Wärtsilä Oyj Abp (WRT1V) Earnings Call Transcript & Summary

Hi, all, and welcome to this theme call, which is focusing on Wartsila's data center opportunities. We will start in my presentation, which will be given by our CEO, Hakan Agnevall, and President of our Energy business, Anders Lindberg. The presentation has been published on our IR website. After the presentation, we have good time for Q&A. Let's start with 1 question with the analyst and then continue with the follow-up questions. I would like to remind you that the purpose of this event is to focus on long-term business opportunities. And as a reminder, we will host pre-silent call on March 23, together with our CFO, RM Barents. So let's leave the questions related to detailed financials to that call. Hakan and Anders, Time to start, please.

Thank you, Hanna-Maria and a warm welcome, everybody. So today, we have me here Hakan and Anders. Anders -- he is President of our Energy business, so to say. And I will say click when we ship the pages. So if we can go back to where we started. I think the whole topic of today is, of course, our supply for data centers. And you could say that last year, we had a breakthrough in the U.S. data center market. We were actually in the European market before, but last year, we had a breakthrough in the U.S. And we do see a significant opportunity for growth. both in new bill and later on in services, associated services have an opportunity right now the biggest opportunities in the U.S. We do see movements in Europe and Asia will soon become as well. So -- and we also have received a lot of questions in the past about our data center business and how we think about it. So we wanted to take this chance to spend some more time. You will also see that we will provide some trends in graphs. I want to make a general disclaimer because this market is growing so rapidly. And it's also -- there is a high -- it's so dynamic that customers, they don't want to talk with their names, et cetera, et cetera. So the full visibility of growth numbers, et cetera, it's not there. But we have been trying at least to give the trends and outlines here going forward. So if we look on today's agenda, we talk a little bit about the market development, and I will start there. Then Anders will take us through more on the technology side. Why in where our engine is the right choice going forward and why have we had a breakthrough in the U.S.? And then also, where are we competitive and why are we competitive? And then we will open up for the Q&A. So I would envision our presentation will take maybe 45 minutes, something like that. And then we will have about 45 minutes for Q&A. So that's the time frame. And we will go through the presentation, and we take the questions after the presentation. Okay. Click Nick. So basically, I mean we are a player in a rapidly evolving market. And we have entered this market for primary power production. And it's quite a recent development. I mean, 2 years ago, you could say that we -- as we actually -- we were not in data centers. And back in those days, the data centers were more focused on traditional storage applications. Storage are mobile photos, et cetera, et cetera. And in those stages, the data centers, they were needing tens of megawatts and our engines, and they normally hooked up to a grid and they had some backup engines and our engines were not the right solution for that type of application. So you could say, roughly speaking, 2 years ago, we were not into data centers. Now what has happened? We all know it. AI is developing and as AI require a lot of compute, they also require much more energy. And now we are talking of hundreds of megawatts of power meeting, sometimes gigawatts. And it's a challenge to get great access and can somebody term is off because it really is going to be the same thing. Sorry, we have some pop-up here coming all the time. So basically, a lot goes off grid because it's hard to get utility access. And this is also where we are coming in because the type of power of megawatts, sometimes gigawatts of power that we are talking about, are very well suited for our technical strengths and also shorter delivery times than some of the competitive technology. Our solutions, they have some intrinsic benefits. They are energy efficient and modular and we compared to competing technologies, and we will talk more about that. They don't rate in hot climates, the world is getting warmer and they require basically 0 water and water access to water is a challenge in many areas in the U.S. and other parts of the world. Also, I mean, we also see on the time in connection to the balancing narrative that for those of you who have followed us, we have been talking about for quite some time. But engines have this superior capability to operate in tandem with renewals, providing the balancing power of a robust power supply. And this also in combination, this property in combination with our development and sustainable fuel side, it supports the data center owners or even the hyperscalers long-term ambitions for decarbonization, so to say. And then last year, we captured orders to total close to 800 megawatts to orders there. And now we started this year very strong with another 429 megawatts. Now that was a utility -- the 21st ones, they were, we could say, data center developments. This was the utility order, but it's a utility with this plant having a dedicated, I mean focused purpose for data centers. So this is the context, click. Now -- we talk about the data center market where we are. And now basically, a big chunk of the market is shifting into our sweet spot. And it's certainly achieve, not all, but there is a strong shift into stand-alone baseload power driven by -- particularly in the U.S., but we see similar development in other parts of the world by the long lead time for grid connections. And if you look at historically, and I mentioned that before when we were talking about data centers, mainly focused on data storage, you were talking about power needs of 10, 20, even to 100-megawatt and the typical power supply was great connected and then you had some high-speed engines for backups. And customer focus was, of course, on power rate, but also CapEx. Now when data centers are growing in size, accelerated by AI and the need of recruit -- the need for additional compute power, the energy consumption and for energy goes up and read interconnection in particularly in the U.S., it's -- for many instances, it's tough, and it takes time. I mean, 5, 7 years, 8 years. And this is where the off-grid solutions grow in importance. And our sweet spot is -- and this is loosely defined, it's between 50 and 400 megawatts. But I would like to highlight our underlying it's lutein because as you've seen, we have taken orders also above 400 megawatts, and we can certainly do that in the future. There is no kind of electrical or mechanical limitation. It's basically a question about how much land you have because certainly, if you go to gigawatts, you will go with combined cycle gas turbines because the energy density for the CCGTs is better. But -- so this sweet spot you should take -- it's not as distinct as we put it here. On the lower end of the spectrum, when we talked about that, our engines are probably too big. They are really -- they are higher than CapEx. They're really good in OpEx, but some of the customers, they will go for high expedience or other solutions set. So above 400, this is certainly loosely defined. But in the high -- certainly on the gigawatt side, you would see a heavy duty gas to in the OCGTs and sensibilities. So this is the landscape that we are moving in. But as I said, we see a lot of need demand for grid power coming into this 5,400 loosely defined sweet spot lots. Click -- now this is the ecosystem, and this is also to clarify how are we interacting because we do work with our, you could say, traditional customer segments, but also with new one. And when we look at the ecosystem, we have the operators and end users to the far right. So there you have the hyperscale and colocation data centers. And we don't deal directly with them because in between, we have the developers and the utilities. And we have that for decades with the utilities and IPPs, but now we also start to deal with the data center focused developer. So this segment, the developers and utilities, they engineer the data centers, the power and they build and often own it and they buy then the equipment, they buy equipment from us. They, of course, buy equipment from other suppliers, and they build these data centers, including the power. So our customers are mainly the developer -- developers and the utilities. And they contract the developers and utilities, they contract with the hyperscaler. Click. Now there is another element also that I think is really, really important because sometimes we get this question, okay. I mean the power line is coming in 8 years' time, then there is no market. And we don't see that at work. I mean our data solution meets the customer demand for quick access, but they are also very well suited for flexible power generation of the future. So now you can say basically is that now everybody is running for power and to be able to grow the data center need and the compute need. And I think what -- I mean, what we see happening a couple of years down the line, 6, 7, 8, depending there in many cases or in some cases, there is a power line coming in. And what our customers are saying then because of the modularity of the solution, they might add a couple of modules, and then they start to cell grid -- sell power to the green as well. So they will still sell the data center, but they will also start exporting to the grid. And the key to manage that type of operation is flexibility. And -- for those of you who follow us, where we really stand now, this is flexibility. And then if it's that 3 chronologically, but we also know the hyperscalers, they want a green power. And I think increasingly, they will be looking at bringing in power for wind and solar and wind and solar are great, but they are intermittent and this is where it becomes more of a balancing solution. And as you know, we are working a lot on the balancing power segment already today. So in all this kind of development aspect, our technology is a very interesting technology, and it has some competitive advantages in its flexibility. Click. Now looking out over the globe. Right now, there's a lot of things happening in the U.S. It's developing rapidly and key customer segments of the data center developers and IPPs, but we certainly see the utilities as well as we've seen here in the beginning of the year. I mean the targeted applications include off-grid and behind the meter data centers. And we had our first 2 orders last year, close to 800 megawatts, and we started this year strong with another order of 430. On the Europe side, we actually started in Europe as well as in data centers. And that was the partnership that we still have with ABK,and we have had 3 orders in Ireland. And then those projects are still ongoing. But I would say the development in Europe is slower currently than it is in the U.S., clearly. But it will come over time. And we provide the equipment and maintenance support. This is for us, both -- I mean, all over the world, it's primarily an EQ business and equipment business. And we see opportunities in Spain, Germany and the U.K. So there are opportunities. But it takes longer time in the U.S. Then Middle East and Asia will also come, I think now the interest in data centers is clearly there. But then, of course, how much power can you take from the grid, how much power will you take from the grid when people land in that, how much additional power do you need and then that will likely translate into business opportunities for us. So this is also coming slower than in the U.S., but it might accelerate quicker than Europe. Let's see. Click. Now the big question is, of course, that is on many people's mind. What's the size? What's the potential volume and gigawatt size of the market. And I think the short answer, I don't think anybody knows. And there is a lot of small people that have made estimates and try to look into the crystal ball and you can see some of the organizations, we have brought forward the numbers of FP, Goldman Sachs, McKinsey, EA, et cetera. And I think the conclusion here, if you look at the prediction, this is U.S. figures, but the spread, which is the key message would be the same for other geographies. If you look at the spread, it's everything from 40 gigawatts to 110 gigawatt. So what's the right number? I don't think anybody knows. It will grow. That's for sure. And when we talk to our customers, I mean, basically, if I simplify the message, they say, okay, this is not about ChatGPT is exciting. But this -- when we see the real growth more over time, it's on corporate AI and what they mean with that is that companies like that or any industrial company, other banks, we start to use AI to drive our internal efficiency, but also to develop our customer offering. And as a, I can certainly relate to that. I see a lot of great things going in -- going on in Essel when it comes to how we run our business and also how we develop our customer base. So these are the drivers. These are the predicted development. But as we all see that there is that's quite a span. Then also on top of that, of course, there is a funnel process. If there is a need of, let's say, just taking a figure of 113 gigawatts at the top of this range, that's the total DC power, then how much will proceed, how much will be gas powered because not everything will be gas powered. I think short term absolutely. I think over time, gas power will have a significant share. But it will not be alone, clearly. And then how much will go into our sweet spot, loosely defined and how much is off-grid or behind the meter. So there is clearly a funnel process. before it reaches, you could say the addressable market, our sweet spot where we are competitive, we can grow. But still, all even considering this funnel process, I mean our conclusion, there are significant growth opportunities going forward. Click. So we have a growing pipeline of data centers -- sorry, if we have a growing pipeline, both data center by a growing pipeline of data center opportunities with attractive life cycle margins. And we do see high activity in the off-grid data center segments with a continuously increasing pipeline. I mean we will not go into the gigawatts, et cetera. But I can say that our pipeline of everything we see has grown with about 50% in the last 6 months. So there's clearly a lot of things happening. Now in that pipeline, the new things coming out, but there are also things going out. So it's a very dynamic development, but it's certainly growing. Data center customers a highly valued speed to power. And clearly, the market is short on equipment supply. Now if you look at -- if you consider our pipeline and how we -- the portion we would translate into business, we expect the revenue recognition to -- connected to the deliveries, it will pick up gradually here over 2026. So it's doing in '26 ramping up towards the end of the 2026. And then the business -- the service business we can, but there is quite a lot of delay we see it's picking up around 2030. And that is, of course, you need to install there is a warranty time. You need to run the equipment for time. that will be a very interesting service business. But it takes time for that to translate into revenues and bottom line. Click. So having said that, that's a kind of overview of the market, very exciting market. I think we are very well positioned, but it all starts with having a competitive technology and great people. But Anders?

Thanks. So let's go to the next page. Here, you can see some of the attributes where we are very strong with our engine technology. First of all, the full load efficiency, we have a good efficiency at full load. But engines also have the advantage that they also are good in Parkland efficiency. And we can also do unlimited starts and stops. -- which is also beneficial as not all engines on the running a full load positive time as load goes up and down. We also have a modular design, and I will show later on how we benefit from the modular design that we don't talk to big building blocks. The heat tolerance, which is, of course, only applicable where it's a very warm climate but we have very little derating with the temperature. So up to 40 degrees senses, we have very, very little derated if we compare with competing technologies and also altitude tolerance, if that should be an issue. So at a higher altitude, we also derate much less than competing technologies. Of course, the efficiency also makes sure that we have low CO2 emissions. And we know since we had a winter storm in Texas some years ago, that we can also deal with lower gas pressure if the gas pressure should drop due to iStone or other things in the environment. And also minimum water usage, which is, of course, very important where there is a dry climate, and there is a lack of water. But even I met with the customer yesterday, I pointed out that actually it's also important where you have water because of agriculture, interest and other interesting communities is not to use that water. So actually, water is important, not only in dry climate. Go to next page. So here, we have looked how we are comparing with the medium-speed engine technology to the high-speed engines, but also the aeroderivative gas turbines and combined cycle gas turbines. So starting from the left, we look at the efficiency. And there, we have put in some rounded numbers of what the efficiency are for the different technologies. So you can see that, of course, the combined cycle gas turbine has the highest efficiency, but the medium-speed engines are following closely at 50%. Then you have the ramp-up to full load where you can clearly see that the engine technologies, both the high speed and the medium speeds are doing better than the turbines and certainly than the combined cycle gas turbines. How many starts and stop you can do, clearly advantage for engine technology, modular design due to the size of the engine, clearly advantageous to engine technology and turban, key tolerance, also engine has an advantage attitude engine as an advantage. CO2 emissions there, of course, the most highest efficiency have the left, which means the combined cycle, but medium-speed engines are following and high-speed are derivatives are the ones with the highest CO2 emissions. And then you have the gas pressure and the water usage. And the reason, of course, why the water usage is very low on engines is that we have closed loop system. So to the next page. Here, we are trying to take in -- if we look in the sweet spot of 550 to 400 megawatts, the trade-offs and we selected a efficiency, modularity and flexibility and compare the different technologies to the left. And then we have summarized that in the background to the right, where we can see then that the media speed engine technology is actually ranking the highest if you to take these 4 attributes together. Of course, I did mention also other attributes, and it depends, as I say, and decide what are the most important. But to point out, these are the 4 that we see as the most common attributes that customer valuation. But as I mentioned before, water consumption heat tolerance could also play a big role, but they are not included in the diagram, but we should not forget that there are also other things that can determine what technology that is best to use in that specific site. We turn to the next slide and looking at the competitiveness of our technology versus some other technologies here, -- so I move to the next slide. So here, we have taken an example and said that, okay, let's look at a case where we have 300 megawatts off-grid data center in Texas. And then we have looked up with CCGTs, arrows and medium-speed engines, what -- how to come to 300 megawatts. So first of all, with the modularity, you have to see how can you scale this up to it to 300. And clearly, if you go with Cities and each CCGT have 72.8 megawatt then you need 5 of them to be above 300, so to say. And then you need the same thing you do with arrows that are 33.2%, then you need 11 units and then you come up to EUR 365 million. But what is then also important is that with the high 99.9% availability that you require from primary power source you then need to consider also the maintenance reserve. And if we then add that on top of the basic requirements, so to say, then you end up with the engines based on our V34 engine at 358 megawatts installed while the are you have to go to 465 and the CCGTs to 510, which actually means that you have 210-megawatt extra capacity, so to say, to all the time be able to deliver that time and availability requirements that the customers have. And with Eros, it's 165 megawatts on top and for the engine, it's only 57-megawatt on top, which, of course, have an impact on the CapEx as we will see going to the next slide.

And Anders, also, if I may make a comment if you go back, I think it's also very important to highlight. This is, you could say, in Akita case and we try to just to clarify the value of molarity, but there is also a note. I think what we will see, especially on the big power plants, you will use a blend of different technologies. So there will be some CCGTs. There will be some engines to get the right equation here. Because to just go with CCGTs will be extremely costly.

Yes. And especially if you go up to 1 gigawatt scale. Yes, we go to next? Here, we have now tried to convert this into what it means in terms of CapEx, and we have looked at the fixed cost CapEx -- and then we have also looked at the OpEx as well. So we have both in here. And of course, as I mentioned on the previous slide, the CapEx is, of course, affected on how much installed power you have and then we have run basically a case here where we assumed USD 4.3 per MB fuel price. And then, of course, the CCGT would have 30% higher LCE than the engine power plant. Of course, if you double that fuel price, the difference will decrease, but there will still be a 16% higher L3 than the engine power plant for the CCGT. And then we have the arrows, of course, that have even higher OpEx. And of course, that gets even worse if you double the fuel price. So this is what you can see in Europe per megawatt hour if you consider both the installation cost with the CapEx and also the OpEx, both in terms of maintenance and the fuel. Go to the next page. We mentioned water consumption and just to highlight that. Of course, it depends on CCGT or ROS. But as you can see here, since we have a closed-loop coating system, we basically require a very, very little quarter consumption, so basically as to top up. And you can see that CCGTs have more than 2,000 higher consumption and the 500 for the arrows. So it's a fundamental difference. And I think that the more the community around the data centers, do care about water and specialist there are agriculture, things going on. It is important, and it's not only in dry places where you have a lack of water is also if you have the need and use of water for other things. So -- very important. And actually, the customer met yesterday said that actually, that's 1 of the most important besides the lead time for them at the moment when they look at how they choose technology. Go to the next slide. I mentioned now lead time being still of course, very important, and you know that. And here, we have indicated what are the normal typical times to plant commissioning under normal circumstances. And then we have also added since this varies. And right now, we can see, of course, that due to capacity constraints, we see very much longer time. And here, we have the difference of going from 30 months, 2.5 years for engine technology, up all the way to 4 months if you look for CCGTs and ROS, I think are somewhere in between. But this is, of course, varying depending on the capacity, cancellations and new orders. So this can bear in time, but this is to show what is the normal time when there is no limitations on capacity versus the current situation that, that is today. Go to the next page. The other thing that I think also is we benefit from in all with our offering, is that we have a very strong service offering. We have built out service network. We have very good references of doing service, and we can also offer all types of services, meaning that we can start with just having parts agreements to optimize maintenance agreements, to full outcome-based agreements and even doing the O&M, and we have experience of doing all of this, and we have a network both in the U.S., but also in the rest of the world that is well built out. So we have references and customers feel confident that we can also deliver on our promises. And since especially the data center customers, for them, it is very important with availability and reliability, but of course, also the security. I think we have a good position with our references and also offering on the service side. And of course, we know that data centers are going to run many hours and then is also a good opportunity for us to offer these services to our customers. Next, coming back to what exactly we have offered in the U.S. So we had 2 orders last year in total of 789 megawatts. It was 42 what we call the 50SG engines, which is a workhorse that is well established in the U.S. and that today can also run on sustainable fuels, hydrogen, for example, up to 25%, but could also do more in the future. And these orders were booked in Q2 and Q4 last year and will be delivered -- starting deliveries in the first order later this year and continuing into 2027. And these orders were followed by, as Hakan mentioned, a utility order, but that was dedicated to data centers from the utility. So we want to mention that as well, and that was for 24 -- 50 SG engines with a total of 430 megawatts in the US. Thank you .

And then I would like to come back to the page that we have also highlighted in our interim report, our recent interim report. And that is with this very positive development, specifically in energy, but also to certain extent in Marine, let's focus on alt today. Now it will take longer time than it sometime back to convert the existing order book to sales. I mean -- and they are basically to kind of mechanists that are playing out here. And we are taking all us further out in the future, clearly. And then you know for those who has followed us, we do more -- much more of equipment deliveries than EPC, I mean, when we do the civil and installation as well. So in the data center of all us, given that they're in the U.S., this is clearly an equipment or EEQ and here, we have a revenue recognition model, which is basically based on milestones. And then normally several milestones, but you can say 1 major milestone is always when you deliver the equipment. And then -- so then you recognize revenues and you also recognize profit. And this is different from an EPC model, and it's a conventional contracting model there. You recognize revenues and profit continues over the project time, all the way from day 1, all the way to the end. So of course, you understand that these different type of models may result in different prioritization. And I would say we will recognize in general in sales later than before. And then we also indicated how quickly in the U.S. data center limits will ramp up. So we start now we ramp up towards end of 2026, but the service business will pick up later. Click. And this is nothing new for those of you who follow us, but we are very happy and proud that we just recently announced our capacity expansion of the industrial system related to our sustainable technology have in Vasind Finland. And we are investing both in the facility that we have, but also in our global supply chain because they need to go internally. We're going to ramp up our capacity. We are clearly working in partnership with our supply base as well. And we will be investing EUR 140 million. It's in addition to the EUR 50 million that was announced in April 2025, so EUR 50 million plus EUR 140 million. And we also said we're going to expand our technical capacity with 35%. So we use the technical capacity as the reference point. And we will expand that until it should come in operation in Q1 of 2028. So there is a certain lead time to get this in place. And also, just to make -- help you to assess this we've also been talking about that in 2025. So last year, we were running at 75% of technical capacity. So if you start from 0.75% and you go to 1.35%, that's actually an expansion with 80%, so I guess it's -- both are correct. It's a little bit how you want to talk about it. But we normally -- we are engineers and we talk about technical capacity and how we work on. So very much in progress. And this is, of course, also part of how we -- it will make a contribution to how we develop our order intake going forward. Click. So -- to sum up a little bit, we are well positioned to grow across the multivelop customer segments in both balancing and baseload. So let's not forget our overall narrative because it still holds. It's not all about data centers. On the balancing side, we do see it playing out a lot in the U.S. as well and it's based on renewables, often being a very affordable source of energy, and we need a balancing to keep the system stable. So that certainly continues. And we are working on the balancing side. We are working with the utilities, our traditional customers and IPPs, new IPPs coming in. So that is very positive. And on the baseload side, I think there are several trends. I mean if I take globally, we have electrification of several industry transport, et cetera. People will use more and more electricity that drives base loan demand. In the U.S. and also in Europe, we have an aging power infrastructure. So there is renewal rate or needed of generating assets that also feeds into this. I mean, on a global scale, the climate is getting warm, so there -- the demand for air condition, it's quite a big driver. Look at the data from international energy agency, it's the same magnitude as data centers globally. So that's -- and then we have the data centers that we talked about today. And here, we work with utilities and IPPs, but also industrials and new -- no new customers on the data center builder and integrated I mean our focus is clearly to support our customer to maximize their value. For us, it's a profitable growth opportunity in both new build and services in a dynamic, and I say this in a positive way, that market for Thermal power.. So if we sum it all up, Click -- coming back to power for data center is a significant growth opportunity for us, both in new building services off-grid power is expanding rapidly. Our technology is very well placed. We see a strong demand also in the existing grid market and baseload power it will be there, we have a high, high value service potential as well. So data center is a very interesting opportunity for us. So that was the message for today. And thank you, Anders. And we give over to Hannamari and the team to -- for the Q&A.

[Operator Instructions] First question is coming from Vivek Midha .

Thank you very much, everyone. Good afternoon. I hope you can hear me well. And thank you very much for doing the call. Really interesting, lots of really interesting data points. I'm interested in understanding a bit more about costs in the Slide 17, particularly on a per megawatt basis, given a, how fluid the pricing situation in this industry has been in the last few years. The fixed costs in the chart also includes things like fixed maintenance and so on. Part of the comments you made around CapEx being about the extra redundant capacity you need on the turbine side. Would you be able to share with us the CapEx per megawatt assumptions that you've used to underpin this case study? And maybe if you don't want to be as specific as that, would you maybe be able to talk about your relative competitiveness on the pricing per megawatt has the inflation we've seen on the turbine side helped you in becoming maybe more competitive on price with the turbine players? Or is that a minor contributor has that not really been a driver

So if I start, Anders, and then please feel free to chip in. And I mean, the challenge here is really to to compare apples with apples and not apples with bananas because I know -- and I fully understand, everybody is trying to assess what's the cost per kilowatt hour and kilowatt meat, et cetera. It's not easy because, first of all, you compare different price points and what's the scope? Is it fully PC? But even if we go to people if it's not, let's say, it's equipment, how much of balance of plant is included, how much of auxiliaries are included. So even if, for instance, it's talking about ourselves, when we -- when you read official reporting, you will see that this varies quite a lot. And even if we define it as equipment and it's because -- the scope varies quite a lot. And that's why it makes it very hard to compare and I understand you want to do it and you really need to triangulate and work with interest that I would encourage. I will not go into all the details. And I think you understand why. Now in terms of -- where do we stand with the [indiscernible] by guys goes? I think we definitely have an advantage of shorter delivery times in general, although our delivery times are also getting long for certain engines for other engines, I think we still have have relatively short delivery times. But I would say because some people say, well, you're selling your engines just because the gas turbine gas are sold out. And I would challenge that statement because of this, I'd like to refer it as the TicTac to, these 9 -- 3x3 metrics of technical advantages. And you also see that if we go to page -- the following page there, if you can help me. Because it's -- I mean the world is not black or white. And you know it, but I want to restate that. I'm coming. I'm just waiting to somebody to put a -- so this 1 No, that's excellent. -- number 14 -- this 1 because here, we're trying to -- how customers evaluate the technology this -- we try to synthesize how does a customer evaluate the different technologies. And in this case, of course, it's our slide. So we come out winning. Obviously, we don't always come out winning, but we are coming out winning. And then it's because customers are clearly evaluating the offerings in different dimensions. And some of the technologies are stronger in certain dimensions other technology like ours are strong in other -- certain other dimension. And then the customers make a holistic assessment. And of course, customers, they have different preconditions. They have different focus CapEx, OpEx, not 1 customer is like the other customer. And that's why because sometimes I don't get challenged. If your technology is so great, which it is, why don't you just win everything? And of course, we will not win everything. But we will -- my key point is that we have a place to play in this market. I mean it's moving into our suites, but this intrinsic advantages, they are there. It's not marketing on a real -- and that will mean that we have a very strong proposition, not for all customers, for many customers. And I also make the point that once customers. We have many new customers that are coming in now. They are trying, we are talking to them. Once they try the new candy engines, I'm critical convinced that they will like them because I have repeat customers in the U.S. They're coming back. Lower Colorado with authority, waken et cetera, et cetera. So -- and they are coming back, not only because we are nice people, they're coming back because they can see that this is real. Now of course, then to pricing and strength, et cetera. I mean, it is, of course, if you look at supply and demand, there is a short as of supply. And this, of course, gives all the manufacturers certain pricing power. But I would also like to balance that a bit because we have a customer and the customer needs to make the business case line for them. So it's a balance. It's a good old balance. But in general, the business is certainly contributing to that overall profitability from a life cycle perspective.

Next question comes from Uma Samlin.

It's super helpful for us to understand a bit more on the data center opportunity here. So I have a question on your slide the same slide from earlier Page 17. So it seems like you're saying you're OpEx cost is a bit lower than aeroderivative turbines as a sort of comparable to the combined cycle turbines. How do you calculate that? Because what we have heard from the turbine makers seems to to suggest that engine maintenance will be a lot more costly because we will have to maintain them more frequently on a frequent basis, like every few thousand hours. So just more to understand a bit more the how you look at that.

So even though it's -- you're right, it's more frequent, but it's also the cost of the major maintenance that we have on the turbines that are much higher than the cost of the engines. And then also the OpEx, it's not only maintenance, you also have the fuel being part of that. So obviously, with the efficiencies, difference between arrows and medium-speed engines, you get the big part of the OpEx difference is, of course, the fuel. And we can, of course, see that also varying with price, but with a big difference in efficiency, a large part of the OpEx difference is the fuel. But the maintenance itself, the OpEx maintenance itself say, it's not a big difference between aeros and engines. But engines have more continuous maintenance while you have bigger maintenance coming up on the areas that are more costly.

But I guess what are the sort of pushbacks you get from your customers? Because if you look at Sineenergy, the book like 36 gigawatts of orders, you're looking at a few yes, less than 1 gigawatt year-to-date. Just wondering what are the pushbacks you get if you both have a short delivery time and also cheaper sort of both from an OpEx and CapEx perspective.

So I would say that the could be many things. One is obviously that I think that many customers knows the turbines and the turbine suppliers very well from from the past, so to say. So that's 1 thing. We are less known, especially in the U.S. than if I compare with GE and maybe Siemens as well. And the other thing is, I think, Waka sort of alluded to it that when you have bigger sites and if you have site constraints in terms of space. Obviously, the density, the power density is higher with turbines than with nets. So that could be another reason. And as we also said, every cycle is different. So there could be specific reasons for that particular site. So it depends. I mean it could be -- I show the emissions here as the overall emissions in CO2. That's 1 aspect. But you, of course, have PM2.5, you have other things and emissions, especially in the U.S., is not only measured on the data center itself, but it is what is in this area. So if you have a coal power plant, for example, you have already used up a lot of the particles already for that. So then obviously, in that specific site, that might then play a bigger role, which might be to our digital vantage. So I would say that it's very, very site specific. What is the criteria on that particular site.

We are to give some further color to the whole topic because -- we need to move away a little bit from this that the world is black and white, and there will be only 1 solution. That would be several solutions. And I think this take to advantages that we have combined with delivery time, that will build us a Hansen business. I can say that. And customers, they believe in it. I had a new customer here recently in the region. I cannot say who it was, but I will give you the context. So it's a U.S. customer. It's a big -- 1 of the big IPPs and they -- on a senior level, they said that we have great engineers in this company and they love the gasolines. I mean they have been working with them for the whole professional life. And I have been trying now because all the constraints to get my team to try the engines so difficult because it's a conservative people engines is a new technology, so this senior leader, he came to I want to do a project and a project with you in another country because then I can use that as a kind of change management to get my engineers to really look at the engines because I know when they look at the engines, they will see what you have been saying for a long time about, it's true. And then once we've done that in country X Y, we will bring that to the use. So there is this journey of technical conservatism you could say 1 area where we need to develop and improve is to get our message out there. And then, of course, the proof is in the pudding. And that's why I bring up this with the repeat customers. I mean, yes, you can sell something, but if you want customers to come back, you need to deliver on your core promises and that's why I've been talking about that over time, and now I'm talking long term, I think this is a fantastic opportunity for us to prove our technology and over longer time, build our market share.

Seem maybe to add to what you said, Hakan. It's a little bit like if you haven't been driving an EV, you worry about a lot of things. And then when you have been driving an I think those of you that have done that, you know that it's not a big worry and you probably won't go back. It's the same thing.

Next question comes from Ante Kansanen.

Yes. Thank you, guys, and I wanted to continue on the same theme a little bit referring to the Slide 6. kind of a value chain that you are looking at and talking about the developers and utilities. And when you guys talk about kind of having some challenges of being less well known and being kind of a challenger with the engine technology, are you now referring to kind of the -- exactly the utilities and the developers who are operating in this space. I'd imagine that the end users, the hyperscalers are not the ones who are making the decisions. So I would want to better understand kind of the scope are you operating with multiple very fragmented base of developers on certain projects, do you believe that there will be kind of a few bigger ones that maybe are already familiar or will become familiar with the engine technology. So there would be some scalability into kind of penetrating better into the developer space. So maybe talk about that a little bit.

No. So definitely, I think there is an advantage, so to say, to work with the ones that have a pipeline of products. We have many small developers. But those projects take longer time and also more unsure. But I think that the ones that have a good pipeline of projects and also prove that they are in this business and have been managed to pull of the projects that we're working with them, which we are doing now. the first order come from such company. I think this is a benefit. And when you talked -- you also asked about the utilities I would say that if I look in the U.S., historically, so to say, we've been very strong with the local municipalities. As also what Hakan mentioned, the big utilities, they have a lot of gas turbine people in their organizations. And what I'm very happy to see is that in the last years here, we're also starting to make good inroads to big utilities. The advantage is with that, and we have repeat customers now on the utility side, the advantage with that is that they also have a long pipeline. If we go to municipalities, they don't have a long pipeline. They are not the same size. So it's good, and they will talk to other municipalities. So from that point of view, we view it's good that we can get new order from another municipality due to a good reference. But typically, they don't have many orders like utilities that are much larger. So for us, it's definitely an advantage that we have made inroads to bigger utilities and also to bigger data center developers that have pipeline of projects. That's clear advantage, and that's what we're doing. .

And kind of if you look at the opportunity for the next, I don't know, 5 years, are there some kind of -- and the focus areas that you want to target with whether it would be choosing certain kind of developers, utilities, certain regions, areas in the U.S. that you would be targeting more .

Yes. We have, of course, now I'm not talking data centers, I'm talking energy market in general. It's quite clear that we have been focusing on certain regions. And if you follow our press releases, you can also see what those regions are. So of course, we target those were having a bigger need for balancing applications. And then on top of that now also the data centers where they are located in just so you can see where there's a lot of renewables and whether it's a lot of data centers to regions where we target. .

Next question comes from Anders Erber.

I wonder if we could throw in energy storage into the discussion, mainly from a technical perspective, I mean, I'm looking at fluence presentations here, they talk about 36 gigawatt hours of pipeline. And I just want to understand how that would impact as well your solution from a competitiveness, they talk about smoothing out the peaks basically. And I understand there could be upside, of course, given that you have the business as well in-house. But mainly, how it interacts with the engine offering would be interesting at this point?

So basically, I think this is an evolving area. I mean the -- you're fully right in the sense that specifically the learning data centers. They have big loads, I mean, tens of megawatts swings in the second millisecond time domain. And these are fairly challenging from a technical perspective. So a different data center developers, they use different technologies to kind of smoothen out, we mean low pass filter and batteries is 1 of the technologies, but also the ones condensers. There are different type of specialty solutions, so to say. So -- and those type of -- I don't know whether colonsmoothing technologies, you need them for you as turbines, you need them for engines as well. But different methods are used -- now battery storage is 1 of the solutions, clearly. And could this be an opportunity for our storage business. Yes, it could. We are still evaluating this from several perspectives. So we -- that's why we haven't brought storage here today. We emphasize the thermal side because this is where we are formulated and then we are starting the batteries.

Next question comes from Sven Weier.

It's coming back to the competitive advantage, but not against turbine makers. I mean, I guess, every 1 of you will talk their own book and how great they are. But I was wondering more about your edge against other engine makers, right? I mean there are a few what's the local ones. What are the -- what do you find there as an edge or as an upside, maybe? .

I think we have -- I think we showed in the table on Page 16 our relatively so the competitiveness to high-speed engines. And of course, since both our engine technologies, 1 is highspeed 1 is medium speed, there are differences, but they also share many of the advantages, so to say, with start-up time and thermal capability and all these things. So of course, there are smaller differences. I think if you look on the high-speed side, Clearly, the size of engine is 1 difference. We also showed on the efficiency that you have some difference. So there are differences to the high speed. And if you're asking because you didn't specify that, but if you're asking to other medium-speed engine suppliers, I will say that we clearly have much stronger network in the U.S. for making sure that these plants can run on a very good service and therefore, run with a very good availability. I think that this is key for data center customers to reassure themselves and not only data center customers back away, it's also true for balancing customers that it is important for them to make sure that these engines start when they should start and they are maintained to the best quality and that they are always able to deliver the availability that is key. And I think we have with the network of service network that we have around the world and especially in the U.S. that I think is absolutely an advantage for us.

And , if I may add fully support ones, but I would say that the high speed, they had 1 commercial advantage to us, clearly. -- because they have been dealing with the data center segment much longer than we -- because as you remember, backups. For backups. -- exactly. So when back in the days when data center were storing, moving to AI -- so they know customers, they have well-established customer relations, the sweet spot is moving into our markets. So of course, we are -- endos and his team is working a lot on building customer relations, et cetera, et cetera. So initially, high speed has a customer relations advantage, but we look at because we have a very competitive technology. .

And how important is fuel flexibility and advantage relative to the gas turbines, we think about later upgrade to hydrogen -- is that important at the moment at all or not really? .

I would say that at the moment, that comes further down on the list. It's another parameter, but I will say that that other things like the modularity for sure the lead time, but the modularity and scalability with that and also the water consumption is probably higher than the future alternative fuels, so to say.

I would say, and if you remember this 3-stage rocket, or where we want to call it, you remember, I put the renewables. I think that is -- that is more of a -- in focus in long-term thinking. I mean right now, everybody is rushing, but I think the hyperscale as they certainly have in the back of the mine, we need to go green, so the balancing property, I think, is at least my perspective, balancing property is more important than alternative fuel capability. Now as you know, because you follow us, we have -- we are well prepared for alternative fuel and also because of our Decor strategy. But I would say that from a market perspective, I mean, the balancing capability will be more important.

Next question comes from Vivek Midha.

Just 1 follow-up on a comment you made earlier around things like PM 2.5. And 1 thing we have heard from certain industry players is there are things like nitrogen oxide, methane slip and so on can also be topics where potentially, there are questions about the reciprocating engines. Is that something which you also here coming up? Is there an issue around, say, data centers, which are closer to population centers and so on. how important is size is a factor?

I think so far, we -- as I mentioned, we have a case of 2.5 I don't know -- I'm not aware about any other cases what you mentioned with Melton Skip, et cetera, that, that has been a deciding factor against us, so to say. So I'm not aware of such a case.

I mean technically, to deal with NOx, we would, of course, apply the SCR, et cetera. So then you would give will eat out a difference. And it really comes back to this slide about different customers use different criteria. They put different emphasis. And the world is not black and white. But we will win and sometimes we will not win clearly.

Next question comes from Robin Fiedler.

I think it's clear from what you're saying today that there's obviously pros and cons for both turbines and engines, and they both have a place in this market. And given what's for now a clear time to power advantage given your lead times and given the value that, that has to the end customers, I'm just I'm just so curious why so far your engine pricing for data center seems to not have shown any upside yet. Is that just a function of when with your initial deals, the pricing terms were set maybe a year ago before there was a severe tightness. I'm just trying to gauge what the pricing opportunity is directionally, if any? Or is this more a volume opportunity?

No, I mean there is a good price realization. But as I said, and I know you're always trying to look what did we invoice and you divide by megawatts, et cetera. But unfortunately, you don't see -- I mean, -- it's very hard to see through the full complexity of that. And as I said about you need to make sure EQ, what type of EQ is it or the balance of plant, including oil, et cetera, we don't communicate that, so it's impossible for an external party to see through that complexity. So that's why 1 should be very careful that saying, yes. We certainly know that establish have increased our prices. We also have a good price realization -- at the end of the day, we have a good competitive technology that the customers can build viable business cases make money on. I think that's the key thing. And I have said it many times that the data center business will contribute to the vases profitability. And also, I talked about deliveries are starting to ramp up this -- gradually this year, et cetera, et cetera. So it's too early to evaluate in the profitability of data centers and looking at our current interim reports in to.

And also can please also note that this -- the comment that we made on the service business in a meaningful way, I mean, in a significant way, will kick in 2030 and beyond, so to say. And there's nothing strange for that. That's just normal power plant business.

And for the new build, we will start deliveries in end of the year and going into '27 4%. So .

But is it -- sorry, is it still fair to assume that you're able to get pricing upside. I don't want -- I'm not asking you to give me a specific number, but let's say the turbine players talk about well over 50% and they're saying most recently that the latest slots are 10% to 20% even higher. And so you must be getting some appreciation. And again, you don't have to give me a .

I understand your question. I respected by any more comments. I will not make those statements that do just on by competition. I'm just saying this business, I mean, supply and demand is giving opportunities for price realization. I'm saying that this business is definitely contributing to the overall profitability of that -- and then to balance, we need to our propositions that are competitive for our customers, and so they can make a good business case

Next question comes from Gary Parker.

Do you hear us? Is it now better? Can you hear me?

Yes, we can hear you. Thank you.

Okay. Thank you. So I think we would agree that the data centers, the at size data centers will be mostly off-grid because of local politics and also you don't need to pay for the grid operator anything when you have a local power supply and steady one. So in that respect, Wartsila's sweet pot is below 100 -- 1,000 megawatt 1 gigawatt. So you are in a kind of a complementary or then you have to scale up the product portfolio. Also, I think it's a wonderful decision to scale up the production capacity in Vaasa, but also 1 could wonder that's the question number one. Is this ambition level that you will have for the '28? Or is there coming more and the second part of my question is on the portfolio because when you look at in your Slide #16, you had competition place there. Siemens 73-megawatt gas turbines 5 of them, GE 33 megawatts, 11 of them. And then you put Wartsila engines and 39 of them. Why don't you put their bigger ones like the 50 SGs which you need only 20 or so? Why is this kind of a comparison that you want to show. So these 2 questions about portfolio, can you stretch the capacity up beyond 20, 30, 50 megawatts. I don't know if it's technically possible. And then about the capacity that would you be satisfied with this 35% increase in Vasa in 28? Or is there more .

So on the engine question, yes, we have different engines to be used. And in this case, we used the 34 engine. We could also have used to be 50 engine that we have sold in the projects. The difference is that 1 is around 9-something megawatt, and the other 1 is 19 megawatts in [indiscernible] . So that's the difference between the 2 engines, but you could have as you say, you could have used the 19-megawatt as well to make that comparison. .

[indiscernible]

That was on the first and other experience. and expansion I would say this is now the expansion that we have decided and communicated. And depending on what happened on the market and what is pointed demand and supply, we are, of course, always looking and evaluating if we should do something else, but this is what we have now decided. .

The next question comes from Louis Bellon.

Can you understand me?

Yes.

So do you have bottleneck for ramping up your industrial capacity? And do you think it's easier for you than your competitors to increase capacity? And maybe in terms of, I mean, how many orders are left before having the same delays in delivery. .

So if I take a first cut, I mean, quietly, I don't know the data of the guest on my supply chain. So I mean, I think we all read about blades, et cetera, but you will have to ask them. I think we are doing a very solid work in terms of making sure that our supply chain can follow us. That's why, I mean, 1 could say proof point of that is one. But we announced this strategic operation with Simple company in December, which is 1 of our major casting suppliers. And so they, of course, are with supply. There are several of those vital suppliers. We are working very actively with that. And when we talked about the EUR 140 million investments, the majority is in the stage, but we will also invest in our supply chain. So there's a structure of work ongoing. And we feel confident we will make it. It's not 100% secure, but made real confident and then you have to have the guests how they see the supply. It's not the same supply chain just at that point because the technology is different, so to say. So I think that was the first one. sorry, if you can repeat.

It's how many orders are left before having the same delays in delivery because you are already 13 months. And how many orders before reaching 40 or 50 months.

Yes, I couldn't even answer that question. What I can say, and this is what we have communicated already for certain engines. And we have not said which types, but just to give you a certain eins, we are looking at delivery times, if you order it now, in 2028. And other types of venue we can still deliver towards the end of this year. So it's a mix of different engine types. I will still say that overall, -- and based on what we see, what competition is communicated, I think the fact that we have shortened times, it will be a competitive advantage with us for quite some time.

Next question comes from Ante Kansanen .

Yes, I had a follow-up question on exactly the same thing, which is kind of the capacity constraints. And for example, in the Marine business, I mean, you would maybe have an extra capacity, but the yards are constrained, so that doesn't really help in terms of delivery time. So are you -- are there any kind of risk? Or are you seeing any signs of something else kind of curtailing the delivery schedules that you would not actually be the bottlenecks regarding the projects that you are now quoting? And then again, the faster delivery times that you might have versus say, a turbine competitor doesn't really matter because the project can't go forward with your schedule anyway. So what are kind of the other limiting factors on those U.S. projects that you are working with? .

I would say now that it's a very dynamic market, and there is a very high level of demand. And when I talk to our numbers, you can also comment when we talk to our customers, they say that this -- I mean, based on the visibility that they have, this will take us to '29, 2030. So from a pure market perspective, and the market feedback, there is no signs that this dynamic will change. Then of course, if you ask overall, what's the biggest risk element for Wartsila right now? It's the geopolitical situation. I mean we all know the dynamics between China and U.S., U.S. tariffs, et cetera. Those are, of course, risk elements where we don't have control over. But when it comes to the fundamental demand side, when we talk to our customers, it seems to be strong, and it's strong -- I mean, all the way to '29 .

How do you currently deal with kind of the tariffs given that, let's say, the delivery times are then stretching to 2029 or 2030. Is it just that the developer or the end client base? Are there some type of share risk profiles in between .

No. This is customer on step risk.

Next question comes from Daniela Costa. .

Hope you can hear me. This is a super helpful call. Thank you for putting together -- just 1 question. I know you mentioned the aftermarket would kick in like later in the decade. But I think some of the experts we have with the flag that sort of these edges are meant to run at peak basically all the time now bit different to when it's a utility type environment. Does that change the aftermarket intensity and your potential penetration of the installed base compared to where you were today? What do you envisage there .

Yes. So we have said, and I want to clarify that this is a base loan application. That means that we have many running hours. However, we should remember that there is also built in an extra maintenance reserves, so to say, for having the availability very high. which means that you have extra capacity that you are not running at the same time to the pit. So the overall running hours are, of course, not 24/7 in the whole year since we have those extra engines or gas turbines in the case of gas turbines sitting there, so to say. You can see that very clearly on the page. So if you have 300 megawatts as the full capacity to be delivered, then you have extra in our case with the size of engine we have used here. We have 57 megawatts of extra capacity that is actually not running down. And then I must also say that also depends on the season and the load of the data center because, of course, all data centers are not running earlier around at the maximum capacity of 300 megawatts. They should always have the possibility to go up to 300 megawatt, but it might also be that they are running at 80% on average or something like that that -- there are many hours that is still the case.

You get the north the logic here. I mean if we deliver something Yes, there is an additional engine, but the other end is there running full speed 300 megawatts. So therefore, it's an application that generates a lot of service hours, so to say, even though there might be 1 or sometimes even 2 engines that are not running at all. I think you get the logic in the cost.

Yes. Maybe actually 1 question on the slide because you mentioned there some of the gas turbines might need a backup power plant. Would that be a Wartsila engine as a backup. Is that what you .

No. In this case, it's pure, it's pure, so to say, CCGTs or arrows or medium-speed engines in this case. What we are saying is that when the sites grow bigger, like the 1 gigawatt, we might as well very likely see combinations where you probably have a big gas turbine to deal with the power density when you have some -- or you don't want so many engines -- and then you have engines for dealing with more ups and downs in the load and the impact. So I think we will see hybrid and we can already see examples of hybrids today. And I think we will see more of that, the bigger data centers get. .

Next question comes from Johan Eliason.

I hope you can hear me. Good. I was curious, you're making a very good case for the medium speed engine in the data center opportunity. Historically, in the power plant market, you've had very high market shares for your supply. Is the data center an opportunity for your sort of more marine exposed guys to get into the power plant business as well? Are you seeing them at all? .

We are seeing them, and you can find press releases from our direct competitors on engines as well. But I would say that -- we have a good market share, a high market share in, so to say, normal market, and I expect us also to add up on the data centers because of the references and because of the service network and all these other things also as having efficient dent. I believe that versus the other engine competitors, we will have -- I expect to have the same type of good market share. But we should also remember that in the overall market, if you also include gas turbines, I think we have plenty of room to grow our market share because it's not so high if you look at the total market. And it's the same, of course, on data centers today.

Then I have received a couple of questions by e-mail. So do you worry about overcapacity given Caterpillar expansion plans in relevant engines. So this is about competitive dynamics, but medium and high-speed engines.

So I mean we don't see high speed as a major competitor in our sweet spot, and we talked that through. I mean, high speeds, they are -- they have suddenly the place to play in the small power intervals, and we are not competitive there. When we go into our sweet spot, and the efficiency is much, much better, and the life cycle cost makes our technology a winner. So I'm not so concerned about Caterpillar expansion. I think they reactivated MIK, I don't know where U.S. time pending and factor reactivating in Germany for export to the U.S. I don't know what they will offer, et cetera, today, let's see. I think in the days when was doing in Marine, I think we had a very competitive offering from technology and commercial perspective. So I think we will have a competitive offering also there going forward. So then, I mean, we should definitely acknowledge that many of the gas turbine players, they're also making significant investments. And you follow it and Miso talk to them surely do. I mean this is also in the 16% and 18%, et cetera, et cetera. So of course, there is -- there will be new capacity coming into the market. And -- but it's also a very hot and buoyant market. And I think we share the same view that this is not driven by 1 data center factor, but it's also we talked about aging infrastructure that needs to be replaced, both in the U.S. and Europe. It's heating, it's electrification, it's balancing power, et cetera, et cetera. So it's not just 1 driving factor as -- and I think also, I mean, I say a little bit similarities with ourselves and some of the -- I think I note that at least when I have seen the press releases, many are expanding their existing facilities. That is normally a relatively efficient way to expand the capacity but also with higher flexibility. And if the demand would come down, it's easier to deal with that type of then adjusting the capacity if you have expanded within the existing facility.

And another question by e-mail. Hakan mentioned the Middle East opportunity, the CapEx OpEx comparison for the different technologies that understored. Have you also made a comparison with 24/7 solar plus battery farms that are now being planned in the United Arab Emirates.

Yes. I mean we do balancing -- so we haven't done the comparison for the UAE. So I have to pull on plan on that one. But I would say that we provide balancing also for wind and also to certain -- I mean term of balancing. -- also for solar. It depends on the application. It depends because batteries are great for handling corepower swings. I mean batteries don't generate energy, we all know that. But they are great on handing short power swings. But if you talk about seasonal swings, the batteries become very large and therefore, very costly. And then you need to go to -- so I have to pull a blank in UAE because we haven't done.

I think that the case in UAE that they refer to is the case where they have built the bridge the whole day, so to say. You know the battery during the day and then you bridge the whole night. So it's not the 2 or 4 hour battery full overnight battery, which doesn't take away what you say will at if there are seasonal swings, it will help. But of course, it is different to what in most other places, they do on our batteries to our batteries, for our batteries or even in some cases of 8. But of course, to do the whole night is a bigger battery. And we have not studied that in detail what that will mean.

nd then the last question comes from Sven Weier..

Taking the question here. Just to follow up on the water usage. I was just wondering, I mean, how much of a difference does it make between wet and dry cooling the need to run the comparison? .

I know that there is a big difference between the 2, and that's why we chose here to show 2 different examples, so to say, there is a difference.

But did you have an overview in terms of the projects? How much can be done with dry cooling and how much cannot in the U.S. pipeline?

Not on top of my head, at least.

Let's take that with us. Let's see if we can do that on this well you're deep into the details. Well, let's see. But maybe too much. I mean we like those engineering-oriented questions. Let's look -- I mean, more from a customer perspective, if I talk about the market feedback that we are having, and I'm sure you read recently the New York Time optical on Microsoft and their challenges related to what you as Anders pointed out, it's not only in these type of environment. But it's a normal environment. Formenthere is increasing considerations for water and this recent customer interaction that we had, that was caring 1 of the variables. So I think water will come in this evaluation, going back to this, looking at different dimensions, water will be increasingly becoming important also in the U.S.

And thanks for the call. Thank you for parting questions and good answers. So I think this has been a very lively discussion. Present call will take place on March 23. So I hope to see you there. Thank you.

Thank you, everybody.