Yeah, sure thing.
Okay. Great. Thank you, everybody, for joining us. My name is Mark Delaney, and I cover Flex for Goldman Sachs. I'm very pleased to have with us from Flex, Revathi Advaithi, the CEO, and Chris Butler, the President of Embedded and Critical Power. Thank you both for joining.
Thanks for having us.
Thanks for having us. Well, data center has been a real big topic for the broader markets as well as for Flex. And so I wanted to start off with some questions on the data center and power business, and maybe starting with you, Revathi. You know, as we think about that business, I think the company said $4.8 billion of revenue in fiscal 2025, and nearly 20% of your total revenue. Maybe speak to the products and services that Flex has for the data center market, and what is enabling the company to be successful.
Yeah. I think what we have shared publicly, Mark, it's not a reported segment for us, yet. But what we've shared is that this year it'll be around a $6.5 billion piece of our business, so 25% of the overall Flex revenue. We said that it would be growing around 35%+, in the year. So you can kinda do the math on the scale of that. What constitutes data center, it has data center and utilities also in it, for us is a little bit of many things. So let me start by saying that on the compute side, we not only do compute integration, you know, which a traditional contract manufacturing would do, but we actually cut the metal, do the metal enclosures like you know, other industrial companies that do in this space.
And then we actually develop the power that goes on the chip itself, and then the cooling, like the cold plates, etc., that goes on the chip itself. And then we also do kinda the auxiliary cooling, like CDUs and things like that. And then we do the power that's outside of the rack all the way up to the utilities. So it's a very comprehensive portfolio, but in a simple way to think about it is that we do compute, we do cooling, and we do power both in the rack and outside of the rack. So that's kinda what makes up that $6.5 billion.
And you mentioned 35% or a bit more growth is the expectation for fiscal 2026. And as you think about what's driving that growth this year, is that just the end market is growing, or is some of that product cycles and traction Flex specifically has seen?
I mean, definitely end markets are growing. In product traction, I would say the biggest product traction actually comes from the business that Chris is running, which is the Embedded Power business. You know, as chips are becoming more power hungry, a lot of technology leaps happening in that space, so that is a significant place of growth for us, and then I'd also say that on the cooling side, you know, what's happening with CDUs and our technology that's around CDUs is also driving significant growth. So I'd say it's a little bit of both. In some places, there is clear technology leaps that are driving it, and then the end markets are, as you very well know, are super bullish, so that's also driving some of the growth.
Yeah. I mean, I think for 2026, the bottom-up CapEx estimates from our analysts who cover the big hyperscalers are for 36% growth next year. So they're continuing to expect nice top-line trends in that sector. Chris, I wanted to go to you and dig a bit more into the power and cooling part of the business, given your role at Flex. And I think the company had said in fiscal 2025 anyway that power was $1.3 billion of revenue.
Correct.
You already talked a little bit about both Embedded and Critical Power, but maybe give us a bit better sense of how that business splits out between the two areas.
Yeah. We don't really break out the individuals, like that, but, you know, one's a little bit bigger than the other, but they're both very substantial pieces of the business. If you think about Embedded and Critical Power, I think first off you kinda need to understand where they fit in the value chain, right? Embedded Power is very early cycle. So if you think about when customers are deciding what chips they're gonna use on their next generation, our Embedded Power teams are designing the power modules and a lot of the hardware that's required, right, to operate those chips in an efficient way. And that's really where the conversation starts, and then we have the ability to actually build out the stack.
So, you know, from those initial conversations, we understand what the power consumption's gonna be in the facility, and then we can start bringing our Critical Power business in and start building on a much more integrated solution.
Okay. And as you think about the growth rates this year, I mean, is there meaningful differences between Embedded and Critical Power?
Not.
Are they both contributing?
Not really. Like I said, from a value chain perspective, I mean, I think you're starting to see a lot more acceleration from the AI investments in the embedded power space, and that's gonna naturally translate into acceleration in Critical Power. The decisions are just made at different parts of the cycle.
Of our overall 35% growth, Mark, you know, obviously the power business is smaller, so it's growing a little bit faster, but our compute business is growing pretty fast too, so both are kinda driving to that 35%+ growth.
Okay, that's very helpful. I wanted to touch on one of the tech themes that we saw pretty present at the Supercomputing Conference that we went to a few weeks back, and it was 800 volt, and I think Flex had a press release out in conjunction with NVIDIA on 800 volt technology, so maybe just to start, why is 800 volt needed?
Yeah. Well, that's maybe just break it down to the 800 volt topic, or even just the physics of trying to get power, right, into these racks that are becoming more and more power hungry. You know, if you think about racks even just a year or two ago, I mean, we were still deploying racks that were 15 kW, right, 15 kW. Now we have customers looking at, you know, hundreds of kW, right, in that same physical space. And there's no real way to distribute that power without doing something differently. And the easiest thing to do, and it's a very linear relationship, you know, we can increase the voltage that's gonna reduce the amount of current that's required. It'll reduce the amount of copper that's needed in that same physical space, and it will allow us to make those racks more dense.
So, there's been this natural drive to increase the amount of voltage that's in those racks. And at OCP, we had displayed our 400 volt DC rack. So this is kinda the first foray into what a lot of people in the industry are calling high voltage DC in the data center, where you know people are deploying power at 400 volt DC, and we have a one MW rack that's deploying a MW worth of power at 400 volts. And you know the natural transition is gonna be people are gonna go to higher voltages because they can continue to increase efficiencies, and they can continue to squeeze down the package. So we're very embedded in that conversation, right?
Obviously we have a 400 volt product that is going through the final stages of validation right now and hoping to, you know, start ramping production sometime probably late next year. Then, you know, the natural transition of that is to get into 800 volt or even beyond 800 volt. I think some of the technical challenges that people have around, you know, even 400 volt DC, you know, there still hasn't been a shift in the industry around how we're gonna handle these things safely. So I, you know, I'm pretty comfortable with the 400 volt transition because we're proving that out. 800 volt's gonna be a little bit longer of a putt. So but the industry is moving towards that direction, higher voltages.
Okay. Is there a specific GPU or ASIC that's gonna require 800 volt? And you can say, "Oh, in 2027 or 2028, we, we definitely need to go to 800 volt," or is it too hard to say?
You know, it's interesting the role that the silicon providers are playing today because in the past they would provide a chip, and they tell the customer, "Hey, you go figure it out," right? And now, you know, to use NVIDIA as the example, they've built this, you know, community of folks that were trying to build reference designs for the chips in advance so that they can be rapidly deployed. So if you look at the Rubin and Rubin Plus, where we're starting to get into 300- and even 600 kW IT racks. So when I talk about a one MW rack, that's a one MW rack of power. And then that'll be coupled up with maybe a couple IT racks, and most of the designs we're looking at today, that's two or three IT racks and then a cooling rack on the end.
But, you know, when you get there, I mean, NVIDIA is driving that 800 volt, because it's gonna allow people to deploy it more quickly and in the most efficient way possible, so.
I'd say end of next year we'll see 400 volt deployments, a calendar year. So when you're thinking about kinda chip technology, I'd say seeing it in 2027 may be quite a putt. So, I would say it'll probably take a lot longer than that. But because it's not just a product design, it's the regulatory environment of safely deploying these and being able to service them, which is pretty significant. So the infrastructure has to go along with it.
Okay. And you think about something like 400-volt coming, so sooner than that, but you have it still, on the come. I mean, what does that mean for your business? Is it more revenue per rack, or how should we think about the implications from a financial perspective?
Yeah. I mean, I think honestly it's quite transformational, at least for the Embedded Power part of the business, right? Because if you think about what we were doing in the past, you know, a hyperscaler would come to us and say, "Hey, we need this particular power shelf, and here's our specs, and, you know, we need you to develop the best design that you can, right, to meet these needs." You know, and it was really, you know, a product play. Now we're talking about a fully integrated solution.
So if you think about it in that power rack, in that one MW power rack, you're not only gonna have the power supply units and the BBUs that we would have traditionally provided, but now we're fully integrating it, you know, the bus bars, the rack itself, you know, how that power distribution then comes into the IT rack. That's all now part of the equation. And the way I, you know, probably the easiest way to think about it is we're gonna have more dollar content per MW than we did before. And, you know, we're continuing to kinda move up the technology stack with more expertise there. So.
What did you think about these big integrated, or these big systems, actually installing it is a challenge? And Flex has been one of the companies trying to help your customers overcome that with some of the prefabrication work. And I think PowerPods is one of the solutions that you bring to bear, to help with that. Maybe just talk a little bit more around what you're seeing around customers wanting some prefabrication done and what it might mean for installation times and what you can deliver to these customers.
Yeah. So we've been in the PowerPod business for several years, primarily in Europe. So, you know, as part of the Anord business, I mean, that was a pretty core piece of what we do, in the Europe market. And they were obviously, you know, rapidly adopting these modular solutions because of the cost of labor and some of the challenges that you have in that market. And we see that same trend coming here to the U. S. Now we were traditionally doing just PowerPods, right? But when I say a PowerPod, we're talking about taking all the electrical distribution equipment that's needed for that facility and putting it into a building that can be dropped on site.
Now, what we announced at OCP just a month or so ago is we're building on that PowerPod idea and that concept and now bringing in IT pods and bringing in our cooling hardware as well. And really the idea there is that customers need more than just the power for the site. And by deploying it in a modular format, while there's probably some technological hurdles that we need to, you know, solve for the customer, in reality, what they're really looking for is speed. And by delivering these modules to site, either in a if you've seen any of these online, I mean, we actually build them like they look like little mini houses, right, that go on site.
Whether we do it in that format or even a skid, it can tremendously reduce the amount of time that's required to deploy and commission this equipment because we'll do all that commissioning on site in a controlled environment. Then when it's delivered to site, they just pull these units in, bring their power cables in, land their power cables, do some final testing, and away it goes. So it's, it shortens the delivery cycle tremendously, and then at the same time, it can reduce the, the schedule risk that you would see on site.
And the real key, I think, in what Chris is saying is that a pod can be the size of this room, or it can be like a tenth of the size. So, you know, a pod can be a variety of different things. But from electrical pods, we're talking about cooling pods and compute pods. And the idea of a company being able to do all three of them and being able to bring it together, the same thing that happens in a rack within a data center, and being able to do that at scale and deploy that with speed is the differentiation that we think makes us qualified to do it.
You both have been in this industry for many years, so hopefully it allows me to ask you some tough questions that are on the minds of the broader investment community, which is on the power to meet all this AI demand, and as you guys look at your business and just the broader ecosystem and from your industry backgrounds, I mean, do you think there's gonna be enough power on the grid to meet all this AI demand?
First, I'll tell you, Mark, I'm so excited that the industry is finally paying attention to what was a sleepy industry, you know, in terms of power generation, transmission, and distribution, right? And, I still remember the days where it's so hard to get utilities to spend any money in terms of upgrade cycles. And here we are in a unique situation where the push for compute and data center growth is really driving utility investment, and it's gonna drive a generational technology and investment change in U.S. utility infrastructure, which is really exciting for geeks like me and Chris in this space.
I would say the important thing to focus on here is one is I'm a big believer that it's true because the world's gonna get compute hungry, compute's gonna get power hungry, and you have to have more power, and utilities have to deploy more. I would say the real important thing to focus on is not just the magnitude of the numbers you see people talking about because everybody's really getting focused on is it a hype cycle. I think it is just gonna be much better than where it is today, right?
For that, I keep saying that whether it's, you know, it's not important whether it's 35% or 40% growth, but, you know, we're talking about pretty significant growth cycles for a few years now at least that we can visibly see that I think is really exciting from an industry perspective. So, yeah, I'm a believer that this has legs.
On the utility market, the company acquired Crown Technical in 2024 and gave you more of a presence to specifically sell to utilities as well as just augment the broader power portfolio. Maybe talk about what Crown Technical did in terms of building out a utility business and if you can help give us a sense of how big that might be for Flex currently.
We don't share numbers, but we can tell you what Crown does.
Yeah. They, I'll just share a little bit about the strategy behind Crown. You know, obviously the primary driver for us to acquire Crown was their ability to do modular systems in the U.S. market. You know, we had a really strong capability of doing that over in Europe, and, you know, we needed, you know, a way to expand here in the U.S., and it was a great catalyst for us. And on top of that, it brought to us some additional product capability. With the Crown technology acquisition, we got medium voltage equipment. And then on top of that, obviously this position in the utility industry, which is closely tied to data center. So, you know, if you look at Crown, they were focused primarily on utilities on the West Coast.
They had a small facility in Dallas, and we were able to invest significantly in our new facility in Dallas, which just opened a couple of weeks ago. We had our first, you know, set of customers in and doing a, you know, a great day of showing off some of the capabilities of that new site. By expanding that, it gives us a couple of things. Number one, it gives us the opportunity to obviously do more in the utilities and focus on utilities that are, you know, a little bit further east. That's part of our strategy, right? Utilities, it's a very long gestation period, right, for utility customers. So you really have to, you know, spend the time and invest the resources in order to capture some of those accounts.
So by now having that capacity and that added capability, we can move a little bit further east. These large buildings too are really difficult to ship. So having multiple geographic locations in the U.S. is really helpful as well. And then also, that facility is also the, the key area where we're expanding our data center module business. So it's kinda helping in both sides, right? We have obviously the utility business that's also growing through Crown, and, and that's growing, you know, again, we don't break it out, so it's still in that 35% category, that we talk about. So that's growing very healthily as well. Probably a little bit longer gestation period than some of the opportunities in data center, but, it's our strategy to continue to grow it and maybe even add to the portfolio, continue to expand.
It's the idea of continued diversification. You know, we added distributed power with Anord, and then we said, "Let's go, you know, through the entire infrastructure chain and then add more utilities," you know, to it. We're looking to build out our services business, which still needs to be stronger. So we're just adding pieces to the overall portfolio and also helping kinda the cycle of that business so we can have a, you know, fairly through-the-cycle kinda portfolio there.
You know, as you think about planning capacity to meet all of this demand across the whole spectrum of power products that you have, you mentioned adding a little bit with Crown, but more broadly, do you feel like you're able to keep up with the demand, and how much capacity might you need to put into place?
The short answer is, no, we're not able to keep up with demand, and we are investing in, we have been planning and investing now in terms of capacity, both for the compute, all three places: compute, cooling, and power. We announced expansions in Dallas. We've been expanding our Mexico facilities in Guadalajara. We have been moving things around, into external warehouses and expanding manufacturing footprint in our sites in Columbia, and Richmond, and in Fontana, and in Milpitas in California, so yes, there's a lot of capacity expansion that is going on today across all three: compute, cooling, and power. And we're very focused, Mark, right now, thinking about capacity for the next year and the year after and kinda where exactly do we have to plan for this kinda growth, so yeah, that definite work going on in terms of capacity allocation right now.
I did wanna talk on the cooling side. Revathi, earlier on in our conversation, you did mention CDUs. I think the company acquired JetCool a year or two ago, and that brought some cold plate and CDU capability. Maybe just give us an update on how that product set on the cooling side is developing.
It's going really well. You know, we bought the category because, you know, Chris described it. You have the chip, you know, we're developing the power that goes on the chip, and then having the cold plate and making sure that you have the right technology so that sandwich that you're developing can be an integrated technology was the real reason we kinda bought JetCool. They did not have a CDU product, but they had a product that was in the pipeline, so we completed it and launched a CDU line that just happened over the last kinda few months. So we're really excited about just completing that portfolio.
I would say it's been a fantastic acquisition for us because there's no way that you can work on generating that kinda heat in our power products and then not have the cooling technology capability to channel all that energy. It's been going really well. I would say it is a definite important piece of the overall portfolio within the rack, for cooling itself, and then the CDUs is a nice kinda auxiliary benefit of all of that.
One thing I also wanna ask about was the recent collaboration that you've announced with LG on thermal management solutions. Could you speak more to what this partnership would entail and what each partner would bring to that and when you expect products to be available, driven by this collaboration?
Yeah. So LG, you know, if you take a look at what we have in our JetCool portfolio, really, it's, you know, at the chip and then the CDU. It's really that secondary cooling loop, right? We didn't have a lot of capability beyond that. With the LG partnership, what's really fantastic about it is it's very complementary. What we do in the rack and around the rack, they don't have, and we don't have what they do really well, which is the cooling infrastructure for the facility. It's a complementary product offer. We're also complementary from a geographic perspective as well. They tend to focus more on the markets in Asia. We're obviously focused more on the markets in North America.
So this is a really great collaboration, and it gives us the ability to provide that entire cooling stack that's required. And it's a really important part of the modular conversation that we were having earlier. You know, we wanna be able to provide the complete solution, both cooling, IT, and power, all in these pre-configured modules. And LG is gonna be a really important component to that. And so we're super excited. It's very new. It's only, you know, about a 30-day-old, you know, announcement. But we had the team in Austin yesterday, and we're really excited about the partnership and where it's gonna go.
Mark, the way we're thinking about the portfolio is we're looking at how technology is getting developed as individual products and as a technology infrastructure as a whole. And we're saying there are some parts of this that we would like to own because they're integral to the design itself. There are some parts of this that we'd like to have partnerships with, because it makes the development cycle easier for a customer. And that's how you would think about kinda we have secondary cooling. Kinda the primary cooling is something that we don't think is a space that we wanna be in, but we wanna bring partners in that can develop an integrated solution. So in some, there will be revenue benefit. In some, there will be the benefit of developing a fully deployed technology for the customer. So that's kinda how we're thinking about these partnerships.
Apologies. I forget the exact number, but I think a year or so ago, you were at 80% or so of the kinds of capabilities that Flex could bring. With some of these new partnerships, where would you sit now?
I'd say we're still around the 80%, because we don't count the partnerships as things that we own and we bring to the market today, so that 80% is a fairly big number, and I feel like the 100% doesn't give you a whole lot more, so we're kinda quite comfortable with the 80%, and then these are just partnership deployments that we'll bring in.
Okay. I wanna talk about the warrant agreement with Amazon. I know that's not specific to the data center, but obviously they are with AWS, one of the big hyperscalers that's out there. Maybe help us better understand to the extent you can when we should think about the AWS warrant transaction or Amazon warrant transaction, I should say, leading to faster growth for Flex.
I'd say, you know, not just the part, the warrant with Amazon, but all of these deployments that you're hearing about, you know, they have long gestation periods, right? So, they're Amazon's already a very big customer for us. It is all about kinda the technology and the things we're developing with them today that will come out in the next year or two years. So I'd say most of these work that is going on around the big announcements coming out from AI infrastructure players, you know, takes time for these things to show up in terms of revenue. So, I would say, you know, hopefully it'll keep adding to the 30%+ growth you're seeing in our business over the next few years.
Right. And would you imagine doing similar deals with other hyperscalers?
Nothing to share at this point.
As long as we're webcasting, we're always happy to break news at the conference, but understand that there's only so much you can say. Maybe just kinda a little bit more broadly, you know, as you think about the data center market and just the competitive landscape, you know, be it Asian-based competitors or as well as North American-based competitors. Can you just talk a bit about the competitive landscape, and is it becoming more challenging?
I'd say not at all. I think what we are seeing in terms of, you know, the Asian competitor thing, not really seeing a whole lot of that. I think we have gone well beyond that conversation today because most folks want local, whether you're in Europe or in North America, you want local capability. And so that is super important. And so having Western companies that have built long, you know, cycles of capability is really important. I'd say in the competitive landscape, the thing to remember is, as we talked about this grid-to-chip solution, we compete in individual pieces with lots of companies, right? So an embedded chip, we compete with another company. In, you know, obviously in critical power, we have all the big electrical players. In cooling, there's a whole lot of, you know, smaller players we compete with.
In compute racks, we compete with some of the larger data centers. In rack enclosures, we compete with some of the electrical players. So you have to think about it as the bits and pieces have lots of unique competitors within that. And then in places where complete technology deployment becomes a critical advantage, then we will bring a more holistic solution to the play. But I wanna be best in class in all these individual bits and pieces. And the selling feature is in places where complete technology, where you want everything stuck in a pod, and you want a compute pod, and a cooling for an entire pod, we can do that too. And that's a plus. So the competitive landscape is one that we feel very comfortable with in terms of our presence there. The beauty of being Flex, Mark, is large-scale manufacturer.
We know how to deploy at speed. What does data centers, you know, companies need today? Scale and speed. And you bring that with kinda the power experience and cooling experience. It's a game changer.
One other trend we've seen in the data center space has been consignment. And I know you, you've given some metrics around how much of your business is consignment. Maybe just to talk about where you think Flex is on that journey. Do you expect consignment to become a much bigger percentage of the business than it currently is, or is it kinda more a steady state? And what does that mean for margins as you do more consignment?
You know, what we have said is, you know, we shared it once just because it's become such a large portion of the overall business. And we really let the customers make the decision in terms of what is the best business model for them. And we'll work with them on that business model. And to a large extent, it makes sense because you don't want to be procuring these very expensive chips and memories and all of that for them on behalf of the customer. You want them to take the, you know, procurement in that. So I'd say, you know, nothing new in terms of change to our overall numbers. But as the business itself grows, yeah, the magnitude of it will grow. It does help in terms of margin rate a little bit, but it hasn't been significant for us.
Okay. I wanna talk a bit about your manufacturing strategy. Flex has talked about having a very significant presence in North America. Part of that's in the U.S. Part of that's Mexico. Maybe just double-click a bit on that if you could. Have you seen any increased interest specific for U.S. manufacturing given the tariff and supply chain challenges that some of your customers are contending with?
Yeah. Definitely, we're seeing more interest in terms of local manufacturing. What we are doing is we're being very picky about where we want which end markets do we wanna be in for that, right? We're already putting so much investment into data center and utilities. We're investing in health. So we're making sure we kinda pick and choose the end markets that we wanna be in. So, for example, if consumer end markets customers want us to put an investment in North America, I'm like, you know, maybe that's not our sweet spot, right? And the same may be the case for automotive today till the industry stabilizes. So yes, we're seeing a lot of interest, but we're being very picky about this. This is the time for us to choose which end markets do we wanna put more capacity in.
And so we're doing that mainly focused around health, around industrials, and around data center.
Specific for the U.S., right? Yeah.
Specifically for the U.S.
Okay, maybe you could touch on some of those other end markets. I mean, automotive, you've talked about being sluggish, and it sounds like that might still be the case from the comments you were just making. But maybe talk on autos. What do you think it would take for that business to return to growth?
I'd say the way, if you step back, the way I think about kinda our overall portfolio, I'd say data center, utilities, strong growth. Health is going really well. We're really happy with that. We're particularly focused on device manufacturing and health, so that's growing well. I'd say industrials where we had a big renewables presence, that was a drag for us as a whole, and mainly residential renewables has kinda stabilized, and we're, you know, starting to see industrial growth mainly around infrastructure deployment. Automotive, I would to some extent, I feel, has flattened out. Then particularly OEMs are more clear about which platforms they wanna invest in because there was a whole confusion on hybrid or EV. So I'd say we have more stability in terms of that, returning to growth. Consumer end markets are fairly flat. I'd say that's anybody's guess, right?
Where does consumer spending go with tariffs and all of that is to be seen, and so that is still kinda pretty flattish, and I'm not kinda projecting any kinda huge return to growth there.
Okay. One technology that's had a lot of focus in the financial community and in the media has been around robotics, humanoid robotics in particular or something. So maybe just talk about to what extent that kinda technology is something that Flex could use in its own sites, and would Flex look to help manufacture humanoid robotics?
Yeah. So for sure, we're in our industrial business, robotics is a big focus. You know, whether it is warehouse robotics, humanoid robotics hasn't had a big presence yet, but other robots are a big focus for us. So it's a very growing business for us within our contract manufacturing space and industrial. So we're super excited about that. And as we have, you know, presence, we'll continue to work with customers on going into other types of robots as they become an important part of the infrastructure. I'd say for Flex deployment itself, I'm super excited about what AI and automation around robotics will do for our business in terms of productivity. I have said this publicly. If I feel like that is a game changer for the manufacturing industry.
Not just hardware automation and robotics, but software automation and robotics, focus mainly around AI is really gonna change the landscape, and if you think about kinda the next leg of productivity for a company like Flex, that's where it's gonna come from, so.
Is it kind of thing where you've got cameras and it says, "Oh, we can see a defect that maybe a human couldn't"? Or is there something even beyond that that you're not?
It's that kinda stuff which is just machine learning, right? Which is more hardware-related, and you'll see more and more of that. You'll see kinda mature platforms getting deployed. But the software integration, Mark, part of this is gonna be a game changer. A typical manufacturing company will have anywhere from 50 software systems, if they're good, to like 200 to 300 software systems. So having data structure integration and then using the power of AI to make end-to-end decisions is gonna be a productivity game changer. So pay as much attention to the software side of it as we're paying to the hardware side of it. That's gonna be a huge plus for all of us.
That's very interesting. Well, maybe sticking with the productivity theme, we can close with a question on margins. The implied 4Q operating margin guide is very strong, in my opinion, at about 6.8%. As you think about that outlook that the company provided, is there anything unusual in terms of timing or mix, that you think's helping the margin in the fourth quarter? Or would you consider that illustrative of the longer-term profit potential of the business?
You know, our margin progression has been very consistent over the last kinda six years, right? Where it's coming from is from two areas. One is from mixing up. So if you're seeing higher growth from areas like data center, and that mix is driving profitability, and from internal productivity deployments that we just talked about. So we've had very steady margin progression. You know, a year ago, you all were asking me about, "Hey, is 6% sustainable? We were a year ahead of schedule." And as usual, I was hedging my bets and saying, "Yeah, we'll see." My answer is still the same. Margin progression is coming from mix uplift in terms of high-growth areas and from productivity, and we should continue to see that trend.
That's great. Well, look forward to seeing how that all develops. Wanna thank both Revathi and Chris for joining us for this conversation.
Great. Thanks for having us.
Thanks for having us.