LGRVFLGRVFOTC
Loading
Operator: Good day, and thank you for standing by. Welcome to the Legrand 2025 Full Year Results Conference Call and Webcast. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Benoit Coquart, CEO of Legrand. Please go ahead.
Benoît Coquart: Thank you. Good morning, everybody. Franck Lemery, Ronan Marc and myself are happy to welcome you to the Legrand 2025 Full Year Results Conference Call and Webcast. As you know, this call is recorded. We have published today our press release, financial statements and a slide show to which we will refer. I begin on Page 4 with the 3 key highlights of this release. First, Legrand delivered a remarkable performance with record sales growth, high profitability and a strong achievement of its CSR objectives. Second, the group continued the successful deployment of its strategic road map towards EUR 15 billion of sales by 2030. Third, Legrand is targeting further sales growth of between plus 10% and plus 15% in 2026, excluding currency effects. Starting on Page 6 of the deck, we fully achieved our annual targets for 2025, which we will detail further by key topic during this presentation. Moving to Page 7, I will start with an overview of sales. In 2025, excluding currency effects, our sales grew by plus 13.1%. This includes an organic growth of plus 7.7%. This growth is driven by an outstanding performance in data centers with an organic growth of close to plus 40% this year. And regarding our sales in buildings, we are resisting very well despite a still muted market over the year. On top of organic growth, we benefit from a positive scope effect of plus 5.1%. I will come back later on acquisitions. Of course, now based on the acquisitions announced and the likely date of consolidation, the 2026 full year scope impact would be close to plus 6%. As for exchange rates, the effect was a negative minus 3.1% in 2025. And based on the rates of the month of Jan, it will be around minus 2.5% for the full year 2026. On Page 8, you will find the key takeaways per geography on a like-for-like basis. In Europe, in a market that remains mixed overall, sales were up plus 1.9% over the year, including, for example, in Germany, Italy, the Netherlands and the U.K. In North and Central America, sales were up a strong plus 16%, driven by an outstanding performance in data centers. Finally, in the Rest of the World, sales increased by plus 2.7% with good growth in Asia Pacific, Africa and the Middle East, partly offset by a retreat in South America. These were the main comments I wanted to share on sales. I will now hand over to Franck for more color on our financial performance.
Franck Lemery: Thank you, Benoit, and good morning to all of you. I will start on Page 9 with adjusted operating margin. We recorded in 2025 a very solid adjusted operating margin of 20.7% of sales after acquisitions. This represents a plus 20 basis point increase year-on-year, including a 10 basis point organic improvement and a plus 10 basis point favorable impact from acquisitions. The group's high profitability demonstrates once again the strength of our strategic model and our solid capacity to execute and adapt I think notably of the volatile environment linked to the U.S. custom policies, which increased the group cost base by around $100 million. Going now to Page 10. The net profit attributable to the group stood at EUR 1.2 billion, represented 13.1% of our sales. The increase coming from the operating profit is partially offset by the impact of financial results, while the corporate income tax rate remained stable. The free cash flow came to a solid EUR 1.3 billion at 14% of sales and a conversion rate of 107% supporting the sustained acquisition momentum of Legrand while preserving balance sheet strength with financial leverage kept under control at 1.9 at the end of December 2025. This is it with the key financial topics I wanted to share with you this morning. I'm now handing over back to Benoit.
Benoît Coquart: Thank you, Franck. Let me now move to our 2025 CSR performance. On Page 11, in 2025, Legrand reached an achievement rate of 110% on the targets set for the first year of its 2025-2027 CSR road map. You will find on Page 12, a few illustrative examples highlighting this performance. For example, Legrand outperformed its targets in terms of Scope 1 and 2 CO2 emission reduction, plastic packaging reduction or use of sustainable materials. All the achievements confirm the strong integration of sustainability into the group's strategy. I'm now moving to Page 13 to conclude on 2025 performance with our dividend. The approval of the payment of a dividend of EUR 2.38 per share will be proposed to the next General Meeting of Shareholders. This represents a rise of plus 8.2% from 2024 and a payout ratio of 50%. Let's now move to the second key topic of this release, our strategic road map. In 2025, Legrand actively deployed its strategic road map towards EUR 15 billion of sales by 2030, combining accelerated growth and value creation. As shown on Page 15, this is first illustrated by the reinforced positioning of the group in energy and digital transition offerings which now represent 53% of our sales compared with 47% for essential infrastructure solutions. Data centers at the heart of the group's growth strategy represented sales of EUR 2.4 billion at year-end 2025, i.e., 26% of group sales to compare with EUR 0.7 billion in 2020. Legrand is recognized as an undisputed major player in this field of activity with a deep offering that is perfectly suited to the deployment of infrastructure for artificial intelligence. Building on nearly 30 acquisitions completed in this field, Legrand has become a leading player and a preferred partner for major industry participants. The side positive impact of all the acquisitions we made in data centers is that they also strengthened the group's existing position in critical power with other verticals driven by electrification, such as infrastructure, industry, telecom, oil and gas and microgrids. I am now moving to Page 16 to 18. As you know, innovation is really the DNA of Legrand. We highlight on those slides a number of product launches carried out in 2025, which illustrates the sustained momentum in innovation across the group's segments and geographies. On Page 19, we underline the group's continued focus on digital initiatives and customer experience with high and improving customer satisfaction in 2025. Finally, on Page 20, we detail Legrand's particularly active M&A strategy with 7 acquisitions announced in 2025, representing EUR 500 million of annualized sales, all in the fields of energy and digital transition. This momentum extends into 2026, as shown on Page 21, with the announcement today of 2 additional acquisitions in data center in the U.S. with Curtis Industries and in Brazil with Green4T. Maybe one thing to add that is not in the slide actually, but in the press release, we recently invested in Accelsius in the U.S., a pioneer in 2-phase direct-to-chip liquid cooling. This investment will further strengthen the group's portfolio of solutions for AI and HPC data centers. Let's now move quickly to the third part of this release with our targets. On Page 23, regarding '26, Legrand will continue to accelerate its profitable and responsible growth momentum in line with its strategic road map. Taking into account the current global macroeconomic outlook, a very strong data center market and a modest recovery in the building sector, Legrand is targeting the following in 2026, sales growth, excluding currency effects of between plus 10% and plus 15%, comprising organic growth of between plus 4% and plus 7% and growth through acquisitions of between plus 6% and plus 8%. Adjusted operating margin after acquisitions of 20.5% to 21% of sales, CSR achievement rate of at least 100% for the second year of its 2025-2027 road map. On Page 24, regarding our 2030 ambitions -- sorry, building on its achievements and taking into account both observed and expected market trends, Legrand is confident in its ability to reach the upper end of its 2030 sales target range around EUR 15 billion with average annual sales growth of close to 10%, excluding exchange rate effects and average adjusted operating margin above 20% of sales compared with the previously targeted level of around 20% in average. This is it for the key topics of this release. Last word before we move to the Q&A session. You will find on Page 26 to 28, our corporate access agenda for 2026 should you wish [Technical Difficulty] management. Let's now switch to Q&A.
Operator: [Operator Instructions] And we're going to take our first question. And it comes from the line of Daniela Costa from Goldman Sachs.
Daniela Costa: Thank you so much for taking my question and the follow-up. I will do them one at a time. But starting on the guidance for 2026, can you give us some help on the building blocks, particularly how much data center growth are you factoring in? And how much of that comes from a backlog you already have? And what pricing are you factoring in there?
Benoît Coquart: Daniela, so our organic guidance, the 4% to 7% growth is basically built this way, a growth in data center of between plus 10% and plus 20% and for the rest of the activity, i.e., the building activity, something basically more or less flat in volume with a bit of pricing. Those are the building blocks. Well, how confident are we on the fact that we're going to grow from plus 10% to plus 20% in data center? I have to say that we are very confident. Not much based on the orders. The lesson of last year is that it's difficult to anticipate what we're going to do based on the orders in hand. Last year in Feb, we -- based on the orders we had in hand, we targeted plus 10% to plus 20%. And at the end of the year, we did plus 40% or close to plus 40%. So it's a bit difficult to that. So we have to rely not only on orders, which are very good, not only on the book-to-bill, which is above 1, but we also have to rely on the feedback we get from the market and the CapEx plans announced by the hyperscalers and so on and so forth. And based on those information, we are very confident on our ability to do something between plus 10% and plus 20%. As far as pricing is concerned, overall, not specifically on data center nor specifically on building, we are shooting for pricing somewhere between plus 1% and plus 2%. Now it is based on our scenario when it comes to the raw mats and components. So we believe that the price of raw mats and components will be up between plus 1% and plus 2% this year. But of course, it can change. And if for whatever reason, the price of raw mats and components was to be higher than expected, of course, we would do a bit more pricing. But based on the scenario we have in hand today, we believe that the plus 1% to plus 2% price should be enough.
Daniela Costa: And then just following up just a bit on the acquisitions and the investments that you have been doing. I think one of the deals today has more of a power distribution medium voltage component, which I believe you hadn't done too much in the past. And then the Accelsius was into liquid cooling, although I guess it's just an investment rather than a full consolidation. But can you talk us through sort of how you're pivoting the portfolio in data centers? Do you want to go into medium voltage? How far out are you in the gray space right now, just to get a bit of a shift on the mix?
Benoît Coquart: Well, it's -- thank you for asking the question that you have because I sometimes feel that we haven't done a job good enough in explaining how deep our portfolio in data center was. So a few numbers. For example, we are already a bit in medium voltage. We've been selling for quite some time, medium voltage, low voltage transformers, cascading transformers to data centers and to a number of other spaces. So -- but maybe let me give you a few numbers. The split between white space and gray space would be something like in 2025, 75% white space, 25% gray space. You could note that gray space was 5% 3 years back, and it's now 25%. So we've been able to rebalance, if I may say, our portfolio as we committed to do at the last CMD. Now the split between gray and white doesn't give full justice to the breadth of portfolio we've been able to build, which is composed of close to 55,000 SKUs for data centers, standard SKUs. And if we put together the customized SKUs, close to 100,000 SKUs. So maybe the best way to look at it is to break down it by type of applications. So in 2025, we had about 35% of our data center sales, which was for critical power. So critical power, it's medium voltage, low-voltage transformers. It's UPS, switchgear, busbar, busway, remote power panels and a few other products. So 35% critical power. 50% of our sales relates to compute infrastructure with approximately half of that would typically be physical infrastructure. So racks, tap-off box, feeders, cable management, stuff like that. And half of that would be typically compute management. It's about monitoring PDU, rPDUs, KVM, consoles, transceivers and a few other products. So 35%, 50%. We have 5% of advanced cooling, which is rear door exchangers, containment, and now we have the ability to be more active on 2 phase direct to chip. And then we have 10% which is testing and life cycle services. That's where we have the power banks, have [indiscernible] power banks, but we also have installation, commissioning, monitoring, field services and so on and so forth. So you see we really have a complete set of products. So yes, we are a bit into medium voltage switchgear. We are already a bit into medium voltage transformers, but it goes down to rack components and even field services. I believe today, we have probably one of the most comprehensive range in the data center industry. And it is set to continue. We have a lot of ideas, both organically and inorganically to continue to build a strong catalog.
Operator: Now we take our next question. And the question comes from the line of George Featherstone from Barclays.
George Featherstone: Maybe I'll start with a follow-up because the color you just gave there was super interesting. And in the context maybe of the way the business will evolve to the 800-volt DC architecture. And perhaps you could give a little bit of color on what that means for you and how you're going to address this new technology in the future.
Benoît Coquart: Well, thank you very much for asking this question because I sometimes feel that the analyst community is a bit lost with these new architectures, and it's the opportunity to maybe a bit fear. So what is basically 800-volt architecture. You have a concept, which is sort of grid-to-chip concept, which we call in the industry, the holy grail, which is still on the drawing board and won't be at scale before 2030, 2031, 2032. What is currently almost ready, if I may say, is a sort of is 800-volt architecture, but slightly different than this grid-to-chip. It is based on what we call a side car. So how does it work? Basically, you have a powertrain with AC components that feed a power side car, which is located in the white space. This power side car convert from AC to DC, then feed a number of racks, and that's where you have the compute components, the cooling and so on and so forth. So 3 characteristics: increased power density up to 500 or 600 kilowatt per rack, use of DC components in addition to AC in the electrical powertrain and a sort of decoupling of power and energy storage from the IT rack and those components are moved into a side car serving one or several racks. So this is the architecture, which is almost ready and possibly at scale in a few years. How will the architecture impact Legrand? We have made a number of analyses, and we think that it will have a neutral to negative impact on about 20% of Legrand sales and a neutral to positive impact on 80% of Legrand sales. So the negative impact typically could be on rack PDUs, for example. It could be on UPS because UPS is replaced by sort of battery storage within the side car. It could be on a few other components. And the positive impact, well, it's on the AC powertrain because you will need more power, more amperage. It could be on the physical compute infrastructure because you'll have a wider racks. It will be on cooling, of course. And especially, it will push 2-phase direct-to-chip cooling. You will have rear door cooling for residual cooling, including actually in the side car. It will be good for commissioning and for Avtron products because not only you will need to commission the electrical infrastructure. But on top of that, you need to commission heat rejection units, CDUs and so on and so forth. So to summarize the impact it could have on Legrand, it could imply for us the theoretical accessible market, let's say, would be between USD 3 million and USD 4 million per megawatt. Now this being said, I wouldn't like you to get too excited by the opportunity because it won't be at scale before '20 or you won't hit our P&L before, let's say, '28 or '29. And more importantly, this is one amongst many architecture. And you have to understand that we are in a world where you have tens and tens of different architecture. And what is important is not for a company like Legrand is to be architecture agnostic. In other words, to have the ability to work with all the hyperscalers, all -- every single co-locators so that our product launches stick to the architecture that they're going to launch rather than betting that the winning architecture will be X, X or Y. And I have the feeling that we have the right relationship with all of those guys. I mean we are working with Meta, the Google, the Oracle, the Microsoft, the [indiscernible] of the world and the QTS and the Equinix and so on and so forth. So in a nutshell, it should have quite a positive impact on Legrand, but not for now, within 2 or 3 years. And again, it will be one amongst many different architecture. Sorry, I've been a bit long, but I thought it was worth taking some time because those topics are complex topics, and we need to bring as much clarity as we can to the market.
George Featherstone: That's very helpful. Maybe just another question on your guidance for data center growth this year. Previously, you've sort of benchmarked yourself to Vertiv and their growth is quite a bit above what you're saying that yours will be this year. Perhaps could you explain what the difference would be this year for you?
Benoît Coquart: Well, I hope they are right. To make a long story short. But I mean, well, this year, we grew close to 40%, which is significantly higher than their growth, right? Because if my reading was correct, they grew 26%. So we did a fantastic performance. And actually, this plus -- close to plus 40% is probably significantly above the market growth. Well, if the market is not growing 10%, 12%, 14% next -- this year in 2026, but much more than that, then fine. Our objective, as we did last year, is to overperform the market. So if the market is growing 20% instead of growing 10%, all good for Legrand. There's no structural reason why Vertiv should grow faster than Legrand. In '26, we grew 40% against '26. And if you look at the past 2 years, the performance is also significant. So again, we are all different animals in this business. So comparing one with the other might not be the right way to do. What I can confirm is that again, we're going to experience a nice growth in 2026 in data centers. We have the right product offering. We have the ability, should we miss something, we have the ability to develop it organically or to buy it. And I think we have developed a great expertise in buying data center assets at reasonable prices. We have the right relationships with the customers. We've been able to scale our business by adding capacity whenever needed. So we are ready to capture any market growth that will come.
Operator: The next question comes from the line of Phil Buller from JPMorgan.
Philip Buller: Thank you for all of the data center disclosure. Just to try and extract one more data point, if I can. You mentioned $3 million to $4 million per megawatt in a higher density architecture, if I heard that correctly. What is the current megawatt in a current architecture, if you will?
Benoît Coquart: Well, it's probably -- now it's between $2 million and $3 million, but probably closer to $3 million now than to $2 million, given the latest acquisitions we have made. So -- well, between $2 million and $3 million, but closer to $3 million.
Philip Buller: Perfect. And then on the guidance, I understood that we are expecting flat volume in buildings. I think that, that makes sense as a planning assumption. Would you see any signs from the ground that there's an improving situation in end markets such as residential in Europe or U.S. office? Any kind of on-the-ground commentary on some of those key markets would be great, please.
Benoît Coquart: Well, you're right to say that the world shouldn't be limited to data centers. It's worth also having a look at building. Well, we're a bit more optimistic for buildings, we're a bit more optimistic for Europe than for the U.S. Typically, for the U.S., we believe that -- and we have embedded into our guidance a slightly negative building market overall. We see no short-term positive signals on resi. But it's only 15% of our sales in the U.S. As far as non-resi is concerned, we also remain quite cautious and the statistics tend to show that the market should be slightly down. So overall, building in the U.S., slightly negative. Now bear in mind that in the U.S., 40% to 45% of our sales is now represented by data centers. So only slightly more than 50% is represented by building. As far as Europe is concerned, we have embedded something flat to slightly positive if you look at the various KPIs, as far as the resi is concerned, well, completions are moving from minus 9% in 2025 or should move from minus 9% in 2025 to plus 2% in '26. Permit should be slightly up by plus 4% in 2026. Renovation should be slightly up by plus 1%. So we start to see positive indicators that of course, we are late cycle. So those indicators do not immediately translate into Legrand sales, but those are rather positive signs that the market should get better. As far as non-resi is concerned, recent updates from experts also suggest some sort of recovery in 2026 with renovation remaining slightly positive and new build also. So in other words, negative building -- slightly negative building business in the U.S. and flat to slightly positive in Europe. As far as the rest of the world is concerned, what is quite a mixed situation. We don't expect a recovery in China yet on the building side. And Africa, Middle East and India should remain quite supportive.
Philip Buller: That's great. There's no pocket of the business that you're concerned about being in a significant contraction territory.
Benoît Coquart: No. I mean it depends what you call significant contraction. The risk is always a bit China because China has experienced a minus 50% decline in the residential business over the past 3 or 4 years. Now China, it's only 2% of our sales. So whatever happens to the resi market in China won't impact much Legrand numbers. So I don't see any reason why there would be contraction somewhere.
Operator: And we'll proceed with our next question, just moment. And the question comes from the line of Gael de-Bray from Deutsche Bank.
Gael de-Bray: Can I follow up on the 800-volt DC discussion? I wondered how you're addressing this potential shift from a technological standpoint, I mean, especially around the potential change from electromechanical circuit breakers to solid-state circuit breakers. And also still in relation to 800-volt DC, can I -- can you go a bit deeper into the breakdown you provided, I mean, especially around the revenue base you have in the rack PDU segment and what the impact could be on this part of the portfolio going forward?
Benoît Coquart: Of course, again, Gael, there's a misunderstanding between what the 800-volt DC architecture, which is ready for deployment, which is the one with the side car and the sort of full DC holy grail type of architecture, which is not ready for deployment, won't be before 2030, 2031, if it is, and which is a full DC architecture. We are addressing both by 2 ways. Number one, by developing products whenever needed. So for example, we are developing OCP type of -- and we presented well actually 6 months back at the trade show of racks by working on DC busways and a number of other things. And number two, whenever we feel that we have a gap, then we fulfill the gap by partnering or buying companies. Good example being the 2-phase direct-to-chip liquid cooling investment we made in Accelsius, which is not only an investment, financial investments, but which is also a commercial and technological partnership that will give us the ability to sell a very interesting product offering to high-density data centers. So you have to keep in mind that Legrand is the only company in this business, which has built from scratch a product offering which is AI ready. Our competitors were either pure play of data centers ready or ready [indiscernible] and so on and so forth. When we started back in 2017, we were doing sales of EUR 300 million in data centers, of which a few PDUs and a few racks. But it was 9 years ago. Since then, we have built product offering almost from scratch by doing 30 acquisitions by developing organic products, which, again, is very suited to high-density data centers. So I don't have the best answer to tell you. We're going to keep developing products. We will keep working hard with the design teams of our customers to make sure that our products are suitable to their needs. We do a lot of ETO engineering to orders. And whenever needed, we'll partner, we'll invest in partners if needed or we'll buy companies, and it will make Legrand perfectly in good shape to tackle the challenges of the new architecture that are going to come. Now as far as PDUs, I cannot be more precise than I was. I don't want to give you sales by product families. I told you that everything which was related to compute management was about 25% of our sales. And within this 25% of data center sales, you have many things, including rPDUs, but not only rPDUs, you have also monitoring devices. You have keyboard video mouse, you have transceivers, you have the console business that we bought a couple of years back. That's it. And it's -- the PDU business is part of the 20% of our sales that should be negatively impacted by indeed 800-volt architecture. But again, you have many products or families of products that will be positively impacted, 80% of our sales. This is our estimate today.
Gael de-Bray: That's great. Can I also ask about what happened in Q4? I think the outcome in terms of organic growth was certainly a bit higher than what you had anticipated yourself. So what surprised you on the upside? Was it just data center related? Or are you also already seeing residential demand in Europe taking higher here relative to the last time as well.
Benoît Coquart: Yes, it's mostly data center again. Yes, the Q4 is optically better in Europe than the full year, but it also comes from data center actually. Don't forget that data center, it's not only a U.S. business, but it's also Europe and rest of the world. And actually, -- maybe data that I can share with you. I told you that we grew in data center close to 40%. This growth is close to 50% in the U.S. and it's about 20% plus in Europe and 20% in the rest of the world. So we are growing significantly everywhere in data center, even though the growth is stronger in the U.S. and elsewhere. Now to answer -- short answer to your question, we did not anticipate so much sales and actually so much orders in data center in Q4.
Operator: The next question comes from the line of Max Yates from Morgan Stanley.
Max Yates: Just my first question is around your pricing. And when you say that's based on your kind of current assumptions, could you give us a feel for kind of what those current assumptions are? Because obviously, it's difficult with kind of copper prices and silver prices. We know they're quite sort of big drivers of direct raw materials. So could you give us a feel of -- are you doing that with kind of $13,000, $14,000 copper in mind? Are you doing that with current steel prices? Or if we do see raw materials stay at current prices, will that number be quite a bit higher?
Benoît Coquart: Well, Max, it's a fair question because -- but frankly speaking, we have so many different -- we are not dependent upon one single raw mat, copper, silver or something else. And bear in mind that the vast majority of our purchases are components. Out of the 35% of raw mats and components, it's about 10% raw mats and 25% components. So it's a mix of many, many things. So we do it with our purchasing team on a very professional manner. We look at experts, specialists. We embed, of course, productivity into that, and it leads to a central scenario. And based on this central scenario, we do the appropriate pricing. The end of the game would be to adapt. The best analogy I can give you is what we did last year for tariff, U.S. tariff. I remember when we did the same call a year ago, we told you that we have embedded only USD 30 million of tariff into our guidance. But I also told you that should there be more tariff we will do more pricing. And that's what happened. At the end of the year, we had USD 100 million of tariff to compensate for. It's actually the reality is that we had $140 million. We compensated $40 million by optimizing our supply chain, making sure that more products were eligible to the USMCA agreement and so on. And the remaining $100 million of tariff were compensated through pricing in value. So the same story with raw mats and components, we have a central scenario. We might be wrong, we may be right. If copper price was to go even up and then the steel and plastic, oil, components, labor and so on and so forth. And if we needed to do more pricing in order to deliver our profitability target, we will do more pricing. And I mean, we've been demonstrating over the years that we have the ability to do so.
Max Yates: Okay. And just maybe a very quick follow-up on your North America growth rate of 7% in the quarter. I'm really trying or really struggling to understand that because you're sort of saying that data centers was better. If I look at your kind of full year data center number for the group, it feels like you did roughly 30% in the fourth quarter for the group. So U.S. must have been higher than that. You're now saying data centers is sort of 40% of your business in the U.S. I know it wasn't that last year, but your data center business should have been growing -- like that should have been a double-digit contributor to growth. So I'm trying to back out how we get back to 7%.
Benoît Coquart: No, no. Actually, in Q4, our data center grew by about 10%. The reason -- so it seems to be slow. But bear in mind that we had a very, very strong Q4 2024. So as early as July '25, we told you that in H2, you would have an optical deceleration, but which was not a deceleration, which was purely coming from the basis for comparison. So this is the point. I mean, about plus 10% in Q4 on a plus 30% last year, I mean, the year before and full year close to plus 40%.
Max Yates: Okay. So -- but then was your -- is your 40% data center growth this year, is that an organic number? Or is that a...
Benoît Coquart: Yes, close to 40% is for the full year, it's an organic number.
Max Yates: [indiscernible] in the fourth quarter.
Benoît Coquart: Yes. But again, the basis for comparison makes the number a bit tricky because we had a very, very easy Q1 comp because of the great pause, which happened back in Q1 2024. And we had a very, very demanding Q4. Now be careful, don't extrapolate one way or the other, there's no such concept as an exit rate in data center, right? So don't extrapolate good Q4 or slow Q4 into 2026. I can confirm that the performance was great in the U.S. and elsewhere in Q4 that we have a lot of orders that are sustaining our guidance for 2026 and that we should see very exciting growth in data centers this year.
Operator: The next question comes from the line of Kulwinder Rajpal from AlphaValue.
Kulwinder Rajpal: So I also wanted to follow up a little bit on the data center side. So we have been discussing about the 800-volt system, but I wanted to actually dig a little bit into the PUE side of things. I mean I know the demand for standard offerings is high. But are you also seeing a traction for your PUE offerings because we know that Europe is already short on power -- to power all the data centers. And then is there a dedicated part of the portfolio that is dedicated to these high-efficiency offerings? And is there something that you can add through acquisitions also?
Benoît Coquart: It's a good question. Indeed, we estimate that we have approximately 40% of our data center sales, which help or can be used to reduce the energy bill of a data center. And it's not only about compute monitoring. It's also, of course, about efficiency within the powertrain and amongst another -- a number of companies. So today, I think the average PUE on a worldwide basis is probably something like 1.5. We know that theoretically, it can go down -- I mean, the best PUE ever, I think, was recorded was 1.025, if I'm correct. So we have a lot of opportunities to help our customers cutting their PUE from 1.5, 1.6 down to 1.4, 1.3, 1.2 if needed. So it is clearly a very important driver behind our data center business, and it should continue to be. Now this being said, it is a fact that you have geographies where access to grid is a constraint. And we -- there are some countries where it can take up to 2, 3, 4 years for a data center to get connected to the grid. That's why you have to take with a certain cautiousness, the numbers, the CapEx numbers, which are disclosed by our customers because some of those CapEx will not be spent in 2026 or not even in 2027 just because they will need to get the access to the grid. So to make a long story short, I confirm that it is an industry challenge to get the energy. It may translate into some data centers being opened rather in '28 than in '26. But we are part of the solution, not of the problem because a large part of our product portfolio help cutting down the PUE.
Kulwinder Rajpal: Right. And so just to follow up a little bit on this. So is there a specific treatment that you get in terms of pricing with the portfolio and also if it helps the profitability? I know that most of the products are centered around the group profitability, but I just wanted to understand if there is a specific advantage that you can get from this particular set of products.
Benoît Coquart: No. I mean we -- I'm not aware of any significant pricing approach between data center and building. Of course, our products in data center are mission-critical. They are extremely important for our customers to deliver their performance, not only in terms of PUE, but also in terms of reliability, compute performance and so on and so forth. But at the same time, there are big customers are negotiating tough on price, and they want to make sure that they have good value for money. So we are not taking advantage of product shortage or the criticality of our products to do additional pricing. We are doing the same pricing. We are doing elsewhere.
Operator: The next question comes from the line of Eric Lemarie from CIC CIB.
Eric Lemarié: My first question on the construction cycle. You mentioned that Legrand is more late cycle, and it's certainly very true for the new build. But is it really the case for renovation? Because I suspect that if I need to renovate my house, I will probably start with some electrical equipment.
Benoît Coquart: No, you are right, Eric. Indeed, for new, you have to -- it depends on the building, but it could be 6, 12 or 18 months lag between the time a permit is issued and the time we sell our products. For renovation, the shorter -- the cycles are a lot shorter even though it depends on the type of renovation. If it is renovating one room, it's almost immediate. If it is renovating a full house, then it takes a full -- a few months now. This being said, the renovation numbers for Europe in 2026 are not very bullish. According to, I think it's your construct. They plan for the residential renovation to be up 1%. So it's a very light increase, let's say. Again, we can have good surprises. We'll see. But so far, nobody expects the renovation market in Europe to rebound sharply. That's not what we have embedded in our guidance. We have embedded, as I said, flat to slightly positive building market in Europe.
Eric Lemarié: And a follow-up on data center. You mentioned this close to 40% growth for the full year in 2025 and 10% for Q4. Could you remind us Q1, Q2, Q3, how was it the growth on an organic basis for data centers for Legrand?
Benoît Coquart: Yes. So we said that it was well above 30% in Q1, well above 30% in H1. So I'll let you, Eric, do your own computation, above 30% in 9 months and 10% in Q4. And that everything was coming from the basis for comparison and that we have kept building a very nice order book over the year. And again, we have a very strong backlog and order book at the end of 2025, which give us full confidence in our ability to deliver our data center targets for '26. And maybe just a word because I want you to avoid a misunderstanding. So the close to plus 40% is really organic. If you were to put together FX and acquisitions, this close to plus 40% would become plus 50%. So the close to plus 40% is really purely like-for-like sales increase in data centers in 2025.
Eric Lemarié: Okay. But the close -- well above 30% you mentioned for the 9 months, it was actually well above 40%, I suspect?
Benoît Coquart: 40% is above 30%.
Eric Lemarié: And just if I may...
Benoît Coquart: I'm glad to give you more disclosure, Eric, but you should have the same level of disclosure to all companies. I have the feeling that we sometimes disclose a lot more than anybody else in data centers.
Eric Lemarié: This is certainly true. And if I may, a last one on margin. Your guidance on margin, do you include in your margin guidance for 2026, some positive impact from acquisition or negative impact from acquisition?
Benoît Coquart: Well, actually -- so first comment, when looking at our guidance, we start from a high base, which is a 2025 margin. And we basically include a bit of leverage, organic leverage, not a lot, I have to admit. But I have to say that we have given a clear priority to growth in 2026 and to sustain growth while you have inefficiencies, you have amortization, you have expenses to do. So a bit of leverage and a few 10 bps of dilution coming from acquisitions. It could be minus 10, it could be minus 20. Those are the order of magnitude. And all that leads to the 20.5% to 21% adjusted EBIT margin after acquisitions, which we have embedded into our guidance. And I'm sure that you have also noticed, Eric, that we have upgraded a bit our 2030 EBIT margin target because at the CMD, we said that we were looking for an average EBIT margin of about 20%. And now we make it clear that the average will be above 20%.
Operator: The next question comes from the line of Martin Wilkie from Citi.
Martin Wilkie: It's Martin from Citi. I don't want to labor the point too much, but just to come back to the data center growth in the fourth quarter, and I appreciate there are obviously some comp effects, both from the very high growth in Q4 and also because of some of the acquisition effects as well. Could you give us -- you've obviously given the EUR 2.4 billion in absolute terms for the year. But in euro terms for Q4, are we right in thinking that data center sales in euro terms were sort of close to EUR 700 million? Just so we can make sure we sort of square the circle in terms of how big data center was in the fourth quarter.
Benoît Coquart: Well, actually, I don't have this level of granularity. And it's a bit complicated because you have to embed FX, which is strongly negative in Q4 because of the dollar versus euro. You have to embed acquisitions and -- so it's a bit complicated to say. But it seems like you are surprised by the plus 10% in Q4, which is far better than what we guided for back in November for the last. And again, you shouldn't look at the data center business as if it was a distributed business, right? One quarter is impacted by the comps by the project and so on. If your question is, is your target plus 10% in data center in 2026? The answer is no. Our guidance is 10% to 20%. But of course, we are targeting to grow as fast as possible. So I wouldn't read the plus 10% in Q4 as any indication of what it would be for 2026. What I can tell you, again, and this is confirmed by the publication of our peers, this is confirmed by the huge CapEx plan from hyperscalers where the CapEx is growing from 50% to 100% between '25 and '26. It is confirmed by industry experts. It is confirmed by the backlog we have. It is confirmed by the book which we have. 2026 is going to be another very good year in data centers.
Martin Wilkie: We obviously see the strength of the orders. I think the debate or the confusion with the Q4 numbers is that the full year seems to be a lot better than people had expected. And therefore, we would have thought the Q4 growth rate would have been higher. But I'll go through the math afterwards. I know there's a lot of moving parts on acquisitions and foreign exchange and these kind of things. Perhaps if I could just have one follow-up on data center. Obviously, you started off the year with a lower guidance. When the surprise comes, is it literally sort of an overnight surprise to you? Just -- I mean, I know obviously, there's a difference between backlog and pipeline. But in terms of when you're having those conversations with your customers, what sort of pipeline visibility do you have?
Benoît Coquart: Well, it's very complicated because, again, we are not as experts as others in reading the pipeline. But again, the connection between backlog and sales it's not that easy because orders are -- can be canceled, they can be moved. If we have to -- most of the contracts do have a pricing clause whereby you can adjust the price up or down depending on the price of raw material and components. The orders we have are usually not 3 years orders. We have an ability -- our lead times are typically 8, 10 or 12 weeks for most of our products. We have almost no products where you have a longer lead time or lead time as high as 8 months, 10 months or 12 months. So customers do not need to pass orders 1.5 years in advance. So most of the backlog we have will have to be delivered in 2026, not in '27 and '28. So it's not solid enough as a leading indicator to tell you precisely, yes, we intend to grow 18.5% in data centers in 2026 because the backlog would support that kind of growth. It's more a trend topic. And this trend topic, again, confirmed the very good '26. I cannot be more precise than that. So in other words, to make the long story short, number one, we shouldn't try to read too much from the backlog even though the numbers are very good. Number two, having a backlog of orders in hand is not a problem to pass on price increase if we needed to.
Operator: Now we're going to take our next question and the question comes from the line of Alasdair Leslie from Bernstein.
Alasdair Leslie: So a couple of follow-up ones on data centers, please. I mean, obviously, it sounds like you saw a similar surge in data center demand to your peers. I kind of -- I appreciate the comments about not overstating the importance of the backlog, but you obviously do talk about a promising order book there. Does that surge in Q4 demand, does that really give you a sort of fast start to 2026 as well? I was just wondering what the outlook for Q1 data center growth was. Obviously, just last year, reflecting on 2025, it can be a little bit lumpy from one quarter to another. So just to help us kind of calibrate expectations for the first quarter.
Benoît Coquart: Well, we're not guiding on quarterly sales or profit, neither on data centers nor for the rest. And again, an exit rate doesn't mean much in the data center business. So no specific guidance to give you, Alasdair, for Q1. We stick to our yearly guidance.
Alasdair Leslie: Okay. And then maybe just the second question was on capacity to meet higher demand in data centers. I mean we hear commentary, it sounds like constraints are creeping back in and on the rise again, obviously, because of the surge in demand. But one of your slides mentioned solid capacity to execute and adapt. I appreciate that's probably a broader level across the group. But -- and it feels like the ability to meet demand, short lead times, that's still very much a kind of competitive advantage right now in the industry. So just wondering if you could comment there in terms of how you assess yourselves relative to the competition on industry.
Benoît Coquart: Yes. So far, the teams have done a very good job, I have to say. So we have doubled our capacity investment on data center in '24 compared to '23 and doubled again in '25 compared to '24, still remaining actually within the 3% to 3.5% CapEx to sales level. So we are not increasing the CapEx guidance. And we believe we will still do a good job. So yes, it's a challenge, and the teams are working hard to meet the demand. But this is a business which is not capital intensive. So you can increase capacity by working on your supply chain, by working with a subcontractor, by adding shifts. You don't have to spend millions and millions in CapEx. So it remains a challenge. We've been able to do a good job in the past 2 or 3 years, and I'm fully confident that it will not be a bottleneck for our business in '26.
Operator: Now we'll proceed with our next question, and it comes from the line of Andre Kukhnin from UBS.
Andre Kukhnin: Can I just clarify a couple of things first, and then I have one question. On -- and sorry to come back to 800-volt DC. But in terms of -- when you talked about AC powertrain being neutral to positive within the 80% of your business, do you see that as in dollars per megawatt or just in absolute terms, given there's going to be a lot more megawatts when we go to that architecture?
Benoît Coquart: Well, it's also dollar per megawatt but again, I'm ready to have the discussion with your experts if you wish to. But yes, we see much more solid redundant with more intensity AC powertrain. So it should contribute to the higher -- slightly higher dollar per megawatt. And on top of that, you will have gigawatt data centers and not megawatt data centers. So for data centers as a whole, it's also a good news.
Andre Kukhnin: Yes, we have no doubt in the growth in megawatts or gigawatts in absolute. It's just a couple of people we spoke to basically suggest that there's at least one stage of switching that disappears in the DC 800V architecture, and that's the AC switching. So I was just intrigued to hear that you see that...
Benoît Coquart: So you have a bit more -- you have less UPS, which is true. But we are not -- we have a very small market share in UPS in data centers, I have to say. But again, it depends which DC architecture you're talking to. If it is the next grid to chip, then it's about DC and no longer AC. But again, this architecture is in the books today. And yes -- so in the architecture, which is currently considered for '28, '29 in some of the data centers, it's still an AC powertrain.
Andre Kukhnin: Got it. And just related to that...
Benoît Coquart: There's a focus on this new architecture, which again is a pretty good news for Legrand, not a bad news, which is too strong from the financial community. You are missing, I think, one point. which is the fact that, number one, many architectures will continue. And even if this one has a meaningful market share, this market share is going to be 10%, 12%, 15%, not 100%. And it will probably be made of different type of sub-architecture. So many architecture will coexist. If you take one hyperscaler, if you talk to one hyperscaler, I will tell you that over the past 10 or 12 years, they probably have 6, 7, 8 different architectures by hyperscaler. So in total, you have 10 different architecture coexisting. There's not one that's going to prevail in the years to come, number one. Number two, you consider a company such as Legrand as static animals, which do have a product portfolio, which we're not able to adapt and to adjust. Again, look at what Legrand did over the past 8 years. So if there's something new coming, in the architecture, if there's one piece missing that we don't have, we will develop it, we'll partner to get it or we will buy it by [indiscernible] company. It shouldn't be such a concern. So it's interesting to see that 1/4 of the questions on this call were on 800-volt DC. I think there's too much emphasis putting on that topic.
Andre Kukhnin: Fair enough. I completely take it. I just -- on that kind of nonstatic animal, I guess that's what you're saying on the solid-state switching and braking that you do not have it right now, but you're confident you will develop it or be able to buy it.
Benoît Coquart: Well, I don't want to be specific on one product family or the other. But again, if we feel that there's something that we absolutely need to have, we will have it. But most importantly, the current Legrand portfolio can address 99% of the needs for the next -- at least the next 5 years. That's a very important message that I want to channel to you. Now what will come in 6 or 7 years, it's a different story, but we will adapt. We will adapt. And if we are ready to tackle only 70% or 80% of the architecture that will come in 8 years, but we will do what it takes to address the remaining 20%, and that's it. But we have a product offering, which is suitable for 99% of the architecture that will come in the next 4, 5 years.
Andre Kukhnin: Can I just ask on that change to the margin ambition for 2030 from around 20% to above 20%. I guess we learned on this call, above 30% can be 45%. So just wanted to understand whether that above 20% is an ambition to continuously improve margin from here? Or is it kind of, hey, the margin is above 20% now and we are kind of okay with it now?
Benoît Coquart: No, we are not guiding more precisely than that because, of course, it depends on the acquisitions we're going to do. I mean, last year, our acquisitions were accretive, but they could very much be dilutive by 30, 40 bps. It depends on the growth rate and so on and so forth. So I don't want to shoot a number. If you look at the past 6 years, -- so '21 to '25, we've been consistently above 20%. And the average of the 5 years, if I'm correct, is 20.6%. I'm not saying that this is a new standard or new benchmark, but it means that it could be 20.2%, it could be 20.5%, it could be 21%. It will not be 35%, if this is your question. So no, we're not shooting a new target, just that it's going to be above 20%.
Operator: The next question comes from the line of Ben Uglow from Oxcap Analytics.
Benedict Uglow: I had a couple. I think a previous question assumed or sort of said you've had an order surge. Maybe I missed it in your opening remarks. But can you just give us a sense of your order development, obviously, in data centers in the fourth quarter? And the reason why we ask is your phasing and your time line in projects can be a little bit different from others. If I look at Eaton, Vertiv others that have reported, they've seen a kind of almost doubling of their orders between the third and fourth quarters and up by 200% plus year-over-year. I guess my question is, have you seen -- I don't want a specific number, but are you seeing exactly that kind of trend qualitatively? And when I think about the phasing of your growth in the current year, is it correct to assume that, that growth, whether it's 10% to 20% or 30% is back-end loaded? I guess what I'm thinking about is how quickly we see any orders come through in the next 6 months.
Benoît Coquart: Ben, well, I'm a bit embarrassed because -- we have seen some order flowing, some backlog building, but I don't want to shoot a number because, again, I don't believe that it will help in any way to forecast what the 2026 sales is going to be. And actually, when we look at our competitors, I don't know, yes, ABB shot an increase in backlog, but they are guiding for growth in data center, which is in the teens. Vertiv, if I correct, is saying that you shouldn't extrapolate anymore the orders and that they will not give it anymore on quarter-by-quarter basis. So I think everybody is more in line to tell you, be careful. You cannot extrapolate an order inflow, a backlog or an order book into the next 12 months sales. Also more for Legrand, as again, we don't have the same order pattern as an ABB, Eaton, Schneider or Vertiv. We don't have orders, and we've never had in data centers orders above a year. 95% of our orders should be in even '26. So it's the very nature of our business, the fact that we have short lead time, maybe the products we are in, I don't know, maybe the geographies we are in, that makes it -- that makes me a bit uncomfortable to extrapolate the orders we have into sales. So I know you don't like the answer, but unfortunately, I cannot give a better answer than that.
Benedict Uglow: No, understood. And by the way, I appreciate all the disclosure. It goes around and around in circles, but I think we're all trying to get to the same thing. The second kind of question is just around North America on the margin side putting 4Q to one side, it's a pretty healthy evolution year-over-year, but we do have -- still have moving parts in terms of raw material tariffs, et cetera. Is there any reason for us to think about the drop-through or the potential growth in North American margin, obviously, given your top line. Is there any reason to think about it differently this year than last year, i.e., that all else equal, we should be seeing some decent drop-through to your margin? Or are there any qualifying effects?
Benoît Coquart: Well, I will let Franck to take this one. Go ahead, Franck.
Franck Lemery: Yes. Ben. Well, as you noted, effectively, the margin on Q4 alone for North and South America is a little bit weak. There is absolutely nothing sustainable. There are many moving pieces in that margin with the mix of business with new acquisition with some one-timers. So what I would -- the way I'm reading the performance is more looking at H2 with the gross margin around 50%, with adjusted EBIT around 20%, which makes North and Central America very profitable. And on a year-on-year basis, the margin is improving a lot despite the tariff challenge that we already shared. And second, as you rightly said interestingly, it's the value, the year-on-year growth, 24% of improvement in value of the adjusted EBIT margin between '25 and '24, and that's in euros, it's close to 30%. So bottom line, a very good performance of our U.S. colleagues.
Operator: Now we take our next question and the question comes from the line of William Mackie from Kepler Cheuvreux.
William Mackie: I have 2 questions, one relating to the structure of your guidance. One, if I can harness your continued generosity on the data center discussion to talk about the definition of the business. So firstly, with data -- you've talked a lot about the profile of the business here and the growth rates. Could you just touch on the customer profile? We've seen a lot of growth across an announcements from hyperscalers, but there's a broad base of customers. So how is your customer footprint positioned between the large scale and the broader cloud providers and other providers? And how is the growth trends differing between the different DC type customers?
Benoît Coquart: Yes. Well, we are a mix of customers that more or less replicate the market that the feeling we have. So it's probably -- it's not always easy to know them, but it's probably 1/3, a big 1/3 to half on the hyperscaler and then cloud 20%, 25% and then on-premise, maybe 25%, 30%. Those are the orders of magnitude. But if you take the market, there's no reason why our sales wouldn't replicate the market. Of course, the hyperscalers have been the one investing the most. So we are growing very nicely on hyperscalers. Now the growth is also very good on co-locators, either serving hyperscalers or retail co-locators. The one segment of the market, which is not growing at the same pace is clearly on-premise data centers, which have a much slower growth rate. But when it comes to cloud, new cloud, hyperscalers, [indiscernible] retail or, those are growing very nicely.
William Mackie: And then maybe moving across to the segmentation detail that you provide annually on Slide 15. Could you talk a little about your expectations going forward for the growth rates across the energy transition category? We've talked about buildings, which largely fall for essential infrastructure.
Benoît Coquart: Yes. Well, let's maybe a word on '25. So on '25, I told you that the data center grew close to 40%, which implies that the rest is slightly growing only. And out of the rest, you have energy transition growing a little bit more. You have digital lifestyle down with Connected Health growing, but Connected Home being down, and this is a result of the housing market in Europe. And essential infrastructure is basically flat. So plus for energy transition, flat for essential, slightly down for digital lifestyle. When it comes to 2026, it's difficult to be very precise because it will depend on the underlying markets. Now I see no reason why energy transition shouldn't do better than essential because it is boosted by structural trend. The fact that the world is electrifying and moving from fossil energies to electricity is a structural trend that is here to last. So it should do a bit better than essential, but by a few points, not by 10 or 15 points. This is our central scenario. A word -- I'm not sure it was completely clear on the press release, which is very interesting. When we buy companies related to the critical power into data centers, it's Linkk company we bought in Malaysia. It's Avtron in the U.S. it's Curtis Industries in the U.S. and it's a few others. Usually, they're not doing 100% of their sales in data center. They are doing 50%, 60%, 70% of their sales in data center and the rest of their sales is made in microgrids, in infra, in industries and so on and so forth. So it helps us building an energy transition footprint in verticals in which we were not. We already had critical power in education, commercial buildings, office buildings, but we did not have critical power into microgrid or industries. So it's a sort of side effect of our acquisitions in data centers, not only builds our position in data centers. But on top of that, it also reinforces and build our position in energy transition.
William Mackie: Maybe a last final follow-up relating to prospects. We've talked broadly about Europe, but I think Rexel last night printed 3.3% growth in France. I know they win market share. What are your thoughts about some of the major European companies -- countries and particularly France?
Benoît Coquart: For 2026, well, number one, we are not growing in France but we're not losing market share. And clearly, Rexel has been massively gaining market share in France for quite some time. So it's a tribute to the teams. For '26, we start to be somehow a bit more positive than we were on France. I'm sure you have heard about this [indiscernible], this housing plan, the so-called Bazooka plan, which the French authorities have stated that they intend to build 400,000 houses a year, which would be a surge compared to the current 270,000. Now we are a bit cautious because so far, it's an announcement. We don't know yet the specifics about this plan, how will it be financed? What will be the measures that will be implemented in order to support this plan. But at least, it signals a change of mood in France and the fact that now housing is considered again as a priority where it was not. So we are slightly more positive on the mood at least. We'll see about the numbers. When it comes to Germany, the Bazooka plan should start to have a bit of impact. Southern Europe has been pretty healthy. We did nice growth last year in Italy, for example. Spain is okay. So again, I don't want to sound too optimistic because we've been waiting for the rebound for 2 years, and it did not really happen yet. But again, the signals start to be a bit more positive. Now before starting to upload, we have to wait for the construction KPIs to flow into our numbers, which is not yet the case.
Operator: Dear speakers, there are no further questions for today. I would now like to hand the conference over to your speaker, Benoit Coquart, for any closing remarks.
Benoît Coquart: Well, I just wanted to thank you for your interest in Legrand. Thank you for the clarity of your questions. And should you have more questions and not only on the 800-volt DC, do not hesitate to call the IR team. We'll be happy to answer. Thanks a lot.
Operator: This concludes today's conference call. Thank you for participating. You may now all disconnect. Have a nice day.