Welcome to today's webinar hosted by Investor Relations. My name is Line Dovärn, and I am Director of Investor Relations here at Munters, and today we will be focusing on Data Center Technologies. I have some very eager speakers with me here in the studio, ready to get started: our CEO and President, Klas Forsström, President of DCT, Stefan, Vice President of Strategy and Finance, Frank, and Director of Offer Strategy and Portfolio Management, Craig. You can ask questions throughout the webinar by using the chat function below, and we will address questions at the end. This is the second time that we do a webinar on the theme data center technologies, and we did one approximately two years ago, so we thought it was a good opportunity to recenter back to this theme. Today we will be talking about data center cooling market, our offering, and how we differentiate. With that, I will hand over to our first speaker, President and CEO of Munters, Klas Forsström. Thank you, Line, and to everyone watching on this right now, then, once again, very much welcome. Let me talk a few words about the history of data center technology. We can talk about history today and the future, or we can talk about stability, profitability, and growth, profitable growth. The first step, the history, that was very much about setting the operation, building ahead of the curve. Innovate, invest in capacity. It was about strengthening our market position, expand our footprint in Americas, and start to expand into EMEA through acquisitions. The second step, then, the current situation, that was about profitability and building a base for the future. Now we talk about expanding profitability, we talk about driving operational excellence, but we also talk about continuing to drive innovation into the market. The new SyCool that has a version both for air and liquid is one example. At the end, it's also about expanding ourselves into the customer base, number of customers, but also expanding it into our existing customer base and setting ourselves up for the next growth journey in Europe, the new factory in Cork, Ireland. Looking into the future, then, it is a little bit of repeat, I can say. It is about grow. It's about continued profitable growth. Expand our portfolio through innovation or through M&A. It is about reaching into other markets besides North America and Europe, perhaps stretching ourselves into South America and so on. But also very, very important, it is about continuing to invest in service and customer-facing functions moving on. If I simplify what we have done the last couple of years, I would like to describe it a little bit like this. It is about we communicated our strategic intent, and we started to deliver that step by step. It was about growing. It was about expanding the profitability. And I think we all are very, very proud and humble about the journey that we have done. Moving forward, then, I see a market that has an underlying growth of about 10%+ for the coming years. I see opportunities for us, as I said, to expand. It could be portfolios. It could be customers. If we talk about markets, I would love to have a stronger footprint in Europe, but I would also love to have a stronger portfolio across the globe. All this, I'm super optimistic for the future, but we lived according to our mantra. We take it step by step. With that, I hand it over to the real experts within DCT, Stefan, Frank, and Craig. So with that, over to you, Stefan. Fantastic, Klas. Thank you, and thanks a lot to the ones of you tuning in. I personally had the benefit to meet some of you in person, and so I happen to know there's a lot of interest to learn more about what we do in DCT and the data center market in general. So coming back to what Line said, why are we here today? What is the intent with this session? Well, it's really about focusing in on data centers as a segment and how we differentiate. So the best way to make sense of this vibrant and tech-heavy industry was to bring in the A team from DCT, Frank and Craig. Fantastic to have you here. So personally, I believe the best way to start, not least for the non-engineers on the call, I think it's really good to actually go back to the core of the customer problem. What is actually the reason, the whole purpose of why companies like us need to exist? Can we start there, Craig? Yeah, absolutely. Just to put it into perspective, all servers have microprocessors, and they're usually either CPUs or GPUs or a mixture of both. Each of those microprocessors has many millions and even billions of transistors. To process data, they have to switch on and off, and to do that, they use electricity. That electricity naturally turns to heat using the first rule of thermodynamics. As that heat builds, the chip's performance starts being affected in a negative way to the state that if it gets too hot, it'll fail. Therefore, that heat needs to be removed to keep those chips in their optimal operating environment. That's where we come in. That's what we do. Good. And to really kind of boil this down, I would like to start by showing four different steps of cooling and heat rejection and also talk a little bit about the difference between the two. Because as some of you know and most of you have seen, there's a big buzz about liquid cooling and data centers as a segment is glowing hot right now. And I do believe there's sometimes misconceptions of what is what. How big is air cooling? What is actually liquid cooling? There is actual liquid in many of the systems. So let's start from the beginning. Step one here, dissipation. That is inside the servers, right? So as you see from the animation, something needs to bring out the generated heat from the actual server, in this case, the chip, basically everything on the motherboard, but more in specific, the CPUs and the GPUs that you are talking about, Craig. So you can do that in a few different ways. You can do it through air and heat sinks. You can also do it by liquid. To us at Munters, we are pretty agnostic to that because we have solutions to couple with regardless of the solution that brings out the heat. That takes us to step number two. When you have taken away the heat from the server, you need to capture it inside the data hall, right? It's still inside the building. And that the animation here is intended to visualize. You can see that the heat is coming up from the server, going through a plenum in the ceiling, and back to what is called a CRAH, a computer room air handler. That's an equipment that we produce. Then you also need to transfer this heat somewhere, normally outside on the roof or to the ambient air or possibly to repurpose it, to reuse the heat. One way or another, you need to transfer the heat. Here you have an animation showing that third step. In the fourth and last step, you need to get rid of the heat somewhere. You need to dump it, basically out into the ambient air or repurposing it somewhere else. I think this is a good way to start discussions, at least according to my experience, because a lot of discussions or questions that we receive is about what we do. Coming back to these four steps, it's very clear that us at Munters, as well as our peers, we come from the HVAC side. We traditionally come from the heat rejection side starting at step two, basically. We are very good at capturing and making sure that the heat generated is taken care of. The liquid cooling companies that are inside the server, they are basically doing what the server fans used to do, right? That's a different part of the equation. We're going to talk about that more. If I turn to you, Frank, Head of Strategy, can you put this in context of what we actually do as DCT in this wider context of how to work with cooling and heat rejection? Yeah, thanks, Stefan. I think when I like to look at us, what makes us great at what we do, it's kind of our focus is creating the optimal kind of thermal management system for a data center, right? And we do that really through two things. The first thing is what we sell, our products. Munters, with data center cooling, started out traditionally as an expert or a niche excellence player in evaporative cooling. So we started out years ago by selling just components for direct evaporative media, and then we expanded that into using our Oasis system with our EPX heat exchanger to use indirect evaporative cooling as a full system. So that was like the first step we made. We've then expanded that product portfolio more. We now sell products that pair up with other companies, air-cooled chillers, to pull that heat out of the data hall. We also have now full split systems as well where we sell both the heat absorption and then the heat rejection component. Our best example of this is SyCool, which works very well with air cooling, with no water. It's industry-leading in terms of dry cooling solutions. We can also pair that up with liquid cooling, which, as you can see in the picture here, we have what's called our LCE or liquid cooling evaporator to pair with liquid-cooled servers. Then we have our service department to both install and kind of maintain and optimize that equipment through its life cycle. So that's what we sell. How we optimize is what we call our value proposition. We do that primarily. Our biggest strength as a company is having a really strong group of people who are very well trained in the application of thermal management of a data center. That allows us to take our core products and then optimize them or customize them for that specific customer's project. That optimization brings out the best resource efficiency we can have. It makes the products more sustainable, both from an energy use perspective and also in terms of the components that go into it. Basically, we try to get the most cooling we can out of the most compact solution, right? Then on a project-by-project basis, we deliver this expertise through one, our team of project managers. Every job we sell from DCT and Munters is assigned a project management team. And those project managers and service technicians both have really strong application knowledge. A lot of them come from our application engineering group and then move into that area. We keep that applications knowledge really throughout the life cycle of the project and the customer relationship. Just a quick question, Frank. I mean, in a very risk-averse environment, can you elaborate a little bit on why this is so important? Because I think that for the audience, this is one key aspect of how we differentiate, that we don't only sell the products, also striving for the thought leadership of this industry. Can you elaborate a little bit on that? Yeah. So we are not a product sales catalog company, right? We have a range of offerings, and then the customer lists their problems to us, and then our team is fully capable of optimizing that technology for them. And then because we have that project manager assigned to that job, they basically cradle that job from the day we get the order to the time it's handover and IST Testing is completed at a job site. All right. Let's continue talking about the market and addressable part of this as well. So, how we look at the data center cooling market or the data center market as a whole. The first thing we do is we kind of size what percentage of the total CapEx that gets announced. And we all hear them. I mean, there's massive projects that get announced every other day, right? So out of that total amount of money that someone's talking about spending, we estimate that roughly currently between 10%-15% of that CapEx is allocated to the cooling portion of the data center. When liquid cooling comes in, that's going to come with a bit of a premium because the systems are going to be a little more complex. You're now getting really close into the servers. So there's going to be more of that CapEx is going to be on cooling. We then look at, okay, how do we size the market and how it's growing? And our best way of doing this is to look at, because these projects are all pretty visible, there's a lot of resources out there that show in all the major data center markets what amount of capacity in terms of power consumption is being deployed. And then we can then kind of just calculate then what the heat rejection needs are going to be in terms of MW s. s. And you can see here we've listed out what we see as the MWs being deployed by region year-over-year. And we, like Klas mentioned, are going with an assumption of about a 10% growth over that time period. It's not going to be 10% exactly every year. Some years could be higher, some years could be lower, and especially in the different types of products that we see could grow a lot more and some could shrink as a result of that. I think it's fair to say, Frank, just as a comment on this, that when we look at all these reports, we do have to add a little bit of a manual filter to it based on our industry knowledge, because if we just kind of accumulate what everyone is planning for, that wouldn't be reasonable given all the permits for land, energy restrictions. And we're going to come to that. But yeah. No. So for sure. So that's why, on the right here, you can see we look at all of them and kind of come up with a sanity check on this. And that blend shows that all of the industry publications on this, on average, are predicting an 11% growth rate. So we're pretty in line with that amount. Good. Let's continue with customer structure. This is something we get a lot of questions on, Hyperscalers versus Colocation and within Colocation and related to technology and so on. So let's talk about this house. I like it. I like the picture. So first, to start out really quick, what are the three major segments within the data center market that we look at? There's enterprise, and we consider enterprise and telecoms together in one bucket. That's a market that, quite honestly, we don't play a lot in because our solutions aren't built for these really small data centers. Remember, an enterprise data center can be the server cabinet down the hall at Munters, right? It's a closet with a lock on it and one small air conditioner in it, right? We then also have Hyperscalers. So these are the cloud giants of the world. They're driving predominantly most of the activity in the market right now today that you hear about. Those Hyperscalers then and the enterprise companies use the Colocation market to deal with the amount of data they need to process. Simply put, I mean, these companies can't invest fast enough to build all their own facilities. So they use the Colocators to build up a lot of this infrastructure for them. We look at the Colo market kind of as two different parts. We have the build-to-suit part. This is the wholesale Colo market, which is basically when a Colo will build an entire data center campus or a building for one customer. We then have the retail Colo. So this is when Munters' IT department decides, you know what? I don't want to manage this myself anymore. There's experts who can do this. So I'm going to go to a company that has a set of racks that I can just go and put our servers in, and then our servers are running in a much more controlled environment. I like to look at this really as a lot of markets that certain companies are in there, if they're segmenting their customer base, their customers are competitors. That's not the case with us. I look at our customers as like an ecosystem. They all have to work together. Hyperscalers couldn't grow the way they want to if the Colocation market wasn't out there. Same thing with enterprise. It would be very hard for certain companies to really scale up with all the servers they need in one place rather than going to a retail Colo provider. So that's how we see the market today. Eventually, you can see the house here. We're going to build on, right? We're going to keep growing. The edge market will be very soon, in our mind, start coming into play, and we'll get into that a little bit more in a few slides. Yep. And I also think it's relevant to say that the Colocation and hyperscale link, as we try to visualize here in the picture, is very tight. So coming back to the question, can you serve both markets? Well, many times it's basically the same thing. We do deliver to a hyperscaler, and they might have the requests on the Colocation to have a similar infrastructure, right? Yeah. It's very important because the hyperscalers can influence what a Colo builds. You need to be very close to all the different segments in this market in order to serve it properly. Okay. Very quickly, to be careful of time, can we touch upon how we are positioned in terms of the competitive landscape? Yeah. So if we look at the competitors, we break the market in terms of our competition up into four different areas. And we kind of look at two criteria: how flexible are these types of our competitors with our customers, and how large of a product portfolio do they have? So if we start out with the biggest guys in the room, we call them the cooling giants. These are the major cooling companies in the world that many of us have HVAC systems in our house that are made by the thing. Yep. They cool and they heat everything. And so they've seen the data center market as a great place for them to sell their core products. And it is. So they have a very specific product type that they can offer, and they offer a lot of it, right? So if you want to buy something that's readily available from multiple vendors at a time, they prevent a very good opportunity from that. We then have what we call the vertical houses. These are also very large companies, and they sell the entire kit of mechanical equipment that a data center operator is going to need: electrical, mechanical, and cooling, right? So it's a one-stop shop. Again, these types of companies are going to be a bit more product catalog focused. They develop full sets of products, and they're pretty rigid with how they sell that, right? Then we'll move over to the other end, the more flexible vendors in this industry. First are the niche players, which we talked about. Munters comes from this. We did evaporative cooling, and that's what we did, and we focused on that, right? There's still a lot of niche players in this market. Then where we strive to be is what we call a full solution provider. Our aim is to have that product range for cooling like some of the giants or the vertical houses have, but also maintain some of that application know-how that actually is how we win in our markets. By combining that, that's where we aim to be. In our mind, we've gone over being a niche player. We have a more broad range, and we want to keep getting higher on that. Excellent summary. I agree. Okay. Good. So in a nutshell, I think to understand this industry, if we take it back on a kind of zoomed-out perspective, I think it's important to see the three most influencing aspects of the data center industry. The first one, obviously, is the digitalization as such. We talk about machine-to-machine. We talk about future mobility, 5G, 6G, and so on. We talk about low latency requirements, meaning that you will have the need to build smaller, more modularized, containerized data centers closer to urban environments to have low latency because you don't want long response times for autonomous vehicles or whatever. The second one is, of course, what we call the iPhone moment of AI. I don't think we don't have time to go into all of that, but I think everyone listening in knows that there's a big buzz about AI, not only in the data center industry. It's actually transforming the whole society, right? So it's everything from generative AI to applied AI in different places, right? So if you put these ones together, you can see the digitalization is like the core platform. And then on top of that, we're going to see an industry infused by AI. And then thirdly, you have the context of all this, which means the climate change and resource scarcity. So you have to kind of balance the growth game with doing it in a very sustainable and energy-efficient way. And there's a big debate of everything from water scarcity to the grid restrictions as well, where you're going to find power. So if I have to summarize it, I would say these are the three main blocks that everyone in this industry are talking about. So if I leave it back to you guys and maybe starting with you, Craig, looking at these ones, I summarize saying digitalization is going to drive changes into location and size. And then you have the AI that's going to be moving more into server densification and, I mean, also as a consequence of that, liquid cooling. And then you have the efficiency and reuse aspect related to not only the requirements, but also what comes with technology. Can we, to pick one of these, to go a little bit deeper on as a double click, can we elaborate a little bit on the heat densification and liquid cooling? Because that is, I guess, what many in the audience are wondering about. Yeah, I'm sure. I'm sure. So yeah, so obviously what's happening with AI is that the servers are utilizing an awful lot more power. So, I mean, if you want to use an example, one that makes it a bit simpler for everybody, a typical server at the moment is generally around 10-12 kW. And what you have is a typical car. If you're going to charge a car, it's using a similar amount of power. So if you can imagine that all of a sudden we are going to, as a minimum, probably triple that up to around 30 kW in a rack, it's the equivalent of having three cars plugged into your same plug. And just to increase that, those 30 kW, there will be racks that are even bigger. And we've even got some clients talking about 300 kW per rack. Now, that is extreme, but it's the equivalent of having 25 cars plugged into your plug. And it just gives you the scale of the growth in the amount of cooling that's going to be required and the amount of heat rejection that's going to be required in this marketplace, really. But besides the actual energy increase in this, why is this a problem for the industry? Why is this a challenge for air, as an example? Well, generally, the fact that air, if you think about the specific heat capacity of air, it's a certain amount. Liquid can carry 4,000 times more energy for the same amount. So one liter of liquid will carry 4,000 times more heat capacity than one liter of air, for example. So it's easier to transfer and to move. You talked about that transfer from the heat source outwards to the heat rejection. So that's why there's limitations. Once you go above maybe 20 kW per rack, you're really kind of going into the liquid-cooled era. So does that mean that everything will now move to liquid cooling? No, absolutely not. Because what you have is there's two things to look at. One is that with a lot of the liquid cooling solutions, you're still going to be requiring some form of air cooling for the actual data hall. The average figure is given about 30%. Now, if you think that the rack's going at least or on average three times, it means actually the amount of air cooling is staying the same, if not slightly increasing. But you've got that on top of the liquid cooling. So from the heat rejection point of view, we're using an awful lot more heat rejection. But the other thing to sort of bear in mind is to be a bit more realistic as well, is that this is a sort of study by Uptime. And this shows the highest server rack density that's deployed in over 600 sites. And what they've got here, as you can see here, the highest server density they've got in 87% of those locations is less than 30 kW. Now, that means probably that the average rack density will be somewhere about 10 times less than that. It'll probably be between 3 kW and 6 kW. So in general, we've got to be realistic about when people talk about these 100 kW racks and 300 kW racks. They are going to be extremely rare. Yeah. Now, I used to say that liquid cooling has been kind of a solution looking for a problem. Now the problem is actually there, which is fantastic. So I think we are talking about a shift, but it's not going to be as fast as the buzz is saying, but it's here to stay, and we have a good compatibility with it. And maybe that can lead us into the next picture, talking about how do we meet this future of hybrid solutions being able to cater to both liquid and air-cooled servers? Yeah. So let's talk about SyCool briefly. Sure. So, I mean, SyCool is obviously a split system we've talked about. You can see there the four stages that you mentioned on your earlier slide: the dissipation, the capture, the transfer, and the release. Now, where SyCool wins and why it's such a low-energy system is the fact we use refrigerant as a medium, which means that the two-phase nature of it—I don't want to go too technical, of course—but the two-phase nature of it means that we can do an awful lot more free cooling. We can get passive cooling all the way up to about eight degrees below the supply air temperature. And free cooling is basically, in this context, it's when you use the difference in temperatures between outdoor and the indoor. Yeah. It's driven by fairly basic physics. It's just the fact that the heat from the servers evaporates the refrigerant that rises up to the condensers. The condensers, the cold air condenses that refrigerant gas into liquid, and the liquid falls through gravity, and you get this thermosyphon. So if it's cold enough outdoor, you can cool through that? You can cool just using outdoor air. That's it. So there's no compressors for the vast majority of the year. And actually, for EMEA, the weather conditions really suit this. They really do. So, I mean, but one of the beauties of SyCool is actually you can see here we've got both the air-cooled and the liquid-cooled option. Now, a lot of designs nowadays are looking at hybrid because they don't really know yet what scale the AI is going to do or the HPC or the higher sort of end of things. So what we've got is that opportunity to do a hybrid system, which is shown here, or actually to flip and actually change from air-cooled to liquid-cooled just by removing the CRAH unit, the computer room air handling unit, and putting in an LCU or a CDU cooling distribution unit. So the flexibility of the system is what you can see in the indoor part here, right? You can have the liquid cooling evaporator or the air-cooled version. And in both those instances, that condenser is still getting rid of 500 kW. Fantastic. Whether it's air or liquid. All right. Let's move forward. To be careful of time, I think this is a little bit of a wrap-up. I turn to you, Frank. I mean, obviously, as any company in this industry, we are strategizing for a very fast-developing market. So in a nutshell, before going into Q&A, can you describe a little bit about what we see in the market? Yeah. So I think both what we see and then how we're addressing it, just to really quickly tie it all together. I mean, the slide says AI is generating a ton of demand and leading to a big growth in the industry. That's true. Cloud is also still generating a lot of demand, too. So overall, there's a lot of demand for new data centers to be built. And that's evident if you read any of the news out there, right? So what do our customers need to meet this demand? They want solutions that can deploy in mass scale. So the days of individual project designs for every data hall seem to be leaving, and then customers want solutions that deploy across their entire platform. And our systems now have this modular approach that does that, right? They're demanding systems that are future-proofed. So that means they can be compatible with air and liquid cooling. And Craig just showed a great example of how our SyCool system can do both. And there's going to be a continued focus on being more efficient with what we got. So that ability to optimize for our customers is going to be huge because every dollar we can save them from OpEx helps us grow with them. And those are really the four key underlying trends in this industry. And we've, I think, shown during this meeting how we're going to plan on addressing those. Very good. Thank you, Frank. And And Craig, we're going to continue. I think that is a good wrap-up to move on to the Q&A section of this broadcast. So let's take it from there. Line, I give this to you. Thank you. And thank you for a very good presentation. Very interesting. I'll squeeze in here with you, Frank. Good. So a lot of questions we have received already, and you can still ask questions by posting them in the chat function, and we will address them. So I think I will just shoot questions, and you will decide who feels they are ready to take the question. Do that. Okay. So can you elaborate on your position within liquid cooling? How do you collaborate with other players in the value chain, for example, NVIDIA? Good one. I think here everyone wants to speak. But in fairness, Frank, I mean, you are leading the strategy work. We do have a lot of collaborations out there, I have to say. But maybe you start and we fill in. Yeah. And obviously, everyone wants to talk about NVIDIA. They're the coolest company out there right now, right? What we do for these types of partnerships is when we talked earlier about that heat dissipation, those are the types of companies we're working with now to see how our solutions pair up with them. We've had a lot of good collaboration, which we've actually talked to customers about. We can't tell everybody we're meeting with, obviously, right? But we are meeting with those companies, with our customers, and with certain chip suppliers altogether to really figure out the right way to manage the servers that are going to demand this liquid cooling. Yep. Now, in short, we are very active. Again, I think you mentioned it before. This is an industry where people are really living up to the expression "smart companies collaborate." So I think it's a common quest to understand what good looks like in new technology. I find it pretty easy to find partners that want to explore this opportunity. We have something to bring to the table through our technology and as thought leaders. I think it's a good position to be in. Perfect. Another question. Are your DCT solutions ready for the change in refrigerants, low GWP? Yes, I'll take that one. Yes. No, definitely. We do keep a close eye on legislation and regulations. Also, we need to keep an eye on local legislation and regulations because you can go within one country, and there's different demands for different municipalities, for example. Of course, with regards to things like F-gas, which the EU has had since 2014, the drive is to get down to much lower GWP refrigerants. So we've designed the SyCool system to use either R123ze, which has got a GWP of one, or R515B, which is an alternative, but it's a non-flammable alternative, which has got a GWP of 299. So the refrigerants we're using in our designs for Europe meet F-gas regulations. Even for the states, there's changes going on in the states that they have to use lower GWP. So actually, we've kind of almost helped in that development for the US market. Yeah. Sorry. So in simple words, I used to say we are future-proofed. Yes. And it's really a final comment. I think it's in our DNA as Munters as well, not only in this area. But yes, we live up to legislation and regulations, but this is also how we differentiate, right? So to push this envelope and try to be ahead of the curve is really how we try to differentiate as well. So I think it's important to say we are not sitting passively and looking at, "Oh, I mean, we have to adapt to something." It's rather like, "How can we use this to a competitive advantage?" Yep. Understand. Good. So modularization has been an important topic within Munters for the last couple of years. Klas, do you want to elaborate a bit about how you see data centers have worked with this? Let me start, and then I gladly hand over to the true experts. But what I see when it comes to modularization, I think we have a beautiful example in SyCool. I mean, just by looking on it, you see it's built up by different modules. You have the roof section, and then you have the indoor section. And that can, of course, be expanded into different modules. So from my perspective, that is a very, very clear way how we modularize and to some extent also standardize. But please, guys, if you could add any flavor to this. I fully agree, Klas. And then I think it's important sometimes to zoom out a little bit and wonder, why is modularity good? Why do we do it? Well, of course, we want to reduce unnecessary variants. We want to reduce cost that we don't see add value to the customer or to us. And this is basically the concept of lean in a big if you zoom out. And that Munters have excelled in, I would say, for years. So I feel very confident that we are chasing efficiency improvements in the operations every month. We can continue to do even more. So there's one part of that. And then we are also in an environment where customization is very important. But you can do that in a smart way. But some products or solutions by nature require to be built into actual buildings. So they are a little bit snowflakes by nature, but we can do it by working smart back end. Yeah. Good. Maybe we can talk a bit about the market, the European market versus the U.S. market. Can you say anything about the differences here regarding technology or customers? Yeah. Well, actually, what's the differences between the U.S. and the EMEA market? I mean, in terms of customers, there's not too much difference. I mean, a lot of the influence that's coming into Europe is done through American customers. One of the reasons the European market's actually growing, I think, in relative terms, a little bit faster than the U.S. is because of digital sovereignty that's come in with GDPR and the rules that we have in the EU to keep the data in the country where it's generated. So that's led to a lot of the American companies sort of moving over into Europe. And of course, it's also led to the expansion of the European-based data center companies as well. So in terms of many differences, to be honest, legislation is one. Obviously, as I said, we talked about F-gas being a driver. We think that there are other pieces of legislation as well. Energy availability is probably more of a challenge in Europe, albeit it's a challenge everywhere. It's a challenge for the world. That's one of the barriers that Stefan mentioned earlier. The weather is definitely more favorable in Europe, particularly in the Nordics and the Scandinavian area. But also, the route to market's a little bit different. So there's certainly more of the early growth in data center in Europe was quite often led by design and build contractors, for example. So there's a lot more power in the contractors than there is maybe in the Americas market. Okay. Good. And maybe two other things to add there, Craig. I think because of the power availability in Europe, we see still a little more demand for evaporative solutions in Europe because those are higher in electrical efficiency. And then in service, we see more of a demand for installation work, whereas in the Americas, that's more commonly done by mechanical contractors outside of our scope. Yeah. Yeah. Maybe then, as you mentioned service, we can take a question regarding service. Can you talk a bit about the service offering and how you anticipate this to grow going forward? You mentioned the installation systems. Do we do this ourselves, or is it something? Yeah. In Europe, we do do more installation work ourselves. In the U.S., it's not as common. We oversee the installation and the startup of our equipment because we're the experts at it. But the actual physical labor is done by typically a mechanical contractor. Our service business is a little different than other parts of Munters. A lot of the routine maintenance for our type of equipment, that's not as big of a value we can add. It's not like a desiccant rotor where you've got to be an expert in a desiccant rotor to really understand how to maintain that. Data center cooling equipment typically consists of compressors, coils, and fans, right? Because we have sometimes hundreds of these systems on a roof, it's just more economical for our customers to do that type of work themselves. Where we're going to start focusing our efforts within service is back to how we win, our application knowledge. Maintaining that optimization, maintaining that energy efficiency is super critical. That's where we can add value. We're looking at ways to expand those types of offerings within our service and being able to provide those either on-site or digitally somehow. Yeah. Interesting. When you mention compressors and coils and fans, I guess the translation really here is that those are low-added value. It's simple procedures. It's not so complex, where we, of course, try to take the service to areas where we add more value. That can be in the more digital aspect. It's definitely to your point in commissioning and startup and things like that. Yep. Good. So a bit on the competition. How do the competitive dynamics differ between SyCool and competitors compared to Oasis and competitors? And also, could competitors copy SyCool? We can start from the back. They can't copy it since it's a proprietary solution, right? Comparing to the customer's second-best alternative, I think that's a question for you, Craig. I mean, you can put this in relationship to maybe air-cooled chillers and CRAHs versus SyCool. Yeah. Definitely. In terms of sort of the biggest part of the market is air-cooled chillers and CRAHs. And SyCool is a direct competitor of that. That's kind of where we thought, particularly for EMEA, where there are pumped refrigerant systems out there which are more common in the U.S. than they are in Europe. But for tackling or the biggest part of the market to tackle was the chiller and CRAH market. Now, the whole advantage, as we talked about the free cooling earlier, is the fact that SyCool will generally use about less than half the energy that a typical chiller CRAH system will use in Europe. So it's a huge benefit. And the fact that there's legislative challenges, there's regulations coming in where they're saying that you can't have a PUE for the whole data center of more than 1.2 for a new build or 1.3 for legacy. The only way they're going to have to do it is by utilizing the most efficient systems that are out there. And it's the right way to go. It's the right direction to move as an industry. So super well-positioned, having, if not the most energy-efficient system, at least up there? In terms of a split system, it is the most energy-efficient system. I love that. They can go evaporative cooling to find even higher levels, which we excel in. That's where we started as a company. Yeah. Good. So new question. Does Munters benefit from a transition from regular data centers towards liquid cooling data centers? The increased share of wallet should not affect your business as Munters is not a supplier of liquid cooling-specific equipment. I think we can keep this one pretty short because I don't think we are by adding liquid cooling. Remember these four stages, the dissipation part. We have never done that. I mean, airflow, yes, but we have never done the small fans on the server, right? So by having something else going in there and fetch the heat now through liquid doesn't really impact us. The key question here is, are we compatible to cater to both those needs? And for that, the question is yes. Yep. Perfect. Broadening your product portfolio, can we elaborate a bit on that? We can. This is so fun. So we all want to talk. But I think we can. But I can start very briefly. I mean, remember Frank's story, right? We started in evaporative cooling as a niche player. And then we have broadened the portfolio now, having our own full solutions. But there are still parts of the market where we only have half of it, sort of, like in the chilled water infrastructure, data centers using chilled water as their main infrastructure. Here, we do the indoor unit, the so-called CRAH, the computer room air handler. But we don't work with air-cooled chillers that Craig mentioned. Maybe you want to elaborate on that. That, I would say, would be, of course, one area where we are not currently in. That's a big part of the market. That's where we compete with SyCool. Good. Yeah. I suppose liquid cooling offers some other opportunities as well. So you have that instead of the CRAH unit, you would have a CDU unit, whether if it's linking out to a chiller. But as we say, we've got the options there with SyCool because we can use the liquid-cooled evaporator, which is a refrigerant-to-liquid CDU. Please, Frank. Go ahead. With the temperatures with liquid cooling too, there's going to be probably an increase in the amount of dry coolers or adiabatic-assisted dry coolers. That's another part of the market we're not in. These are areas that would be open for expansion. Just to summarize it, I mean, I know that one of your main objectives, that is, to widen the product assortment and to widen the markets that we serve. And if we then can do that by organic or if it is by M&A, I mean, in my book, it doesn't matter really. I mean, we will expand. Yes. So we evaluate the product portfolio often. And we want our portfolio to fit in with those how-do-we-succeed criteria. So if we have a product that has a good market and we can succeed in, we'll look at expanding into it. One quick add-on comment about liquid cooling. This is really important. I see that a lot of people externally, they see like, are you a liquid cooling company or not? I think it's really important to underline that basically, I stick my chin out here and say none of the HVAC players are liquid cooling companies. We all come from a different angle. And now we step into that part through acquisitions or collaborations, right? We have chosen to stay agnostic, to collaborate with the step one in the picture I've shown. But it's not like some of the competition has been in this forever. It's been small niche companies, often university spinoffs, smaller companies that now start to scale together with companies like us. I think that's important to remember. Absolutely. Good point. So you talked about a market growth of 10%. How can you connect this to Munters DCT growth? If I start there, then perhaps I mean, I used to put myself out on the plank. I used to say to those guys then. But you made us come with you then. Yeah. But really, I mean, we are very confident that the underlying market, as you can see by statistics and so on, is growing 10%+. Then we are not saying that, I mean, we will grow this and that. But I feel extremely confident that the market is there, and we are gaining market share. So with that, I mean, that is supportive, accretive to our long-term growth targets within Munters. Perfect. I'm speeding up here because we have a few questions left, and we're running out of time. So I will try to catch one more question here. It would be interesting to hear your view on the current and future power situation. Given the massive demand outlook, the data center build-out, the increased power density, etc., do you expect the industry to be able to meet the demand? Could regulations limit the build of data centers? I think just yeah. No. I mean, what you've seen if we just take EMEA as a market, what you're definitely seeing is a lot more investment going into the sort of tier two markets. So they've already realized where there's struggles for getting power. So what you see is a rise in Milan, Madrid, and Warsaw, and Vienna. And Berlin in particular, there's a lot of growth in those areas. So the data centers are kind of moving to where the power is a little bit. And actually, with AI, as I say, with most AI, they could get away with low latency, so they don't need to be right there. So you're starting to see locations being moved. And I think we mentioned it earlier. They're either very big data centers like 300-MW campuses, for example, or they're getting small as well, where they're moving towards the edge. Those areas, the interesting part is also the heat reuse aspect of it that we've also got to talk about. Good. One last question. The separately announced large orders in 2022, 2023, would you be able to split the mix between complete new builds and retrofits of existing centers? And how many data center projects could one expect, for example, in North America until 2030? Yeah. I mean, if we start with the simple part, correct me if I'm wrong, Frank, here, but I think that the vast majority, if not all, of those were, of course, to new builds. So none of it was retrofitted. The announced ones. And our equipment typically is going into new builds. So 97% or 98% is going to be Greenfield. So that's the easy part. And the second part of the question was what we see in the pipeline of the general market to be built. Well, according to building on what Klas said, we see a good growth in general. The exact amount, of course, is really hard. It's only guesswork. The problem here, not excusing myself, but even when we look at general market institutes and the ones doing this service, there's a very big variety of information. But to summarize it, we do see that the market is growing in a healthy way. We're going to see it's going to be very diversified from areas being flattish, even maybe shrinking, to the ones being glowing hot and growing a lot. But overall, for data centers, yes, there will be more data centers built, probably more in planning right now than what is already deployed. That says something. But then you need to find the land, the energy, the permits, and all that. That's going to be the true challenge of the industry. Challenge. Understand. Good. Thank you very much for all the good questions. We have a few questions that we have not been able to handle, but we will reach out to you separately. This webinar, together with all our other webinars, are available on our Investor Relations webpage, where you can also find other presentations. The next two events that I would like to mention are our Q1 report on the 23rd of April and also our Capital Markets Day on the 21st of May here in Stockholm, where we hope to see a lot of you. If you are not able to come here to Stockholm, we will also be hosting it digitally on our webpage. A very big thank you to all of you guys for presenting today. Thank you. Thank you for all the questions received. Also, thank you for listening in, of course. Bye-bye.