement. When you think about a data center, you will always have, of course, a cool, sorry, a heat extraction from the server. And that's basically what liquid cooling does.
If you think about an air-cooled server or chip, well, that part of that TAM didn't exist for us because that was just a fan on the chassis of the server. So that's a net TAM addition. Air is a given. A given for two reasons. One is that every, let's say, rack or even cluster has an air-cooled and a liquid-cooled portion. A year ago and still valid, we were talking about, hey, think about a 100 kilowatt rack. And now, of course, it looks even small. But we said, OK, there will be probably 30% of that if it's liquid-cooled that will still be air-cooled. So air and liquid coexist and will coexist for the foreseeable future. And I'll elaborate in a second what that means. But then again, you have to do something with that heat. And that heat needs to be ejected.
So you can have rejected, let's say, that heat in the atmosphere or, we believe, increasingly so as time goes by, reuse. Now, uber simplified way and not necessarily liquid, you will have air in the halls or in the aisle. And you will have chillers or other heat rejection type of tools. And it's probably, think about a liquid cooling constituting probably at maturity 30% of the total cooling TAM. That could be kind of a reasonable expectation out there. But an interesting thing that I think I want to comment on is that it's not just about how you cool your perfectly designed data center today. It's how you make sure that your data center will have a kind of the asset will have probably 15-20 years of useful life.
In those 15-20 years of life, there will be all sort of mix of high-density racks, low-density racks, liquid-cooled GPUs, air-cooled GPUs, CPUs, all sorts of things. So what we see is a lot of the design now over-provisions cooling. So you will have, if you put together the liquid cooling and the air cooling, that will be more than the nominal power installed in the data center because people want to build the overtime resilience that the infrastructure needs to really weather all the possible different mix of high-density, low-density loads that will be thrown at them during the life cycle. So it will stay. It's there to stay.
Then strategically, I guess, you took that decision to get bigger in liquid cooling. I don't know if there's sort of other areas within the data center that you think are attractive that you're not in today. Then sort of more broadly, beyond the data center, do you see interesting applications for electrical equipment that you'd like to lean into just from a risk mitigation or breadth of end market perspective?
Starting from a comment from the founder, we certainly believe we have, and we know we have a very, very strong portfolio in the data center infrastructure and a very complete or the most complete portfolio in the data center infrastructure industry. Certainly, the technology is evolving. Just like we did with liquid cooling, we invest a lot in innovation. But that investment in innovation is always checked against the, let's say, inorganic option when that is needed. You saw us doing a relatively small but very focused technical acquisition in December last year with DSE, focused on the chiller market and chiller portfolio. And again, that really corroborates what we were doing organically and accelerate. There will be more things like that. And the technology will evolve, the technology, the demand. And we will stay ahead of that curve.
But as you were saying, a lot of what we do is quite fungible outside the data center. So switchgear, UPS, power distribution, but also chillers, et cetera, et cetera, are all technologies that can be applied outside the data center. Our commercial industrial does exactly that. It's 10% of our revenue, more or less. If you think 2024, we like the way it's growing, but it's certainly outpaced by the growth in the data center. Having said that, it's a space that we like. And I think being vocal about the fact that we will continue to look at that space with a lot of interest.
How about the sort of electric utility type markets? Or is that too removed from the core?
You know, critical infrastructure for us predominantly is a behind-the-meter type of activity. So we do behind-the-meter very happily microgrids or battery storage systems. And we have quite some success there. But we will not go into grid-level type of market. It's not our market, if that was the question?
Yes. That's clear, and then you've had this sort of equipment boom. Why is there? I think you're also talking about services and spares and sort of making sure you're kind of putting your best foot forward on that front. On the services front, kind of is there any way to make it more recurring or contractual as opposed to kind of spare parts, and what are you doing to flex that?
Absolutely. The vast majority of our service businesses, more or less 22% of our revenue, is predominantly contract-based recurring. That's our model. We are very diligent and have a long tradition of mining and going after the installed base, not only because it's good business, but also because that's a fundamental promise that we make to our customers in terms of the service level that accompanies our value proposition, and it's an important element of our value proposition and even more important as the infrastructure becomes more critical, and that contract type of service business that, again, is the majority of our service has been augmented by more digitization of that offering, more telemetry and predictive maintenance. It's growing. I mean, it's maybe early stage, but certainly growing and very central to our value proposition and to our strategy.
Last quick one. Many sort of industrial companies have been seduced by a push into software inorganically. Just your broad thoughts on the merits or demerits of that.
Let's say, as I said, we believe that the digital element of our offering is an important value contributor for our customers. Never say never. But in this moment, we know what creates value for our customers, and that's what we focus upon.
Very clear. And with that, we should switch quickly to audience response questions. So the first question is around current ownership of Vertiv. So 60% no. Second question is around sort of general perspective towards the name. So fairly positive bias. Third question is around earnings growth outlook versus the multi-industry average. So 90% above peers. Fourth question is around uses of excess cash. So it's a bit of a mishmash, but generally share buybacks. Next question is on the PE, the stock should trade at. So generally a premium to the market. And then the last question, I think it was sort of 60% non-owners, kind of why is that? What's the single biggest reason? So core growth. But I think it's been OK. Anyway, with that, thanks so much, Gio for being here.
Thanks a lot. Thanks everyone.