Hi, everyone. I'm Alex Duval. I head up the Europe Tech Hardware team for research, Goldman in London. Delighted to be here with Pekka Lundmark, CEO of Nokia. Thank you also all for joining, and look forward at the end of this Q&A to be doing some audience questions as well. Just like to state the conversation is not intended for the media and is off the record. So once again, Pekka, thank you so much for joining. Maybe we can just kick off with some high-level questions. Firstly, we could talk about the 2023 outlook for the wireless market. I think you pointed recently to the broader industry trend of inventory digestion and lower CapEx at some players in North America. So, what's your latest view on scope for recovery in the second half of this year? To what degree is there some conservatism baked into your guidance, that second-half mobile networks and network infrastructure revenue is gonna be roughly flat? The big picture in the operator investment is really that because of the supply chain crisis that they experienced or the industry experienced in 2021 and 2022. In 2022, they bought more than they built, and in 2023, they are buying less than they built. The big picture is that in 2024, we expect this kind of to normalize. Then another question is then that how much do they build? And there we are also seeing some cyclical downturn in terms of their network build speed. And this is something that I believe they are still themselves figuring out, that how will then this look like in 2024, 2025. Of course, the macroeconomic development is playing into this, inflation, and high interest rates is affecting their spend. As we all know that the industry is in a way cyclical. The good thing is that for us is that we are gaining share in a market that is structurally a little bit weak at the moment, but we are strengthening our relative position there. Ultimately, it is a question of timing that when they will have to start spending again, because the data traffic keeps on growing 20%-30% per year, and if they want to stay competitive, eventually they will have to start investing again. Makes total sense. And you talked about hopefully getting to an equilibrium level by next year. What are your thoughts on the sort of growth level for the wireless market as you go into 2024? You know, some investors ask whether there could be a risk that some of the sort of weakness in 2023, as it were, spills into 2024. So how much confidence do you have on growth next year? It all depends on their network build-out speed. I believe the inventory digestion question will be more or less behind us in 2024. That's a very difficult question to answer. If we look at, for example, what Dell'Oro is saying about North American wireless infrastructure market, this year is down significantly, 35%. And I think the latest they are saying is, was it +14% for next year? So this is kind of consistent with what I said earlier, that 2022 was, in a way, a peak year, then more than normal, then 2023 was below normal, and then 2024, hopefully, would be somewhere in between. Got it. Makes sense. And if you were to think on a sort of 2-3-year view, how are you thinking about the market? What are your thoughts on the drivers and growth level? Well, well, there are big differences, of course, in terms of those drivers between the CSP market and the, then the enterprise and Webs cale market. The, the CSP market as a whole will not be a fast growth market. Low single digit kind of trajectory through the cycle at best. So if you want to grow in the CSP business, you have to take share, and that's exactly what we've been doing. We've been investing a lot in technology, and we have been improving our relative position there. So it is possible to grow in the CSP space as well, but then, of course, the real growth driver will be enterprise. And there are two, three big things there. Of course, on the Webs cale side and data centers, AI is driving significant investments there. And then the other part is then the enterprise digitalization, where so far the investments have been driven by IT needs when you put ERP workloads on cloud. But then the thing that is now happening is that operational technology gets connected, gets networked, all these machines are getting connected, and that increases the requirements on networks. So the enterprise networks are also presenting much more than before, kind of mission-critical, CSP type of network requirements, and that is an extremely important thing for us. Makes sense. I'd like to now zoom in a bit on North America. We've kind of talked holistically, but if we think about North America, you know, how realistic is growth next year? I think one of your peers is talking about sort of positive rebound in growth rates, despite the fact Dell'Oro has just, you know, fairly meaningfully revised down growth this year for even chunkier declines. So what gets the market kind of growing next year in North America, given that that was one of the earliest markets to start adopting various flavors of 5G? Well, again, I'm referring to what Dell'Oro is saying. If you ask specifically about 5G infrastructure market in North America, what Dell'Oro is saying, -35% this year and +14% next year, so they are expecting a rebound. But ultimately, this goes back to operators' decisions as to their rollout speed, but the reality is that in North America, only about half, slightly over half of the base stations in 5G have been upgraded to mid-band. So if operators want to stay competitive, they will have to continue to invest. Got it. Let's perhaps pivot to India, 'cause that's been a big story this year. You know, we've seen it emerge as a strong growth driver for the wireless industry. Can you talk about how much sort of further opportunity for growth there is? You know, is this the peak? Because obviously, things rolled out perhaps a bit faster than people thought. Or is this still an opportunity for, for future growth there, either driven by the market itself, or perhaps your strong market share position getting even better in, in the region? Of course, the big needle mover for us in India is the fact that we broke into the 5G network of Reliance Jio, which was 100% Samsung network in 4G, and now we are a significant supplier in 5G. And why is this important? Jio network is the largest mobile network in the world outside of China. They have about 300,000 base stations in their network. The US carriers, I think they are around, what? 80,000-90,000 each, to put things into perspective. So this is a significant needle mover for us. And remember, our Indian revenue in 2022 for the whole group was EUR 1.3 billion. Now, we have over 300% growth in the first half of the year. Of course, that type of growth will not continue, and we are expecting some normalization in the second half of this year, and then into 2024 as well. But, the big picture in India is that if 2022 was here, EUR 1.3 billion, then... 2022 was EUR 1.3 billion, 2023 will be exceptionally high, then the new run rate, new normal, will be somewhere in between, but significantly higher than the EUR 1.3 billion that we had in 2022. Makes total sense. Could you just help us think a bit about the kind of performance requirements you've seen in India? You know, is there anything you'd particularly call out, and how does that play into the strengths of, as it were, the new Nokia, where you have, you know, higher quality silicon and better performance? I think it's a little bit of a myth that customers in countries like India would have much kind of lower requirements and so on. There is some difference, and if I generalize a lot, I would say that in some of these countries, in some cases, kind of off-the-shelf products are okay. But at the same time, also, Indian customers are extremely demanding in terms of their requirements. But the difference is perhaps that some of the Western largest operators, they have had traditionally significant R&D departments of their own that have really kind of taken the network specifications to the perfection, and they have always had a lot of customer-specific requirements that we are not seeing quite to the same degree in countries like India. Makes sense. And how should we think about profitability? 'Cause that was also a topic of conversation, I think, before these rollouts started. Can you help us sort of think about that a little bit? Well, there are obviously margin differences between regions, and it's no secret that North American margins have been higher than in most other parts of the world. India has been a little bit lower, but still they are not horrible margins in any way. India, for us, is gross margin dilutive, but it is still operating margin increasing accretive because of the high high volume. Of course, we have to remember that this, the key characteristics of this business is high fixed cost because of R&D, EUR 4.3 billion last year. And to be able to cover that investment, you need high market share, you need high volumes. Brilliant. Look, I'd like to shift on to network infrastructure in a minute, but perhaps my last couple of questions are on wireless, and particularly on the sort of market share situation. You know, as I alluded to before, you've done a ton of work to improve your product. It feels like you're clearly moving ahead quite well. And it'd be great to get your assessment on sort of where you are now, and to what degree you see further opportunity for share gain, either as a function of that particular investment or continued opportunities from some of these Chinese vendors. You know, where do you see the opportunity for share? Because it feels like the last few quarters or even years, you've actually improved your market share position in wireless in quite a number of markets. Mm-hmm. So in Network Infrastructure, just in case there are people in the audience who have not noticed, but, Network Infrastructure business in 2020 delivered an operating profit of slightly more than EUR 400 million euros in 2020. In 2022, the operating profit was EUR 1.1 billion. So in two years, from 400 million-plus to EUR 1.1 billion in operating profit, and that is a combination of two factors. It's a significant increase in technology competitiveness in pretty much all the segments, and that combined with a higher market share, and you will have seen the growth, top-line growth, in segments like Fixed Networks, for example. There is now some normalization in operator spending, also in the Network Infrastructure side. But also here, the fundamentals are extremely good because the data traffic, as discussed, will continue to-... grow and, for example, fixed broadband penetration in most parts of the world is still extremely low. Once the fiber is in the ground, it will provide you an internal upgrade of opportunities for new capacities, and so on. So we are extremely bullish of the overall kind of outlook of this business. There are market share gain opportunities also here, including from Chinese competitors. Of course, there are in some countries political considerations where people do not want to have vendors from certain countries. We are positioning ourselves as a trusted vendor that you can rely on in all situations. Of course, we come from a trusted country, and we are a trusted company. That is an important thing in many customer decisions. But then there is, of course, a technology race, also. We are now moving to 5 nanometer technology, and then in the coming years, to 3 nanometers, and then all of the competitors will not be able to follow, that development. And then one final point, kind of on, on where the demand for all of this will be going. When you look at the, where the world of, technology is going, there's one hugely important thing, obviously, as we all know, is, is cloud compute. The world is becoming a, a, a massive decentralized computer. The other big thing is AI, that is crunching all this data. These are the two anchors, but the third thing that you do need is the network that connects all those nodes. So the three big things you need is cloud compute, AI, and the network. That network, obviously, wireless is a big driver there, but in relative terms, an even bigger driver for AI will be data center interconnect and all the routing and optical connections that will be needed to connect all the users to the, to the data, data centers. That will be the fundamental driver of NI demand for many, many years to come. That's brilliant. I think that segues very well into my next question, which is just to ask you to give us a high-level view of the latest dynamics you're seeing in each of the kind of main segments within network infrastructure. For example, if we look at the fixed access portion of your business, there should be some opportunities in terms of fiber investment, governmental investment, and things like that. But it'd be great to get a bit of a picture across that, and IP networks, optical, and so on. Yes, as you know, we have four businesses inside the Network Infrastructure business group. And, kind of to simplify this a lot, we have had two businesses that have, for the past two, three years, performed really, really well, and they are IP Networks that has been on high profitability for several years already. Kind of high teens in terms of operating margins. There, we have been gaining share, especially in the edge routing. We are now penetrating to core routing and very importantly, penetrating into Webs cale switching. We had a switching deal with Microsoft, for example, last year, which is really a breakthrough into a new growth segment. The non-CSP share of the IP network business is about 20% at the moment, and that's clearly where the growth will come from. So profitability has been good. Now we need more growth. Then the fixed access is something that has had extremely strong growth in both 2021 and 2022. Now we are seeing some normalization of that growth because the comparables are getting much tougher. But also, the fundamentals are extremely good because, as I said, the penetration rates of homes passed and homes connected in most parts of the world are still very low, and we are clearly the technology leader in passive optical networks. We have over 40% market share in OLTs, which is the central office side of the broadband connection in the whole world. And again, as I said earlier, once the fiber is in the ground, you first upgrade to one gigabit capacity, then you upgrade to 10 gigabits, and then 25 gigabits, and so on. The U.S. government has said that their goal is to have a one gigabit service for every citizen in the country. And that's why they are, for example, investing in this BEAD program, which is connected to the Buy America initiatives. And we are now the first vendor to meet the requirements of Buy America for broadband, which means that we are the first one to qualify for the BEAD funding, which has now been allocated to the States. It's all together $43 billion, I think, is the fund, and about 10% of that will turn into addressable market to Nokia. And we have currently about 60% market share in the U.S. for this type of application. So we believe that, that, that we will be a key beneficiary of this this government spend programs for broadband. And of course, the BEAD program in the U.S. is only one example of programs that we are seeing in different parts of the world. So that's the fixed broadband business. Then Optical Networks is a really interesting case because that is unlike the two others, that has been subscale in terms of size and market share, and also consequently, profitability in the recent years. But there is strong improvement there now. We have invested a little bit in the same way as in 5G. We have invested a lot in technology development recently, and now we have our two latest generations, PSE-5 and PSE-6, which will start ramping at the end of this year, which is our 5 nanometer technology solution, which gives, for example, superior ranges for 500 gigabit Ethernet services. That will help operators to optimize their network architectures, is getting a lot of traction. We have a good growth in Optical now in the first half of the year, and gradually that is turning into profitability as well. This is a business that, despite the overall good profitability of NI, but, Optical has clearly been behind. It has been in low single digits, and our goal is high single digits, in that business. And then the final, final, business, which is, subsea, networks. Obviously, there is a lot of demand in this business, driven by, web-scale investors who are building their own subsea networks. There is a lot of political interest in this business also, because obviously, most of the traffic in the world is touching or going through a subsea cable at some point, and there are quite significant geopolitical and also also defense and military connections to the security of this network. So this is also highly relevant in today's world. The nature of this business is different from our other businesses. It's in a way, offshore project business, where you are laying and maintaining cables, cables undersea cables. Demand is good. Profitability has so far been on the low side, but in the new deals that we have made, we have gradually been pumping up the margins and profitability, so that is also going to the right, right direction. Good dynamics, actually, in all businesses in NI at the moment. Very helpful. I guess just to go back to this point about sort of inventory clear-out, I think you sort of helped us understand that on wireless. To what degree is there sort of different timing or timeline for that? If you think about network infrastructure, is it broadly similar? How should we be thinking about? It is. I would say it's broadly similar. Operators stockpiled, especially customer premises equipment, in 2022, when they were afraid that they would not be getting enough. So in a way, compared to the rollout, they over-invested in or over-purchased in 2022, and now they are under-purchasing this year compared to their rollout, and then we expect that to normalize in 2024. Got it. And it sounds like you're sort of characterizing dynamics as pretty good in network infrastructure. Obviously, we've seen some players, like Juniper, guiding a bit more conservatively than people had perhaps expected, and some of that probably reflected orders, to my understanding. So, you know, how should we sort of conceptualize your positive view, in that context? Is it more just that, as well as market dynamics, you need to overlay, the kind of share gain opportunities and the product cycles which you're undergoing? Of course, we have cases where we are gaining market share, but still, I would say that talking about this inventory digestion dynamics and network rollout speed, I would actually say that our view is pretty similar to what we have seen from some of the competitors. And clearly seeing some of the dynamics that are slowing down the orders, especially this year. Because, again, operators did overspend a little bit in 2022, and now they are maybe underspending in 2023. Great. I'm really keen to leave some time for questions at the end of the fireside, but maybe we can discuss two group-level topics. Firstly, margin, secondly, cost control. So I think you recently reiterated your long-term EBIT margin target, and talked about sort of 14%, if I remember rightly. What are, what are the key levers and drivers that kind of underpin that? And then at the same time, despite recently adjusting your 2023 top-line guidance, you actually only narrowed your EBIT margin target. So can you help us understand a bit the, the moving parts? We have, we have taken a lot of proactive actions on the cost side, because we were kind of expecting that this slowdown in demand would be coming. As some of you may remember, we announced already in 2021, a EUR 600 million cost reduction program, where we said that the pace of that program, or the speed, execution speed, would depend on the overall market demand, which was really strong in 2022, which, as a result of which, we did not move that quickly in 2022. But we are clearly accelerating that now, and we are still keeping the targets that we published when it comes to the cost position at the end of 2023. And as you have seen, despite the weak demand and actually extremely weak demand in our most profitable market, which is North America, which was -40% in the second quarter, a catastrophic number, I would say. Despite that, we delivered 11% operating margin in Q2, which shows that we have been able to increase our cost resiliency quite a lot. My message has been very clear that should it be so that this weakness in overall demand from operators would persist longer than expected, we are ready to take more action on the cost side, if needed. This is not time to announce anything like that, but we are following this very carefully, and there is more to come if needed. ... Great. Well, I think we have 10 minutes left on the panel. I see someone with a microphone over there, so if anyone would like to raise their hand. Yes, perhaps the gentleman at the back. Can you comment a little bit about, your strategy for private networks? Enterprise private networks. Sorry, what's the question about private enterprise networks? Enterprise private networks. Yeah, yeah, absolutely. Thank you. That's a great, great question. We had, as I said, first of all, in terms of, the market, demand, since CSP market will not be a growth market and the only way there to grow is to take share. It's a completely different situation on enterprise, markets and, and non-CSP markets, where the first driver, obviously, is web scalers, data centers, and AI, where we are making big moves. But then the other thing is really, really then the kind of enterprise networks for different verticals, where the big driver is that, that when in the past, this, for example, routing investments, were driven by IT needs, where you put your ERP and other workloads on cloud, and you need to invest in routers. Now, the big thing that is going on in industrial campuses is the digitalization and making the operational machines connected. And that puts a totally different set of requirements on the networks. The enterprise networks are actually becoming mission-critical networks in a completely way, a completely different way than before. They start to resemble some of the requirements of the CSP networks. And that's fully understandable when you all of a sudden, if you start connecting cranes in a container terminal, or moving machines in a mine, or robots, or in the future, industrial drones, and so on. It's totally different requirements on bandwidth, on quality, on latency, on end-to-end security of the network and everything. And that is really the sweet spot for us in enterprise, because we know how the CSP networks work, and we want to bring all that expertise to the mission-critical networks of enterprise. This is still a fairly small business for us. It was EUR 2 billion last year, less than 10% of our top line. We had 27% growth in the second quarter, even higher than that in the first quarter. As I have said, we want to make non-CSP revenues to be, as quickly as possible, 10% of our sales, and then after that, it needs to go to 20% and then to 30%. Time for another question. Gentleman over here. Thank you. In wireless, do you see the rebound in customer orders coming in 2024, more driven by macro sites or small cells? Where do you expect the rollout of 5G to come? Hmm. There is still, there is still a lot of macro site deployment that needs to take place. I do not want to name any carriers specifically, but even in the U.S., which is one of the most advanced 5G markets, there is still a lot to do on the macro side. And, and, and all you need to do is to just look at this number, that how many% of the 5G base stations have been upgraded to mid-band? A surprisingly large part of the 5G services or what operators call 5G services are FDD bands and dynamic spectrum sharing. Then you get the 5G logo on the phone, but the user experience is not that much better than 4G. If carriers really want to push their ARPU up so that they can monetize 5G big time, they will need mid-band capacities, and they will need Standalone Core. That will really then enable true 5G services, and I'm arguing that we have only scratched the surface of this opportunity so far. Time for another question. Hi. How much more expensive is building products in America for Buy America? What does that do to your margins? Thank you. It is more expensive, but obviously, we have calculated our business case very carefully, and we believe that this is a good investment. Remember, we are not building our own factory. We are relying on contract manufacturers. But of course, this has been incorporated into our margin plans and the subsidies, the government subsidies. They are not coming to us. They are going to whoever builds the network. But of course, our goal is that we would not take a hit in terms of margins, despite the slightly higher cost. But then how many% higher exactly that is, that is a trade secret. We can squeeze in one more question. Anyone? Maybe I'll just finish with one question. You made a lot of progress on your semis engineering. You obviously- ... really focused on a complete different way of doing that in the last few years. Can you just sort of give us an update of where you are there, and perhaps how far you think, if at all, you're behind other competitors, or have you fully caught up at this point? We have pretty much caught up, and then there are examples where I believe that we are taking a lead as well. If I take one step back and explain why we are doing all this, I mean, there was a belief some years ago that general purpose compute and CPUs would in a way rule the world for everything. That is not happening. If anything, it's going to the opposite direction at the moment. When you look at data centers, it is purpose-built compute that is ruling. When you look at AI, the same thing, I mean, look at NVIDIA. The same thing is true to things like 5G networks, where we absolutely believe that purpose-built compute for the most demanding parts of the network, like L1 processing and beamforming algorithms in the radio, gives the best power consumption, the lowest cost, and the best performance. This is why we are investing in our own silicon. We are building significant IP blocks of our own, then we are working with partners like Marvell and Broadcom. Sometimes they bring some of their IP blocks, and then they package the whole thing together. But our very clear conclusion is that for certain parts of the network, you need purpose-built compute, and a significant part of our competitiveness needs to come from a strategy where we possess significant own IP in those blocks. Maybe just as a follow-up, you know, one of, one of the ideas of using sort of Open RAN and Virtual RAN was that you could actually use a sort of general purpose compute. Would it be fair to say that the progress of those new technologies has perhaps been a bit slower than expected, and maybe that was one of the reasons, the ability to get the energy efficiency and performance, and so on? Well, these two are necessarily not mutually exclusive. I mean, as you know, we support Open RAN and Cloud RAN 100%, but the reality is that it is coming more slowly than most people thought still some years ago. I think the latest Dell'Oro forecast is that 15% of the radio market would be Open RAN in 2026, 15. The way it will most likely go is that when you take the CU and DU of the base station, which are the parts that you could put on cloud compute, you will need hardware acceleration there, which is then purpose-built compute. Otherwise, power consumption will be far too high, and performance will not be enough. Again, we are ready to deliver all alternatives that customers want. This is not a religious question for us at all, but these are some of the technical realities that that O-RAN and Cloud RAN are facing at the moment. Great! We'll wrap on that note. Thank you so much for joining us. Thanks everyone in the audience. Thank you very much. For listening in. Thank you. Thank you.