Trane Technologies plc ($TT)

Hi, I'm Andrew Obin, BofA's multi-industrial analyst. And with us, we have management of Trane Technologies. We have Dave Regnery, our Chairman and CEO; and Chris Kuehn, Executive Vice President and CFO. I guess we'll just jump into the fireside, right?

Great. Thanks, everyone, for coming today. Appreciate it.

Always a pleasure to have you here. So maybe we can start with tariffs and inflation now more visible than 90 days ago. Where do you see pressure points on price cost as we head into the second half?

Chris, I'll let you start?

Yes, sure. As we said on our call a couple of weeks ago, it is a bit more inflationary from a tariff perspective and from a raw material input perspective than when we started the year in January. We think we had that capturing our guide. We raised our full year guide to 7% revenue growth and EPS growth, 13% to 15%. But look, we're going to leverage our business operating system. Fortunately, we've had to get really good at this over the last 5, 6 years from inflation and our tariffs and looking to mitigation first and then ultimately, if we need to, to deploy pricing. But I think, Andrew, from a relative perspective as a company that has been in region for region for well over a decade, we have 21 factories that serve our North America business and 20 of those factories are in the U.S. So over 95% of what we sell in the U.S. is either manufactured and/or assembled in the United States. So we like having that framework. But I said it's a little bit more inflationary based on some movement in tariff. It's a dynamic area, but we think we have that captured in the guide for the year. Maybe some near-term pressure on price versus inflation. But over the medium to long term, we'll make sure that we're driving to the spread.

And I think the industry announced a number of price increases in early May. Would those be normal? Or would those be sort of to reflect inflationary pressures?

I can't speak for the entire industry. I'll just say for us, what we did is we announced the price increase in February. It was effective April 1. We have that baked into the guide when we had our call just a couple of weeks ago. And again, it's a very dynamic environment in terms of cost, and we'll keep taking those inputs. We'll understand what we can do to mitigate with suppliers. And then one of the areas we'll have to -- may have to use as pricing to have an offset, but we're going to take all those inputs. And I don't want to get in front of our teams in terms of what they're thinking about.

Yes. Of, course, of course.

Yes. I mean just to reiterate a little bit what Chris said here. Look, we've been -- our business operating system, we've proven how we could stay ahead of inflation. And we had a lot of practice for the last, I'll say, 5 years. But we're very comfortable we'll stay ahead of it this time as well.

So maybe we can talk about Commercial HVAC. So applied HVAC bookings were up over 160% in the quarter. How much of that strength is data center driven versus traditional verticals like health care, education and government?

Yes. Look, our applied bookings have been up over 100% for 3 quarters in a row. So they're up 160% the last quarter, but they've been up over 100% for the last 3 quarters. And data centers are very strong, okay? But I would also tell you that we're strong in many other verticals as well. Health care, you mentioned for one. Certainly, higher end is another one that we've been very strong in. Tech, industrial, we've been strong in. Government, we've been very strong in, whether it be municipal or state or federal. Look, we're the company out there that doesn't talk about the size of our business in data centers. And just to be fair, I'll explain to you the why, right? We've made the investment in a direct sales force and we've had that investment for decades now, and it's a very expensive model to go to a market in. However, when you have a direct channel, you have a very strong channel. And we're not going to disclose the size of any vertical, whether it be data centers or health care or you pick another vertical just because we don't need to be able to tell others that have not made that investment where we see opportunities. And it's that reason why. But I would tell you that we're very strong in data centers. We've been very strong in data centers for decades and will be very strong well into the future. But you should also know that this 100%-plus growth that we've seen in applied over the last 3 quarters is more than just data centers. And if you think about it, 95% plus of all of our account managers or our sales force does not call on data centers. So there's a reason why we have strength that's broad-based.

Right. And it seemed that's compensated for -- regionally for the entire portfolio.

Absolutely. And there -- most of our account managers or sales force are around 100% commissioned. So they're going to go out and become experts in particular verticals or in particular geographic areas so they could help serve our customers.

So maybe just talk about pricing in applied HVAC, I assume it's a price cost positive?

We haven't done that or provided an info on an individual P&L or product line. I would say, for the enterprise in the first quarter, we had a little over 1.5 points of price. We've guided the year. It's probably approaching 2 points of price. But when I think about Applied Systems, you are not only thinking about the equipment for day 1. We're thinking about the service tail that comes with that over a many year journey. And if we'd stick with data centers, for example, that service growth is really in front of us. So what you're seeing with converting to revenues is really equipment. And you can see, I think, strong performance in our Americas business and our European business in terms of margins. So we like the margins on the equipment. We like making sure that we have that customer for life because it isn't just one application. It's a journey in terms of how they are thinking about their footprint and whatever vertical they're in. Build out capital expenditure plans for a customer for the next 5 years and get through all of their assets and build with them a journey. That's how we think about customers for life. So I think it's there.

And just as applied is growing faster now than unitary, how strong, how should we think about incremental margins, particularly once Stellar and LiquidStack are fully integrated and operating at scale.

Yes. Look, we like the 25% plus framework that we start generally every year with. And we do that because the pipeline of investments that we have is very, very strong. And we like ensuring investors that we have the minimums we're going to deliver. And if we can do better than that, we will. Last couple of years, we've done a little better than that on the incrementals. But that's where our guide was for this year and where we're holding to 25% plus. Let's see what happens by the time we get to July, we're halfway through the year. And obviously, at that point, well within what we think for the balance of the year will be. But you mentioned Stellar, you mentioned LiquidStack. I think we should talk a little bit about Stellar. It's a great acquisition. They have some very modest EPS accretion this year. We said about $0.03 of accretion in year 1. But we see that business in the next 2 to 3 years, growing from $350 million of revenue that they generated last year, okay, prior to our ownership to $1 billion of revenue, we think, in the next 2 to 3 years with mid-teens EBITDA with a plus side after that.

Yes. The other thing about Stellar is it's serving data centers today 100%, and it's modular chiller plants. So think of it as a chiller pumping station as well as the electrical infrastructure that would get assembled on a job site. So it's basically taking a lot of the skilled labor that's required on a job site and moving it to your factory where it's very controlled, and it really speeds up the installation process as well as eliminating what they would call stick build on the job site. Well, if you think about the scarcity of skilled labor, it's not unique to data centers. So we're seeing this in many verticals. So we believe that in the future, Stellar will be serving more than just a data center vertical. This is very applicable in all of our vertical markets that we serve today. So we're really excited about Stellar. It's only been about 60 days now, maybe 65 or so, but off to a great start, and we're excited about what the future is. Think about it. This is a business that's got a $1 billion backlog. Half of it will ship this year. And in 2 to 3 years, this will be a $1 billion business at mid-teens EBITDA.

And think about the investments. So a business that needs to ramp and scale, we know how to do that very well. And so think about our business operating system being deployed to the business, learning from them, what's worked for them, but also at the same time, let's think about procurement. Let's think about factory flow. There's a factory and set of factories we acquired in Florida. We're actually expanding given the backlog that we see in the pipelines. We're expanding to a new facility in Texas. So as I mentioned my numbers earlier of how many factories we have in the U.S., we should be adding 2 to that later this year in terms of our investments. But it will take some time. We're going to have investments this year. They'll probably carry over into 2027. But let's make sure we've got a nice scalable business here, and I think we know exactly how to do that.

Can we just sort of talk about lead times. And I think they are extending at the customer decision level. Does that give you better visibility or raise risk around lumpiness and execution?

Yes. I think I tried to answer this on our earnings call. I think I got a lot of comments that maybe I didn't do a great answer. But look, lead times, we have published lead times or what we would publish to a contractor. And think of it as some of our products, the lead time could be weeks. Some of our products could be up to 6 months. Think of a centrifugal chiller. That's the lead time. Now over time, lead time is actually contracted, right? So we've actually improved a lot. In fact, for the first time in a long time, we have quick ship programs that we're able to meet our customers' needs should they have an emergency. So that's lead time. The second question is, what are customers asking for, right? So a customer would say, "I could give it to you in 6 months." But the customer will say, "I don't want it in 6 months, I want it in 12 months, I want it in 18 months." So they're giving us a lot more visibility to what they need, specifically in the data center vertical. So they're out there saying that here's my plan for the next 12 months or, in some cases, even 18 months as to making sure that we have plenty line of sight to what their needs will be. As far as risk factor, I don't really see a risk factor. We only put something in our backlog when we have a signed PO. So it's a signed PO. Our backlog of $10.7 billion at the end of the first quarter is all equipment. We do not put service in our backlog. So it's just equipment and there's all signed POs. And once there's a signed PO, you could assume, especially in the data center vertical that there's a green light to start construction and power will be supplied to that data center.

And from that perspective, I mean, were you been hearing anything in the channel some concerns about just delays in the field with the EPCs? Are you seeing that?

Well, I mean, I think that in the data center space, they're pretty aggressive build cycles. So we haven't necessarily seen any, what I would call, delays that I would say are chronic, but there could be delay on any job cycle. We haven't seen that become a trend yet. But I would tell you that there's very aggressive build cycles in data centers. Still, that progress. But I don't see that becoming a cancellation. I see that being a pushout, right? Once you have the approval and you have the power source, that data center is going to be built. And we won't get a PO until that clearance has happened. So I don't -- I see it maybe there could be delays, but I don't see any cancellations.

Right. But these delays, is that enough to create volatility quarter-to-quarter versus your plan?

It could, but I wouldn't see these things being -- these could be, I don't know, 4-week, 6-week, 8-week delays. But eventually, that will start to lap itself. So your backlog will just move out. I don't see it as a major risk going forward.

Got you. I appreciate it. How sort of to -- as you go into data centers, right, you've extended your data center toolkit with Stellar and LiquidStack. So how do you think about sort of growing your presence inside data center in terms of product, systems, life cycle services? And another thing I sort of want to add, there was an interview with Jensen and he sort of said that he wants to have partners in the room and sort of to be able to design next-generation of product and effectively sort of saying that specific product capabilities become less relevant, what becomes more relevant is to have the total system capability, design, install, commission, service. How do you maintain the seat at the table with NVIDIA and what capabilities do you need to sort of to grow your presence there?

Yes. I would say, well, we were in the room, okay, number one. So that's a good thing. I'm glad you brought that up. We are part of the reference design. In fact, it's actually published on NVIDIA's website. We've had 2 different reference design. We did one about, I guess, 8 months ago, we did another one that was just released and published probably about 2 months ago. And he's right. You got to think of it at a system level. It's not a component, it's out a system, right? This is the same philosophy we've had in buildings forever, right? This is how we go to market. With the breadth of our portfolio, we're able to think of things at a system level, right? It's not a particular component within the system, it could be the greatest efficiency you want. But if the whole system isn't delivering that efficiency, you're at a loss. And that's where Jensen is getting it with, I'm assuming, by making that statement. So we're really good at the thermal management system within the data center because we have many of the components and/or we have a partner that will have another -- some of those components, especially at the cold plate, but we know how to interface with it. And if you think of it at a system level, you're going to open your mind as to what the possibilities are and how efficient to make that particular data center thermal management system. But this isn't new to us. This is how we win in our core verticals, right? It's funny because when I was younger in my career, I was always used to have regulations about components within a product like the compressor has to be at a certain efficiency level. And I was always thinking to myself, what the heck does that matter about how efficient the compressor is if the entire product isn't efficient. And it's now -- so that everyone's gotten smarter on how they regulate products now, and it's still a longer at a component level. It's now at the product level. But you need to take the product and think of it as a system level. And that's what we do in not only in data centers, but all the verticals today. And there's a massive service opportunity, by the way. You were talking about how to think about it at a system level, a massive service opportunity that's in front of us in data centers.

Some of your competitors have published data that sort of indicates that perhaps service opportunity related to data center is greater than regular. Would you care to comment?

I would say, it's probably the same. These are applied systems that are very complex. I think that data center customers are very risk-averse, so they want to make sure the OEM is doing the service work. So maybe that's what they're referring to, but we're connected to all of our applied systems coming out. So it's not -- for us, it's probably the...

Right. So maybe for some, but for you?

I mean, I think one of the other points you touched on, Andrew, just to go back to it is this commissioning capability. This is a competitive advantage that we have. It's not being able just to be able to supply the components to the system in a data center, it's being able to make sure they're going to operate and you commission them to make sure they're operating the right way. And with over almost 7,000, 6,500 global technicians. We have a workforce that's geared to do this. And we had a major colo come to our operations in Davidson, and I was telling them about how we commission and how we train technicians to commission data centers. And I guess they came to verify, trust but verify. And they came in and we took them through our training center, which is now state-of-the-art and they were just like, "Oh my gosh, you do all this." And there was actually a class going on with technicians about how to commission in a data center. And it's just the depth that we have, the knowledge that we have, the ability to invest in the technicians so they become the best possible. This particular colo gave us 100% of their order, which is somewhat unheard of in the data center space. But it was all based on our ability to make sure that we can commission the products.

Sure. Maybe talk a little bit of M&A. I think M&A related to data center cooling has been picking up. How do you think about your strategic positioning? And what would it take for you to consider a larger deal in this space?

Well, I mean as a major HVACR player, we're going to get a chance to see, obviously, everything on a global basis. I say everything, but the majority. So we're going to take a look at it. Obviously, we will evaluate it. Look, we did acquire Stellar. We did acquire LiquidStack. Those are in that space. We always want to make sure that we want to add value to our shareholders. At the end of the day, that's one of my and Chris' job is to make sure that any acquisition is going to be accretive to our business in the long term. So you always have to be careful about what you're paying for these businesses to make sure that you're getting the proper returns. And I would tell you that we've been very disciplined. I think you've all seen that over the last several years that we're a very disciplined orchestrated company here that's very -- has added a lot of value with the M&A that we've been able to do.

And just another sort of topic, fairly controversial topic, we're getting a lot of incoming questions about the evolving...

You'd like to give me the controversial.

Because you guys do well. Evolving architecture of the data centers, sort of more dry cooling, maybe less chiller content, absorption cool -- absorbing cooler chillers. A, how do you see the technology evolving? And second question, which I think is sort of more important. What do you think it does to thermal content per megawatt and service opportunity, right? Because it goes back to the sort of extreme designed thing, like deploying solutions will evolve, but I think a more relevant number, I think, is content per megawatt, your ability to capture it and how much service does it generate. So 2 ways of looking that.

Yes. I mean at the end of the day, we're part of reference designs, which is some of the terminology that you use or the data centers of the future, some hyperscalers will use that terminology. And we are designing data centers. Think of it as what's the data center going to look like in 18 months or 24 months. And there are different technologies that are being deployed in those developments. I guess, a general rule, I would say is think of chillers being smarter and more integrated. And as the temperature, as the chips can have a higher water temperature to be cooled at, the chillers will be smarter as to how they remove the heat that's there. Think about it as if the water is hitting the cold plate at a certain level, right, whatever that level is, it's leaving at a temperature that's much higher, right? Think of it as 20 to 25 degrees up higher. That's the heat that needs to be removed. So regardless of where you started, if it's 45 C, it's still going to have to be removed. And how you remove that is why I say chillers will be smarter. So think of it as a chiller farm being able to understand what mode that chiller farm should operate in. So should I be in a free cooling mode, okay, depending on the ambient air, maybe should I be in a vapor compression mode, maybe, maybe not all the chillers are operating in a particular mode, maybe only some of them are, and you're worried more about the leaving water temperature out of that chiller farm or segments of that chiller farm. Those are all things that are evolving, that we're working on. I haven't seen any reference designs or data center of the future that does not have a chiller in it, okay? All those chillers are operating. Yes, I've seen lots of different variations on that.

But how do we think about [ dollar count, ] right? Because if you think that is a small portion of the total cost for data center, and it's so critical like the feedback we get is that it's just going to get more and more complex.

It's certainly getting more complex because what you're doing now is you're building like dry coolers into your chiller, right? Maybe we won't even call them chillers in the future, right, because these things are different. The use of water in a data center is you don't want to necessarily use evaporative cooling, right, because you use a lot of water. So many colos and hyperscalers will say, we're not going to -- we're going to have all closed loop systems. Well, there's a different complication associated with that. We have some that are looking for how can we go from chiller to chip, right? Interesting concept. We have a hyperscaler that's working on that. And how do you do that? What do you do with deionized water? You do it with stainless steel components because it's about purity of water. How do you maintain water flows? So all that is part of this chillers getting smarter in the future kind of mindset that you need to have.

And maybe last data center question. How much exposure do you have to neo clouds? And is that portion of the market evolving differently than sort of more hyperscaler focused?

I think neo clouds are getting -- they're getting bigger, so we can see if they are or not. I think we deal with all chiller customers. So you'd have to be specific with the name, and then I probably won't tell you anyway. But we deal with everyone. So everyone is our friend if they are a customer.

Maybe on residential, since first quarter earnings, what has the residential market look like? How has the start to the selling season been?

Let me start?

Sure.

Yes. I mean first quarter, we had guided residential down around 20%, and the business did better than that. It was down about mid-single digits. I think the most important part was entering the year, we wanted the inventory and the channel to be at the right levels, okay? And we took some drastic actions in the fourth quarter internally from a production perspective. We took about 1/3 of the production days out of our residential business in the fourth quarter. We also decided starting this year, we would level load the factories. And the path usually has been you ramp up production through the first quarter into the second quarter to build inventory to be ready for season. This year, we said, okay, we're going to take the inventories to the right level in the channel. We're going to level load production this year. And ultimately, had a good start to the first quarter. It is just the first quarter, right? It's maybe the least significant quarter of the 4 for residential. But what we're doing this year, again, that level loading, it will put a little pressure on the first half of the year when you think on a year-over-year basis, incrementals. That will be a nice tailwind for the second half, though. And also, we have easier comps going into the second half of the year just given the challenges we saw with 2 pre-buys last year, refrigerant change and ultimately, maybe not whether that was as cooperative. So we raised our guidance to say flattish for the year for residential. It means volumes will be down, right? There'll be some positive price in there. But very happy where the channel started the year and happy where the channel ended at the end of the first quarter.

Yes. Look, we're probably the ones that are more bullish on residential than maybe others. And like 2025 was a really, really strange year for residential, right? You had a pre-buy with refrigerant. You could argue you had another prebuy with tariffs. And then all of a sudden, you had a refrigerant problem, where you couldn't install new units. So you couldn't burn through the inventory and you had a really short selling season, meaning it was a lot cooler last year than it had been in a long time. So we could argue about whether that's going to repeat in 2026, but the first 3 are not. So we'll see how the year progresses. But as Chris said, it's the first quarter, but we're off to a good start in that business.

But with level loading, does that mean that in so far as sell-in?

No. Think of it more as what your manufacturing output. So instead of ramping up your factory output in the first 4 months of the year and then letting it drift as you see demand in the back half, we're basically just going to manufacture at the same level throughout.

Right. But from that perspective, it means that there's not going to be a lot of variability to second quarter unless demand is really well outside your range?

You have inventory that buffers that, your own inventory. So if you had additional need, you'd be able to satisfy that with the inventory level. Yes, and then you would just ramp up your rates. What it does is it means less unit volume through your factories in the first half of the year, which means less absorption. And then you'll have more absorption in the back half of the year. And if you remember, in the fourth quarter, we made the decision that, okay, we're going to get this excess inventory in the field adjusted. So we took 1/3 of our production days out of our factory. And that was not an easy decision for a lot of reasons, forget about the financial side of it, just think about the impact to our employees. So we wanted to get that over with. And we're happy to say right now, inventory is in a good position right now within our residential business.

But in terms of demand. So the point is that if demand is stronger, you will work down the inventory?

Yes, we'll just buffer with it.

We have the availability for that. But we did guide the second quarter to about flattish eventual and growth in the second half of the year really off of easy comps. And then let's see how the year kind of plays out. But the team is really executing well there where they started the year and ultimately managing the inventory and then taking the decisions around, as we said, the factory production to level load.

How should we think about replacement versus discretionary upgrades in residential as sort of affordability pressures persist? And do you see a mix shift towards things like Oxbox?

Is it -- okay. Are you talking about like higher SEER to lower SEER? Or are you talking about repair versus replace or?

Well, it's just people -- that's right. It's like people just -- because things are so expensive.

Repair versus replace, I don't think you could tell in the first quarter, right? So it's a shoulder season. So we'll see how that plays out as we go through the year. We have not seen a mix change. So we're still -- we lead in high SEER products. But we haven't seen -- we really haven't seen a mix shift. Although it's just the first quarter. So we'll see how the year progresses. But right now, we haven't seen that.

Excellent. Maybe just focus on transport. I think transport remains one of the slower recovering pieces of the portfolio.

And we're in year 5 of a 2-year downturn.

Yes. Anything about...

I think we are seeing green shoots there, okay? I mean, look, I know this business very well, and the fleets are very old. And the good trucking companies and their -- most of these trucking companies are very well run. They know there's a cost to having an old fleet. And that will -- so as an underlying macro, I would say, that's one thing that's out there. The other is if you look at what spot rates are doing, they've actually are increasing. If you look at rejection rates, meaning that I have a contract, but you go show up and it's like, "I can't deliver because they want you to go on the spot side." that's increasing. And utilization of fleet is increasing. So those are all positive signs for that business. Now I've been wrong because last year, I was probably up here, maybe not at this particular conference, but in a conference saying, the back half of the year is going to be a lot stronger, and we're going to have a recovery, and that didn't happen. But I guess I'll say I'm more optimistic this year than I was last year about Thermo King recovering, which is a good sign for Trane Technologies. As you think about it, we had a residential business that was -- had a terrible 2025 for the reasons I explained. You have Thermo King that's been down for at least 4 plus years now. And both of those, we're seeing at the back half of the year, those are going to be positive. And we think we're going to have positive momentum going into 2027. On the Thermo King business specifically, if you look at ACT, their projections for 2027, we don't -- we're not as bullish as they are, okay? We believe that 2027 will be a strong year, but we don't believe we'll see the inflection point up that they're predicting. I think they're -- the demand might be there, but I don't know if the trailer OEMs will be able to meet that. And because we're a reefer that goes on a trailer. If you don't have a trailer, you're not going to sell the reefer. But we'll see how that plays out, but we're optimistic about the future in Thermo King.

What does the normalized margin profile look for transport once volumes recover relative to historical peaks?

I would say that we like margins in that business today, and we'll like the margins, I think, even better when you get more volumes coming through the factories with more units. At the same time, we made some really good investments in this business for the last 3 to 4 years when markets have been down. We're not going to be out there trying to sell the same products we had 4 years ago and say, here they are. They're maybe a bit more expensive due to inflation. They're more energy efficient, more options around electrification, to hybrid, to more efficient even gas-powered units. So I think we've been -- that team has done a phenomenal job outperforming in markets. And the same would be within our EMEA transport team as well, right? There's a market that hasn't necessarily grown either for many, many years. And they've outperformed those markets, and it really comes back to the innovation flywheel of where they've invested. And I think as we see these markets start to recover, we'll like the incrementals and I think gives us a lot of confidence we should be 25% plus. And that's why we like the plus side after that. It's a commitment that we'll be at 25%, and can it get better, we'll see if we can.

And for EMEA Thermo King, what's driving the outperformance over the past several years? What are you doing to win market share versus your competitors?

I'll start. I mean, look, I think it does come back to the innovation. Again, this is a team that, as Dave said earlier, right, Europe is still very focused on sustainability. It's focused on decarbonization and especially when you get to city centers, there's also noise abatement concerns. You want quiet systems. You want running on electricity. And that team has led with that innovation, all the way from trailers down to truck units to smaller units. But they have a fleet. They have a choice. If you want something that's hybrid, you've got that and you can go full gas as well. What the team has also done is they've done a number of acquisitions actually over the last 12 to 15 months. And where we've typically been 2-step distribution, we've actually bought back some of that distribution and go direct in certain markets. And this is bilateral with the owner. They wanted to sell. We said, okay, we're going to come in and we'll buy back the business. And in some cases, it's a really well-run business. Let's go grow what we have. In some, there's opportunities there for share. And so let's invest in the business. Let's go direct here in those markets. And let's ultimately grow back the share where we need to and/or keep expanding where we're at. So we just actually closed on the fourth acquisition just last week in Germany, one of the larger providers of transport equipment in Germany, we bought back that distribution. So they've now got a combination or a mixed model and more opportunities now to get touching to the service side. So think about the dealer is the one that we do service for the units. Now by going direct, we have access to that service and how do we make sure we're servicing customers effectively.

And just talking about services. Services is now 1/3 with the revenue growing low teens. So the question is, do data centers accelerate service growth through the end of the decade?

I think that any time you expand your installed base, it's going to accelerate service. And obviously, we're doing a lot of data center work. So I think it's going to be good for our service business. Our service business is 1/3 of the company, and it has demonstrated results of a compound annual growth rate of low teens for the last 7 years. So this is a very, very resilient business, a very proven business. It's got an operating system built around it that is very, very robust. And look, we love investing in our service business. I love what we're doing on the digital side of our service business. I love what we're doing on the virtual engineer within our service business with ARIA, making our technicians smarter in front of customers. This is all part of our philosophy, and it's a business that's 1/3 of the company that has demonstrated compound annual growth rate of 11% over the last 7 years.

As And as maybe the last question. As service mix rises, what's the ceiling for enterprise margins without sacrificing growth and investment?

We love the 25% plus, right? Look, could you leverage more in a particular quarter? For sure. I would tell you that we like optionality, we love to invest in ourselves because we like to think about the long term. And I've probably told this group before, the easiest decision a CEO can make is to cut investments because in the short term, you look like a hero. In the long term, you don't have a business or you're not going to be able to hit the growth rates. Look at our growth rates, look at the model that we have. We love investing in ourselves for the long term. And that's a model that we've been able to prove and it's a model that you're going to see from us well into the future. So our model is 25% plus. And we assume that within that, we're always going to have lots of ample capital to invest in ourselves, and that's what we've been able to do, and that's why we've become the growth company that we are.

Well, on that note, we'll end.

All right. Thanks, everyone, for your attention today, and talk to you all soon.