Nokia Oyj ($NOKIA)

Good afternoon, everybody. Thanks for joining us. I'm Sandeep Deshpande, I cover the European technology space. And I like to welcome Justin Hotard, CEO of Nokia. Thank you, Justin, for joining us this afternoon. Thank you all for joining us.

I will just jump into the Q&A and then maybe a few of you could ask a few questions as well. I mean, clearly, the interest, Justin, in Nokia is your exposure in the AI cloud market at this point [indiscernible]. So I want to just -- at [ OFC ], you announced a whole bunch of new products, and that is potentially going to change your outlook in the optical market going into '27. Maybe you can talk through some of those new products that you've announced and how you see that impacting your revenue in the next few years?

Yes, absolutely. And thanks again for having me, Sandeep. So I think if you look at what we announced in Optical at OFC, first was -- the first big shift in this was really the benefit of bringing together Infinera and Nokia. That was number one. Number two was we -- by doing that, we enabled the -- us to build a very differentiated portfolio. So 4 DSPs, 13 unique platforms. And the key thing for us in announcing that was giving us an opportunity to engage early with customers on those products in bringing them to market. Now we announced that most of those will be in market in '27, obviously, leveraging the DSPs. And each one targets a different unique application across the [indiscernible] from what we see in scale across all the way to what we see in ultra long-haul networks.

And you see that ramping up from the second half of '27 into '28?

We see volume ramp in second half of '27. So early units in first half and volume ramp in second half of '27.

Understood. Maybe moving on to the Q1 results. I mean, you had this amazing set of results. I have people like myself who have covered Nokia for a long time. I mean this is an amazing set of numbers from Nokia in Q1 given how your orders in AI and cloud were [indiscernible]. You had this EUR 1 billion in orders in Q1, which is almost 40% of the orders you received all of last year [indiscernible]. So maybe the question I have is what changed in Q1, which caused this huge increase in the order intake?

Yes, I think first of all, if you look back at Q4, we actually had pretty strong order momentum in Q4 as well. I mean the first thing we did -- the big thing that we changed is we actually started to give you visibility into the segment. So that's probably the first step. The second thing is obviously what we're seeing is really strong momentum on the back of the demand we see in 800 gig. So we've had traction in 400 gig. Just a reminder, we really weren't a player -- no presence in 400-gig pluggables, some presence in 400 gig in terms of ILAs and some of the other systems in that space. But the momentum we're starting to see and the demand we're seeing for 800 gig was really what drove the order for [indiscernible] Optical. And then on the IP side, we've had a little bit of traction there largely in routing with AI and cloud. But not in the order book in Q1, but what we also announced during earnings was also that we've got a few design wins getting into IP networks, including on the switching side. And so that's been really encouraging for us in terms of the opportunity that we see.

Yes. Since you raised the IP networks market, I mean, you mentioned that you've got a couple, I think, of design wins in the first quarter. You also clarified that they were not in the orders in Q1. That's right. So you expect to see these in the orders later this year?

Yes.

And I mean, clearly, you're not going to answer this question directly. But in terms of the orders in Q1, would that be the sort of orders you expect for the rest of the year? Or this was a big one-off in terms of this EUR 1 billion in orders?

Yes. I mean I think a few things on the orders as we look ahead. I probably won't give you a direct answer on that, Sandeep. But as you rightly said. But the way I would look at it is there's a couple of things that are happening. One is [indiscernible] starting to see more orders into '27. If you look at what happened in Q1 and obviously the second half of '25 as well, what we had there was really most of the orders that give us visibility and confidence into the assumption we have on revenue growth for '26. Now we're starting to see things that are maybe a little bit in '26, but much more demand into '27. What I anticipate is that as we see more orders come in this year, we're going to continue to see some of the time line elongate. We traditionally been a company that's just shared order numbers. Part of that is if you look at our historical business, particularly in telco, but also mission-critical enterprise, we've really not been a business that's had long lead time orders. We typically have got things within a 12-month window. So in telco, we win a frame agreement, for example, but then the orders come in, and they're usually fulfilled in like a 6-month period What's changing in AI and cloud is that we're now seeing orders, obviously, orders grow and orders increase, but also orders elongate. So one of the things we're going to try to do is we're going to think through how we dimensionalize that for you guys so you can get visibility to how much we see over what duration. And we'll think through that. But whether we give you kind of a 12-month rolling forward, 4 quarter forward forecast or something like that, but just to have a sense of what that is. Because I do think things are going to get lumpier, particularly given the constraints that we see in the supply chain, both in optical and IP.

You talked about lumpier in terms of orders, or lumpier in terms of shipments? Or rather...

Lumpier in terms of orders. The reality is -- I mean the constraints right now are all on supply. So the shipments, I think, will be quite linear, but I think orders will be quite lumpier.

Understood. I mean now coming back to those 2 customers that you've got in the first quarter. I mean, would this be correct to assume that the -- you've announced in the past, you have Microsoft as a customer in this market, [indiscernible]. So is it beyond Microsoft? Or is it essentially more wins at that customer, [indiscernible]?

Yes. I mean I think we've announced that we have a win in the past with Microsoft around [ Sonic ]. What I can confirm for you is this is not -- this is a different -- these are different design wins, incremental design wins. In terms of customer details, as you know, we won't disclose that at this time.

Understood. In terms of the wins in the switches, again, I mean, where are you winning? Are you winning inside a data center with top-of-the-rack switches, or [indiscernible] switches? Or is this scale across switches is where the wins are?

Yes. I mean, I think I'm not going to disclose too much on that front, but I think we're -- let's just say we're pleased with the progress we're making across the portfolio, including inside the data center.

Understood. Now when I do the numbers on this, I mean, I mean, clearly, you [indiscernible] not going to entirely answer this question, I understand that, Justin, but you had an order intake of EUR [ 2.5 billion ] in AI cloud last year. You reported only EUR 350 million in revenue in Q1, as such, really which is a very small percentage of those [indiscernible] that you received last year, or even as a percentage of the orders received in Q1. How would this order intake now, I mean, translate into sales through this year? Because if you just -- those of us who model, I mean, 350, 500, 600, something like put that through, you would come to even a much bigger growth for your -- rather or you've guided 18% to 20% growth in IP and -- plus Optical businesses. I mean we can come to a number which is much bigger, given that if that EUR 1 billion order number was to be replicated through this year, you should have a run rate of EUR 4 billion in this business next year potentially in revenue. So I mean, any comments on this?

I mean -- I think there's a lot of really good modelers probably in the room and listening. So I think you've dimensionalized it in a good way. I mean, look, I think for us, this is a huge focus. Obviously, the growth that we saw year-over-year was very strong on revenue. And I'm pleased with the progress we're making, but a lot more work to do to continue to scale. And again, the other thing I'll just reinforce is that if you look at where we are in orders through Q1, we've got visibility to demand through the year. So it's really -- when I think about -- as I think about it, from my perspective, I always think about risks, right, risks and opportunities. But if I look at the assumptions we gave you on growth for the year, what I would think about is that we're basically -- the only risk we really have is on supply chain. And that's not to say that we have risk. It's just that that's where the risk is, it's not on demand. And maybe to put it another way, if we had additional supply, we could probably fulfill that given the demand that we have.

I was in a meeting just [indiscernible] your main competitor in the [ switches ] market. And I mean, whole meeting was discussed -- a large part of the meeting was discussing supply. Maybe you can address -- I mean how are you facing -- TSMC is a huge bottleneck for many of your suppliers. So how -- where is -- where are you on the [indiscernible] side?

Yes. I mean I think for us, it's a little bit broader across the business. But obviously, I mean, I think everything you've heard of from an industry perspective around supply lead times tied to leading-edge nodes of fabs, TSMC. Of course, that's a driver for us as well. The other things we're focused on are in different parts of the business, but important ones are memory. Memory has an effect. There's memory content and routers. There's memory content in some of our consumer premise equipment on fiber -- on fixed networks and fiber networks. There's memory -- there's some member content in radios. We still have a small business in our core software platform where we resell third-party hardware that are servers, that obviously have memory content, that's a business that we're continuing to deemphasize and want to shift to direct fulfillment. You provided, is the right thing for our customers. So those are places where we have memory dependency. And then obviously, as you've heard from many of the industry players in Optical and certainly in [indiscernible] we're scaling at a tremendous pace as an industry, right. [indiscernible] manufacturing. I think it's -- if you look at it, it's just a massive. It's 100 to 1,000x scale that we're talking about in terms of meeting the demand in the market over the next few years and from where we were just a couple of years ago. So obviously, that drives constraints across that side of the supply chain, and that's one of the things that we need to mature as a -- as an industry but as well as a company playing in that. So we look across all of that as well as scale for contract manufacturing. Particularly in Optical, we're moving into this volume. It's a very different it's a different problem than it was just a few years ago. So all of that are areas that we think about in terms of constraints, as well as smaller components in different parts of the risk. But -- but I don't think there's anything from our perspective that unique vis-a-vis where the rest of the industry is. And I would say we feel pretty good as a systems provider, particularly in optical of having vertical integration, particularly as we ramp pluggables. Because we think that's -- we do think that's an advantage to scale under ramping.

I mean one of the things which came up during this conference, I mean, one of your suppliers momentum was here yesterday. And at the [indiscernible], they mentioned that many of your other competitors placed orders quite early in the supply. And one of the points he made was Nokia is suddenly emerging as a major player in this market. So the question is whether you placed orders well enough in time to get your supply, because you seem to be a late sudden bloomer in this market. You were not much in the data center market until a year ago, right?

Yes. Well, again, we weren't in the 400-gig transition, right? And so if you think about the transition on pluggables and the early part of the scale across demand, it's all been on 400 gig. But the pluggables side of that and the system ramp has happened, we were smaller. We're now catching -- we believe we're well positioned in catching pretty strong demand trends in 800 gig. But obviously, we're also working on scaling those relationships to support that demand. And Michael and the team are obviously great partners, and we're working closely with them.

And in your own [ indium-phosphide ] capacity, I mean, you've got your new 6-inch fab ramping up in San Jose later this year. Where are we in that ramp in terms -- because remember, normally, semiconductor fab ramps have problems. Even momentum has had problems ramping up its capacity. He was talking about they have choices, whether they will go back to 4-inch or 6 inch. So the question is, there are issues there in overall in the industry in ramping up. Does -- where does -- how comfortable do you feel with your own ramp-up at this stage? It's clearly early days [ yet ].

Yes. I mean we're early. I think we're pleased with the progress. We'll continue to provide updates on our progress, but we still feel good about early ramp at the end of the year and then scale that production volume through next year. It's aligned to the supply that we see. A key thing for us in this and maybe just one thing to remind everybody of is the majority of the volume that we're building is this, and particularly in that facility, is this photonic integrated circuit for pluggables. So it's a bit different in two dimensions. One, it's a full photonic integrated circuit. So the die size is a little bit bigger. So that creates a different -- those of you that understand semiconductor manufacturing, you understand the yield dynamics of that. But the other side is it's largely one platform. And so as we look at the business, that probably has a little bit of a different dimension than others, right, on both sides. But so far, we feel pretty good about where we are. And as we make progress there, we'll continue to share that and unpack that for us of visibility.

In terms of that capacity, I mean, I think at some of the previous meetings, you've talked about that you will have as much as '25 [indiscernible] capacity by '27, versus what you had in '25. This is quite a lot of additional capacity. Are you going to consume that all yourself? Or are you going to also be a component business selling this externally?

Yes. So today -- I mean, today, when we look at the capacity we've talked about, it's all been in line with selling our own pluggables. And given the demand we see right now, we're largely consuming that demand within wind forecast. But that's not to say we won't do something different in the future, but right now, that's what we're focused on, and obviously delivering for our customers.

I just wanted about touch base on margins in this AI cloud/networks, and Optical networks. Optical networks as stand-alone business was -- has been indicated to be double-digit margin post synergies by Nokia by 2028, as such really. The scale of the revenue growth is much faster than what was expected when you gave that guidance. Does this change the guidance firstly? But secondly, where are you on that -- those synergies because that is also critical in terms of you achieving that double-digit margin?

Yes. Yes. I think first of all, in terms of guidance or assumptions that we outlined we set a set of those assumptions out in -- for CMD through '28. Fundamentally, those a whole, those were at the NI level, and we provided some visibility to IP and optical growth underneath that. So those continue to hold, obviously, as we talked about, we're not going to update those every quarter, but we'll give you visibility into what we see going into next year and provide an update on it from that perspective. And we gave you an update, obviously, on what we see this year versus expectations. In terms of the synergies, I think very clearly, if you go back to the case that I inherited on the acquisition, it was largely a consolidation case. And so if you think about it, it was -- there were sort of 3 elements to the case. One was consolidate two players to become a #1, #2 player, certainly in the western market. But 1 of 3 big players globally. The other part was get increased exposure to the U.S. market, both in AI and cloud, but also in telco. Because we were -- Nokia was relatively less exposed. Infinera was more heavily exposed to those 2 segments. And third was largely around synergies, and it was [ $200 million ] in 3 years. But focused about 2/3 on OpEx and 1/3 on cost of goods sold. As I came in and we started to see the opportunity and scale across and the relevance of what we're doing in pluggables, and I talk to customers and obviously, part of the decision we made around the leadership and pivot we made was -- we basically made a complete flip on the integration plan. We said we're going to go 2/3 COGS because we see higher revenue than we expected, which led to obviously the updated guidance we shared first at CMD, and we've updated for this year in Q1. Second thing we said was we're going to still hold to getting the G&A synergies, which is really where 1/3 of the OpEx comes out because more broadly across the company, as we've talked about, we need to continue to be become a leaner, more agile company. And so this was one of the places where we need to make sure we capture the synergies is something that as a company, we haven't done consistently in the past on acquisitions. So the discipline that we have right now is maximizing those two. What we decided was not to sacrifice R&D. And so instead of cutting R&D, we basically merged the R&D teams together. And the first thing we did was -- first decision we need was to move some of the Nokia engineers to go work on getting the [indiscernible] the 100-gig pluggable out. It feels like that was a pretty good decision in terms of synergies. So we were able to get much more competitive and catch up in that market where we were quite far behind. Second decision we made was to keep the 2 DSP teams, and that's why you see us building 4 DSPs. Typically, the 2 companies would each have released 2 DSPs that were largely the same. Now we're able to deliver 4 DSPs in the same period. So that's the shift. And so that's really the net-net of what we did from an integration plan, capitalize on the opportunity. The last thing that happened, and I'll just touch on this is we -- and David alluded to this in CMD, is we basically said we're 9 to 12 months earlier in terms of realizing the synergies. So we've accelerated our commitment against the synergies. And I think -- and obviously, we're in a position where we're delivering far greater growth than we anticipated as touched on earlier, Sandeep.

When I -- as a tech analyst, we learn that when companies scale up like this in revenues, I mean, the gross margin doesn't necessarily -- because you're outsourcing a lot of the manufacturing, the gross margin doesn't change much. So if you're going to do 42%, 43% gross margin, that might be 45%, but it's not going to become 50% or 60% gross margin [ at that ]. But because of the gearing in the OpEx, the operating margin goes up a lot. I mean when you see your guidance at the moment to '28 is 13% to 17%, which is better than where you are today, no doubt, or rather in '25, but it doesn't show any dramatic gearing. So is it that you're going to spend a lot more OpEx in the next few years? Because you seem to see in orders and which will eventually translate into sales. We should see Nokia seeing much better operating margin in the NI business.

Yes. Again, a few things on this. So first of all, I'm going to go back to CMD where we provided the multiyear guidance, right? So [indiscernible] the set of assumptions for the multiyear against the operating buffet guidance. First of all there, we had a set of assumptions on growth and performance. We had already assumed that we were investing the R&D -- making the R&D investment on Optical and also reallocating capital to invest in IP networks at that time. So if you think about that, this is important because it's across NI. So if you think about that, that was the assumption then. Since then, we've talked about we're incrementally investing, and that's largely around scaling, manufacturing, scaling capabilities in that in optical, we're continuing, obviously, to maintain the investment in IP. But also, we haven't -- we're also going through all of the -- the investment, or the efficiencies investment we're making to drive cost out of the company at a corporate level for G&A and driving productivity. We're also -- the other thing we talked about a little bit is we also made a decision around the integration of our joint venture in China to then go ahead and capture synergies on that. So we have quite a few things kind of on the move that touch the business. And so therefore, you don't see this big swing in terms of the operating profit drop for this year. What we haven't done is come back and said anything about future years right now. And to be honest, my focus, first quarter into the year is not to update 3-year assumptions because I'm not going to do that every year. But I do think this is a year where we've got a lot of work to do to make -- to position the company for longer-term operating leverage and structural growth. And it's not just within the business. It's also across the business, across the company as we drive the efficiencies, drive simplification of systems, modernization of systems, complete systems integrations that, in some cases, go back to the Alcatel acquisition, so that we can build a much cleaner, simpler company. And of course, in parallel all of that execute the exit and the transition of the portfolio businesses that we outlined at CMD of which we've announced now that we've got a transaction for one of those -- with the FWA business [indiscernible]

If there are any questions on NI, maybe we can take a couple of questions before I go to mobile infrastructure.

I just had a couple of questions. Just had a couple of questions. The first is, now that the AI boom is ramping up, how do you see the economics of the industry in terms of the split in profits between infrastructure providers such as yourself, and data centers? Is there a risk of commoditization of the infrastructure players? And the second question is on [ indium phosphide ]. You mentioned both boom in indium phosphide demand. Do you see a bottleneck in indium phosphide demand? And you think independent foundries could help you offload some of that demand? Or you're just going to try to do everything yourself? And is it going to be another bottleneck like memory where everything just goes crazy?

Okay. So I think those are two questions. Two questions, no problem. So first of all, on the first one, look, I think there's always a place for innovative infrastructure companies to make -- to capture value. I think what you have to be really disciplined on is where is our differentiation and value capture. And that's my focus. Right now, we've got multiple vectors of that across the board. And my message which has been clear across the company is we need to focus our R&D capital where we can differentiate, where we think we can differentiate for the long term. And when we're not differentiated, we need to we need to transition and partner. And in this space, I think we've got multiple vectors for differentiation and how that -- we'll let that play out over time, but we'll continue to stay focused there. On [ Indium phosphide ] manufacturing, look, it's not that easy to say. I mean two things. One, it's not that easy for someone to just ramp up in Indium phosphide fab. So I think that's number one. So if you -- at a component level, the reality is I think you've got a few scale players in the U.S., you have some manufacturing in China. You don't have much else. And I think the reality is the problem ahead of scaling supply is also one of [ mature ] in the manufacturing process. Because we're going from tens of thousands of units to hundreds of thousands to millions to potentially tens of millions in a period that's incredibly accelerated. So I think the key thing for all of us is to figure it out and as an industry to figure it out, right, and solve that problem. I don't think there's other people that will solve it better. Now are there lessons we can learn from other parts of the industry? Of course. Are we working on learning those lessons? Absolutely. I'm sure that my peers that manufacture indium phosphide would tell you the same thing. The second part of that, though, is in any situation like this in semi, and you guys know this because it's a long cycle investment for short-cycle demand. When you have this kind of demand spike, you're going to have constraints, right? And that's just the economic reality of the dynamic. It's [ not ] solvable. Even if we're scaled as an industry, and we have spikes, we're probably going to see this problem. So I think that's the reality of where we are. Our focus right now is how do we derisk that ramp as much as possible. How do we scale and mature it and ensure we can deliver on the inherent demand in the market because that's what's most important.

Yes, question on the competitive environment, both in Optical and data center switching, right? How do you look at that? And what gives Nokia right to win?

Yes. First of all, on optical, I think we have a few different elements. One, if you look at the business, we've got a broad business across systems, components, software. The fact that we've got the DSP capability that we have, the engineering resources we have, we're able to build far more tailored products and at the scale that we're talking about. Literally, we think of every application inside of a customer is a product, which, by the way, is exactly how the customers think about it. If those of you that understand what's happened in storage or compute in the past in AI and cloud, you know that the hyperscalers look at it that way as well. So I think that's one advantage we have. And the other, I believe, is the vertical integration, particularly as we ramp in pluggables is a competitive advantage for us. It simplifies the design. And for us, it gives us the ability to optimize between the DSP and the photonic integrated circuit. On the IP switching side, a couple of things there. One, we make our own silicon on routing. That's obviously a core part of the scale across story is routing. So we believe that's an important [ vintage ] for us. But also we got -- we think we've got a very competitive and robust system design and software stack, and that's the feedback we get from customers, and the ability to tailor that to specific applications.

Thank you, Justin. I just want to go out to Mobile Infrastructure.

Now we're going to talk about Mobile?

Yes, we need to talk a little bit about Mobile. So on Mobile Infrastructure, you've talked about this new strategy where you're going to go to standard product from an ASIC-based solution. Now clearly, that could be [indiscernible] margin accretive to Nokia going forward given the cost of developing ASICs there. My question there would be, can Nokia remain competitive in the Mobile market? Because if I remember, when Nokia transitioned to 5G, the issue was you did not have an ASIC solution in time and that made you lose share to Ericsson and Huawei and some customers. So how do you see this competitive environment play out in a standard product world -- a standard -- chip world?

I think first of all, just for everybody's recollection, Sandeep, and I obviously wasn't there at the time, but my understanding of it is if you go back to the story of what happened at that point, this was -- the company had a misexecution with a third-party partner, right, which was Intel at the time on the ASIC, and it fell back to FPGA, which I think wasn't competitive. But Intel wasn't building a general-purpose product. At the time they were building [indiscernible] This had to do with a fab transition that didn't have to do with the product. So obviously, I feel pretty good about our partner's ability to execute on their silicon time lines. I'm not worried about risk there. They're excellent at this. Also, we're not on the leading edge compute platform. So if you think about -- our first product is going to be, kind of, at the tail end of what they're building in terms of their silicon time line. So that's more favorable. But the other fundamental piece, if you look down the road, is the cost of building an ASIC is going up. And so you have to amortize that cost against your volume. And so you have two penalties you got to consider. One is capital -- deploying the capital involved building that unique ASIC. And then the amortization of all of the mass cost and everything else against the volume that you're going to see. And the reality, when you look at radio is -- and [indiscernible] is we're just not going to see that. I mean, there's already a movement underway. There was a movement underway with x86 around general purpose silicon. When we look at what's going to happen into the next generation with leading edge [indiscernible] we can get so much more -- we can get equivalent performance and so much more upside by building on general purpose. And that's before we even start talking about the capabilities we'll see on AI. So probably later this year, as we trial with customers just see us come out and talk about some of the performance backs and metrics, but we think it's extremely competitive. The shift for us, much like you saw the shift for us in core, is to become much more of a software business on the on the baseband side. And that's really where we spend most of our investment anyway and where we deliver most of the value for customers. The other side of that, which I think is really important is we also need to change the arc of the industry's capital deployment. And what I mean by that is today, the way that we invest in this infrastructure is we basically require a customer to build a completely new radio network every time they make any kind of generational shift. I think that's something that has to change in this industry. I think if you look at one company here that's -- that's quite visible on this, albeit not with us is AT&T, who's committed to [ Open RAN ] and been a leader in this. And I think to their credit, I think that's the right strategy because we need interoperability at the [indiscernible]. I mean it's almost like -- the analogy I'll give you, it's almost like if you came into a data center and said, I'm going to put more power into this data center. Oh, by the way, you got to upgrade all your electrical systems and transformers. I mean that's the logic of what we're doing right now. We're taking something that you should be able to upgrade on a compute cycle on the baseband because it's a computer. And we're saying you need to upgrade that like it's industrial hardware which sits up on an antenna, which is the radio. So unless there's some compelling reason to replace the radio, generational transition, new spectrum, massive reduction in power efficiency which there is, I mean generation to generation, there's power benefits. But unless you're getting some massive benefit forcing the upgrade of the entire system seems very complex. I think this is something the industry needs to change because our customers ultimately aren't returning their cost of capital. So I think this is where we need to pivot strategically as an industry. I think we're leading this pivot. I think what we're going to be able to deliver based on the early validation and test we've been doing over the last couple of years with NVIDIA and [indiscernible] is powerful. I think it positions us to be much more software-centric as I've outlined. And candidly, I think this is an industry we also need to ensure is healthy because from everything I see, I don't see Mobile -- Mobile networks going away. In fact, I see them being far more critical in an era where you've got physical AI. You've got dynamic connectivity, and the reliability and the security and the importance that these networks play in economic and national security for countries is only going to increase.

Just quickly, we've got a minute and left. On the potential for Nokia to gain share in Europe because there's quite a lot of in the network still and European Union is thinking of removing [indiscernible]. Is this a potential in the next couple of years? Or do you think is it still further out?

Look, I think for me, my focus right now is on planning for the market we're in. If there's an accelerated demand increase in Europe because Europe decides that sovereignty extends to their radio networks and that national security on a radio network is very important. I think we'll be well positioned to win a share of it, but it's not something in our base plan.

Understood. And then finally, clearly, I mean, gaining share overall, I mean, not just my way, but in the United States, for instance, where Nokia used to be a much bigger player in the mobile network market. Is this somewhere high on your list? Or is this not very high on your list at this point really?

Look, all 3 major operators in the U.S. are very important customers of Nokias. The two of them are not major radio network customers. But they're very important customers to us, they're important partners. Obviously, we want to grow our market share with them across all parts of our business. That's from -- as the CEO of Nokia, that's my focus. My focus, they're some of the biggest customers in the world. They're very strategic. They're leading in different ways they're innovating, they're leading, and we want to continue to be a great partner for them and grow our business with them.

Thank you so much, Justin

Thanks, Sandeep. Thank you.