Status Update

Jun 3, 2020

Thank you, Alexandra. Welcome everyone to today's call, part of the series of calls where we where our business area heads will present current topics, new contracts and other interesting news and development. The format is straightforward. We have planned for maximum 1 hour session, and we intend to have more or less fifty-fifty presentation and Q and A. Today, we hold the second call, and it's my pleasure to introduce Peter Laurin, Head of Business Area Managed Services. Peter, please go ahead. The stage and screen are yours. Thank you, Stefan. And first of all, great to have you virtually on board for this session. Interesting times in many ways. But I would like to use the following 20 minutes to give you an update on where we stand in the Besink Air and Management Services, build on what we showed also in previous Capital Market Day, but also give a sort of a new sort of a view on what this offering is all about because we have now doubled down on a new offering that we talked about before as well on AI and automation and what this really can mean for operations of the networks of the future, meaning 5 gs and IoT. So we think that this is super exciting, and I hope you share that view after this call. And all right, if we dig in what are we talking about first. So if you go to the first slide there, Ericsson, a global leader in managed services. We have 28,000 operating networks 20 fourseven in the world. We have around 1,000,000,000 subscriptions that we manage, 700,000 sites. So we are one of the largest operators in the world if you look at it that way. What has really changed over the last, I would say, 2 to 3 years is that we have not only taken the old processes and done sort of what the operators did, we did in a centralized offshore manner, but we more or less did it the same way. What has happened in the last couple of years is that we have used much more data driven approach, really been driven by a lot of tools like open sourced as well as in house developed to be much more AI and machine learning driven. That has also led us to higher and competent lift in a different way. So we have some 300 data scientists now working on these AI and machine learning use cases, but also automating in a much broader way. I'll come back to that. But it's really a fundamental step change in how we do this business. So it's not sort of just an evolution. It's really in house. We more or less see it as a revolution, if you call it that way. All right. But let me then go to the next slide. This is a little bit to explain and share the journey. And many of you that have met me as well as Ericsson and the services area, the managed services area in particular, you know about this journey. When we started in 2017 and Borje organized Alsat segment, we also realized realized that we were not making money. If we singled out this specific area of managed services as well as MBO, network design optimization, we were not money. So the turnaround has really been about moving out contracts that were not strategic fit or not profitable and zoom in on our strategy, our simplified 300 contracts, the 42 were completed in 2018. And then we significantly automated. So we went from some 36,000 staff down to the 27, 28 that we are today. But then the second part, I would say, is the most interesting. This is really about the R and D. Services has not had its history of having a lot of R and D, but now we have invested a lot in R and D. 3 is then what I would like to talk more about today, which is how do we package this now? What is the offering? How do we scale this? And how do we sell it? We have done a lot in this space, but this is really where the battle will be fought, so to speak, in the years ahead. How do we scale our AI and machine learning? And how do we really use a federated learning, so to speak, where you have different algorithms being in one part of the world and then being sold in different parts of the world. And then those trained algorithms are coming back to the buyer, so to speak. I'll come back to that, but this is really the whole change process, which is so exciting. But nothing of this would have been possible without R and D, money and investments. So this is a slide that we showed at the investor update it it has started on the net sales how yet we will actually have a pressure on the top line because we're exiting deals. And that's sort of we aligned on the on how to exit it in 2018. But still, those exits are sort of happening most in 2019, but some of them are coming out now. And then we are also taking new business, which is great. You heard about some of them like Indusat and TDC and so on. So that's the top line. But on the operating margin side, if I start on the baseline there, the 6.4% on the operating margin that we had rolling 4 quarters in 2019, as we showed this slide, that is then being improved by delivery efficiency because we are really taking cost out of the whole chain. But what I really would like to do is, at the same time, invest. So we've said that we would reach the 5% to 8% operating income by 2020. I'm still committed to that. We still say that we can do that, and we will do that. And I think the interesting piece is that, yes, we will reach those good margin levels, but we will not sort of compromise on the R and D investments. So that's why we will continue to invest. And we are investing even in these challenging times of COVID-nineteen. And we are actually seeing that we can sort of get traction during getting traction and we we are getting traction and we are able to sort of spend the money and also get the output of it. So this is just a chart that, yes, we are delivering to what we said. If I go to the next slide, this is a bit the cost customer view. I could imagine that you meet and listen to many customers, and this is really a customer reality and this technology you're IoT, you're virtualizing your core network, you're introducing a lot of new devices, smart factories, industry use cases. This is a complex world to not only to manage and to bring to market, but also from an operational standpoint in the back end. So our goal and our value proposition is really help us simplify that. But to do that, we need to be thought leaders and experts in automation and AI because we believe that AI here is not a nice to have, it's a must have. We would not be able to cope ourselves with this. And we see that from 5 gs and other introductions just enormous, right? So we need to both filter the alarms in a different way, but also then be able to go as far as to go to what we call closed loop. So it's not only about human guided anymore. It's about that the alarms will be fixed and triggered by the platform itself. So that's complete new level of automation, so to speak. And that is what we are super committed to. So and one aspect of this is also that we don't believe that the operators, even the large ones, will have the scale to do this their own. There is no one really, as we see it, that could be the subject matter expertise and the latest data driven platforms mirror those better than we can. It might sound sort of bullish, but we believe that we have a unique position there. Not the large IT players, not our competition have the same access to data and trusted and have invested and been so focused in this area as us. So here, we see that we a very interesting dialogue with our customers. If you go to the next step, this is really about the journey that we have done because our offering to solve that customer challenge is Ericsson Operations Engine. So this Ericsson Operations is Ericsson data driven offering that is sort of the encompassing proposition to the operators of how to do everything from managed services networks to managed services IT to network design and optimization. And it really started in 2017 on the one here with it's very process driven, and that's also been a huge asset that it was quite, I mean, I say easy, but when I that's always a caveat to that. But there we use the industry processes in this space, ITIL and the other ones. But these 26 processes, we have reengineered. So that is sort of the first step of it. But then the second part of this to sort of fast forward here has really been to automate and deploy new tools and platforms and make it data driven. We have had data lakes. We now cloudified our solutions, so we can scale our tools platform. So if you are, say, an operator, if I take Telenor in Asia, then we can easily then plug them in, get them onto our tool suites for everything in service assurance and then operate their network remotely. That was quite a long process before of transformation, but now we have sort of simplified that and made it much more future proof, I would say. And even so now, if I take another example, is the cases we have in Indonesia, where we've taken the operations for Indusat. There again, we can do a lot of that transformation even in the COVID-nineteen times where travel is difficult, we actually have the, I would say, the muscle memory, if you call it that way, or the experience to do so much with our experts from remote to sort of still operate and use this new platform remotely. So that it's been a huge experience in the last couple of years, taking it from first step of process driven and then data driven. We are now aiming for about we have 300 automation use cases, 6,000 of those are then rules that are developed and reused. Because the thing is here that when we have done an automation for 1 operator, we can then scale that over the whole network because the issues that we see are quite similar. If that is a voltage dropped calls or if that is anomaly detection in a telco network, same AI and machine learning use cases can be used across. So this is our sort of reuse that is so valuable. All of this, these are the huge investments that we made, has led now to that we have 40 AI use cases and these five capabilities. So if you keep this in mind, the yellow part here, as you go to the next slide, because the data driven processes, the application platform, the people, because this is not only about software. This is really also about the subject matter expertise that we have been upskilling. So front office and back office is more or less obsolete, you can say. It's more higher sort of skilled engineers in a service desk that we call it. So handling multiple customers before it was more single island, one team doing one operator, for example, from India handling the U. K. But now it's one team in India handling multiple customers. And then automated execution and then 5th, the insights that we can make. What's the why is it sort of less bolt dropped call rates, for example, in Bangalore versus in Washington, D. C? How can we then take these learnings? So the and do that in a very data driven way. So these are the capabilities. And if I go to this slide then, the next slide, Ericsson Operations Engine and Stefan. So what is it then? When we say Ericsson Operations Engine, how do we make this simple? And I really think this is a great slide because it shows the capabilities that I just showed that we have invested in for the last 3 years comes really down to these 5 capabilities. And those capabilities then lead into 4 offerings. And we talked for many years about network management, sort of the network managed services. That's one piece. I mean, this is where we operate networks on behalf of our customers, primarily mobile, where we can do fixed as well. The enterprise services is very much linked to what Orsah talked about last week also in her session. How can we enable this enterprise service? It could be SD WAN or it could be a lot of different services where dedicated networks and so on, where we provide support and the operations of that. So it's really about making Ericsson sort of we are enabling the enterprise offerings. And then cloud and IT services as well as network design optimization. So these are the 4 offerings, super simple, very straightforward, then you can double click on each and then the offerings underneath. I will come to that. But I think it's good to keep in mind that 2 offerings and that leads to certain outcomes then for the operators. And we are purely operating focused in that sense. And we are aiming for low TCO, lowest total cost of ownership. And then we have the improved customer experience at the center of our offering, and we also then focus in on new revenue stream, efficient transformation and trusted and secure. This is, of course, super important now when we talk about data, how do we secure it, what is the all of that would be, of course, key. And we have changed our contract for a different sort of consent and different relationship. Important to keep in mind here also that we don't use end user data as such. We use network performance data and honor mine. So it's a little bit different. But of course, it's still sensitive, and it's still something that we have put a lot of focus on. And all right, if you go to the next one, because I really would like to double click on the offering because this is new and this is something we have not showed before. So if you keep in mind those 4 green boxes on the previous slide, now I'm on the Airsoft Operations engine portfolio, therefore, is these new offerings. And this should be then super simple for our sales force and for our customers to understand that in network design optimization, we have the dark gray here is a base pack. That's sort of when you buy something, a large base being an operation of a network, for example. We take over 300, 400 people or whatever size they have in the network that they want us to operate. So that goes here across network management, network managed services as well as NDO. That is the base pack. But then you add value packs on top. So the light gray here is value packs that you can buy, network planning, etcetera, etcetera. And these are new. And the good thing here, I mean, network planning, design and tuning are quite traditional, but the new part is that you can buy them as a value pack. As you sell it, you double click, and it's clearly how you price it and what the outcomes are. Same thing with the new things that are more software driven, cognitive planning and so on. So the main thing with this slide is that we have 19 offerings now that are much more data driven AI based. Same thing goes for sort of the other parts. But the main thing is that the gray one, the dark gray is base packs with clearly defined offerings. Same thing with the light gray that's a value pack. Some of the value packs, you can buy standalone and some, you have to have managed services contract in the base because it's an add on that requires that we are fully integrated tool wise and so on. But I will show others here that are not requiring that you are a managed services customer, so to speak. And that is also a benefit now as we go out because not of course, not all customers are outsourced to us, right? So we would like to have our offering to be independent, so to speak. But we are multi vendor. That's the other key part. So when you talk to other parts of Ericsson, they might be more focused on Ericsson's hardware and software and offering. We are in a true multi vendor environment. We can offer them that we operate networks that are not fully Ericsson. We want them to be part Ericsson, but they don't need to be fully Ericsson. So we can take an end to end view as an operator has us and other vendors installed. All right. So then we go to the next. And I just wanted to make some examples of this. What do we really mean when it comes to operating networks of the future, so to speak, because we have done this for many years, but what is different then? So if I take Telenor in Asia, I think that's a great example. They are, as you know, doing quite well in Asia, but also being quite focused on what they called CDC, Common Delivery Center, trying to optimize their delivery. And there, they have used us to be their partner. So we are operating all the 3 assets in these countries, Thailand, Malaysia and Myanmar, from the same place in India with the same tool suite, with the same KPIs and SLAs, then being able to cross fertilize, but also use the best of breed from Ericsson to all of them to get costs down and experience end user experience up. So this 54,000,000 subscribers, we are doing this operation for 7, and we've done that. We started with Myanmar back in 2017, and now we have continued and prolonged and then taken both Digi in Malaysia as well as DTAC in the last 2 years. And it's really about base packs and then around value packs on top of it. So you see here capacity management and optimization, network security monitoring, network promoter score, improvement in energy infrastructure operations. So these are value packs on top. And it's quite easy then also to measure this operation. You can measure the web experience. You can measure the NPS score. And it's been a very interesting journey with a very demanding customers. So it's been a great innovation journey with Telenor. And one part that I would like to then double click on is the last value pack here, which is Energy Infrastructure Operations. As you might know, Telenor has been very bullish on the carbon dioxide statement and sustainability ambitions as well as taking cost out of the energy piece. If you go to the next there, and it might come across as a busy slide, but bear with me here because I think this is super exciting. If you look at a site today, it's around 5,000 watts on the site. So I sometimes compare it with a Christmas large sort of Christmas tree with 100 light bulbs, 50 watt light bulbs shining every hour of the day, 365 days a year, and that's just one site, right? Half of that consumption is active equipment and half of it is passive. What we do here, we take that holistic view on the site, and we've done this with Telenor and it's now been a it's now a commercial offering, but we have created over the last year with the Telenor. And what it is all about is that we put about 8 to 10 sensors on a site, and then we put a site controller on the site. All that data, both on the passive and active, around 34 data streams, we shoot that up into the energy management server in the cloud. And then from there, we can shoot back actions to the site. It could be, for example, that you don't need to have all technologies, 2 gs, 3 gs and 4 gs and then in next step 5 gs on all the time. Maybe you know from the machine learning that in this very area, there is no traffic in the middle of the night between 23 or 3 or 4 or whatever time slot. So then you might not need all technologies. You keep maybe, say, 2 gs and 4 gs on. And you can switch between whatever is most efficient from a cost perspective between the grid and the batteries and so on or solar and so on. Or you might say that air condition does not need to jump on now because we have cool weather coming diesel car when you it sort of stops or shuts down. So there is tons of things we can do. And I sometimes compare it with your car and your diesel car when you it sort of stops or shuts down and goes to sleeping mode when you come to a red light, right? Similar to the site, we can be much more dynamic here. And we see now savings of up to 15% on the consumption. And we see savings of about 5 metric tons on an annual basic on the carbon footprint. So super interesting and a great opportunity for the whole industry, I would say. So more to come on this, but this is what we want to achieve and how we use data in a very practical value pack. And this we can now sell not only to the managed services customers, but to all customers. All right. If you go to the next then, Stefan, if you'd say over here we go. Yes. We don't need to go into all of these, but I just wanted to give you a little bit of a sense what we do with another large Tier 1 customer in India with 300,000,000 370,000,000 subs. So in Pan India then, we are doing the charging and expansion into the cloud. So we are very multi vendor capable, but also very capable in the MSIT and ADM domain. So this is about helping the customer into the cloud, and we have that capacity and capability as well. Okay. Docomo, I also would like to take one case on the design and optimization. DOCOMO is an interesting one because they have very high demand. And we have been together with them, and they have 8,000,000 subs in Japan, and it's a multi vendor network. And they went out on an RFQ on AI based optimization vendor for the RAN nationwide. And we won that in steep competition, and then we've been deploying our AI use cases during the year. And this is we're very proud of this deal, and it also is a testimonial to the AI capabilities that we have invested in over the last couple years. And it's very powerful, as you see here, 98% troubleshooting accuracy, 6 100,000 sell to be optimized automatically and then multi vendor KPIs, of course, and now ranking on vendor guys, of course, and now ranking on vendor selection in the trial, as I talked about. So this is a great case. It's been press release, and we feel that this is also just the beginning of something bigger when it comes to AI and machine learning in the optimization of existing networks. So these were three examples. But I also would like to not only talk about what we have done, but also talk about what we will do. So if you go to the next one, and I talked about R and D investment, but what do we invest in then? We have the cases that we do now to get a lot of our customers into the Aeroxel Operations engine. But what we put into our sort of R and D box is also around these four areas: managed algorithms, was partly what we talked about in the case, but it's also more there that we can do. AI for 5 gs is sort of quite self intuitive, but this is really about also about 5 gs slicing and how you can use data to see the problems before they occur. I will come back to that, but it's extremely powerful. And we also see that this is where operators could add more value than only connectivity. So when an operator goes and offers, for example, a full network to an industry, example, what Orsar talked a lot about, we can operate that for them and also then use data not only from that industry in that country, but from similar industries in other countries and provide additional value and also use that value and that data to see if the fault will happen before it happens. I'll come back to that. Machine reasoning, this is a lot of research that we're doing together with the Aerosome research around how can we when you don't have all the it's easy to make an assumption and a decision when you have the full data set. But how do you make a decision when you don't have all the data? That is hard. Then it becomes sort of reasoning. It becomes logic. It becomes almost like a human. So when you have different faults happening at the same time in a network, how do you make sense of all that? So that we are taking now to the next level, and that is super powerful as well. And again, this requires so much subject expertise and so much research, so it doesn't make sense for every operator to do this on their own. It's much it makes much more sense on the data we do it and then we provide it to these operators. And managed algorithm, this is really about when you have an algorithm in a use case. For example, say that you have an algorithm to do that energy consumption in Telenor in Myanmar, for example, and you really optimize that, then you can use that same data to another operator, say, in the U. S. But we don't use the data as such. We just use the weighting in the algorithm. So it's federated. And here, we use that algorithm in a marketplace. You can have multiple algorithms then trained and being better and better all the time as they go around. So it's being part of that community. And that brings me to the next page because this is a bit of an illustration of just that, managed algorithms. So it's almost I think the picture is quite intuitive in that way that these algorithms for one specific use case, it could be what I talked about the force keeping cells, tower client prediction, anomaly detection, KPI degradation or what we talked about now, the energy part. That algorithm that does that work and optimizes that work in, for example, one country then goes into the the accuracy of that algorithm could be, say, 85% as it left the first customer. But then as it comes back after being sort of trained by the others with a larger data set, it comes back with 95% accuracy. So it's really a federated learning, and we're going to put this into a marketplace where we then can offer it more or less as a smorgasbord, you can say, but it's also an an opportunity for us internally to use it between the different parts of the company. So here, we use we work very close with BGS and BIMU, the sort of the other areas as well as BTEGs. So all business areas are using the same thinking when it comes to architecture, security, privacy, life cycle management and so on. And last but least then, I would like to end on what I believe is a pretty funny, I maybe shouldn't say, as an interesting, exciting use case. So this is about the 5 gs slide thing because I get a lot of questions on what is data driven operation in 5 gs? Why is that different from a 4 gs operation? And I believe that it's very much around the 5 gs slicing. And if I take an example that we're now doing the Encore network in Canada, Encore is the government owned network where they use academia and others to test and trial 5 gs. What we have been able to do there is to significantly load test the network, and it is a 5 gs radio and core trend, and it's sort of the best of breed network. So it's a very good test environment. And in this case then, it's a simulated factory where the robot is walking, carrying hazardous material, a dangerous material. There is no human involved on the shop floor, so to speak. This is a 5 gs factory and 5 gs managed and virtual environment. So what we do then is that we have 4 slices sorry, 3 slices. 1 is the motor heuristics, so the balance of the robot. And the other one is the video streaming, which is the eyes of the robot. And the last part is sensors being just safety sensors for smoke and fire. So the safety sensors, they just need high uptime, never to be down in any way. The video streaming, that needs high uplink. They should see if there's anything in boiling or what type of material it is carrying. And the motor heuristics is really about latency. Anything below 10 milliseconds, the robot will fall. So then the operator would contract this industry with these 3 slices and with SLAs accordingly. And we would then keep track of that because you can just imagine how many different slices and SLAs that will be, right? That's one part of it and that's really industry standard, and we can help with that. But then the real value comes in the data. So what we have seen is that in the beginning, with machine learning, with about 4 months of data, we could have a couple of minutes, around 15 minutes, we could see that the robot will fall, the latency will be above 10 milliseconds. Then what we have seen is that the more data we get, the more predictions we can have. So now we're up to an hour in advance. We can see end to end. It could be the Internet link into the factory. It could be a congestion in the baseband. It could be several different things. And then if it is, for example, a congestion in that baseband in that factory on a Friday afternoon that will cause the robot to fall, we can take action in a closed loop and shift that. We can sort of shift the traffic to a node that is not congested in an automated way, right? And this we could not do before. So this is what is different with 5 gs that we can, in 5 gs, slide seeing provide different values and bring much more predictive and not reactive as we've been on 4 gs and other technologies when it comes to operations. So this is an ability to make the networks much more resilient with smart operations and AI. And on that, I would like to end and open up for questions. Thank you very much, Peter. Yes, like you said, it is now time for questions. But before we start, I have a request, and that is that you please focus your questions on Peter's areas of responsibility being managed services. The other company or group questions, IR will be happy to help you out with a little later. With that, Alexandra, would you please start the Q and A session? Thank you. Ladies and gentlemen, at this time, we will begin the question and answer session. And our first question is from Andrew Gardiner from Barclays. Please go ahead. Your line is open. Good afternoon. Thank you for taking the question. I had a high level one for you, Peter. Just as we sort of think through the current phase of business in which you find yourself where you've been exiting some of the more legacy contracts and revenue has been under pressure. And you acknowledge in one of your earlier slides that there's limited net new business, net new revenue at the moment. But as you get to the end of that phase and look to 2022 or perhaps beyond that, why wouldn't the industry challenges that you've highlighted on the technology complexity slide result in growth for your business? Presumably, given the R and D that you're putting in place, you do think that's coming. I'd just be interested in how you could frame the longer term growth opportunity. No, thanks for that question. We see that we have growth in many contracts, right? Then it's a little bit harder to see how big is that growth. We have, I think in our earlier statements, sort of showed like a 2% to 4% COGS. And that we still have that prudent outlook, if you say. But if, of course, if this becomes as powerful as we could hope, right, I mean, there's also upside in that. Then, of course, there's always price pressures, and we will, of course, be price competitive. So there are a lot of these automation savings we share with the customers, right? But I have big faith in this business. So it's here to stay as I see it. And then how big will the growth will be, it's a little bit hard to tell. Okay. But I suppose also from what you've described with the lead you think you've got with the operations engine from a competitive standpoint, the way you've described it here, it feels like you should be able to monetize it or get a better margin out of it than in the past when it was more of a managed services was more of a headcount business? And how much could you take on? How lean could you get if you have this competitive lead at the moment with the operations engine, then presumably the ultimate margin should be better? Absolutely. Absolutely. And what we said in the previous analyst sessions has also been that we believe in this business. We have gone from a sort of a turnaround from negative to positive. 2020 will be 5% to 8% on operating income. In 2022, we said that we are 8% to 10%. So we will gradually improve margins. But I think it's also important to say that we are doing that at the same time as we are doubling down on R and D for the long term. So I think that tells you a little bit about that. We are both increasing margin at the same time as we are investing. And we're not tying up any percent, the point of view, the return on capital employed is very good. Point of view, the return on capital employed is very good for us. Thank you very much. And our next question is from Jorgen Vester from Nordea. Please go ahead. Your line is open. Hey, Peter. Thanks for taking my question. I have a couple actually. One, I'd like to double click on Andrew said. So when it comes to the maybe long term profitability here and the increase, if you could elaborate a little bit where the biggest contributions are going to come from? Is it a mix shift? Is it the lower share of problem contracts? Or is it generally just better a better product offering to give us some sense where it's coming from? And my second question is, if you could talk a little bit about the emergence of private networks and how you think about that revenue opportunity, if you're thinking about going direct to enterprise on that or if you're going on behalf of operators or if it will be a mixed? Super good questions. If I start with the margin profile. When we have cleaned up the contracts that were not profitable and deals that we were not able to sort of turn around. We did that in 2017 2018 and now some of them are coming out in 2019 as well. The margin profile is quite stable on the base business. And then what we see is that when we offer these value packs, they have a higher margin profile. So you can say that we are doing a couple of things at the same time. 1 is, of course, to be more lean and lean in the base business, which is all the installed base that we have. And that is sort of, I would say, toppling along, if you call it that way, if I may say so. And then the new offerings comes with a better margin than the base. And that's really a little bit also how it should be in the new offering that we are more AIM and more software based, you can say. So in the past, as you rightly said, service has been very much a labor arbitrage. But now we are moving away from that. The labor arbitrage will be there as kind of a base and a table stakes, but the differentiator is really the AI and ML as well as the all the data driven. And it might sound sort of a little bit of a high level, but it's a lot of things that need to come to play actually do that, data lakes, cloud for the tools to be able to scale quickly and all of that. And we have taken those investments. So now we can improve margin steadily as we go forward. But we also said that this is profitability over growth has been a strategy. So let and maybe a little bit back to Andrew's question that we will take deals, but we will not take bad deals. So we feel also we are in a position of strength in that sense that we don't need to sort of take business for the sake of taking business. We need to take good business, and we know that we will provide value to the operators. That is sort of the first question. Second question, if you should repeat that question, I'm sure I answered it correctly. No, it relates to private networks like CBRS in the U. S, etcetera. Yes, yes. No, thanks. And I think also if you heard Alfa talking, the focus now on private networks is really that we can do more of an end to end, right? We similar to what we do on IoT, where we operate for Orsad. So she is the go to market, and you can see us as the engine room. In Romania, we're operating all her IoT connections. So the DCP or the HDRI Connected Platform, the underlying engine of so to speak, we do from Romania. And that is the base for the IT offering. When it comes to the private networks and the dedicated networks and Industry Connect and so on, all of that space, that would be more of a cookie cutter to the through the operators. It's also you talked about the channels there. We are focused on through the operators, right, then we will be the engine house for that as well. So that is, of course, an offering that is developing as we speak. But we believe that it will be very much a services play in that. So we are working together with this enterprise this enterprise team. And that's also why you saw in the Ericsson Operations engine that we have this new offering called enterprise. And that is it's really also us enabling the larger company. Is that a 2021 opportunity or 2022 or even further out? What are you thinking? Yes. We have one of the testing deals actually. I mean, some of them are publicly announced with Deutsche Telekom and so on where we are doing this. So commercially, it's out there. Then to your question, which I think is fair, when do we really scale this and when do we see this to be significant? I would say that would be probably 2022, 2023 where we see big volumes coming in, but the foundations is set right now. Okay, Jorgen. Okay, happy with that. And our next question is from Hamid Kalavandi from Citigroup. Amit Hazhindani from Citi, and thanks for the opportunity and the detailed overview. Two questions, if I may, both sort of big picture. The first question relates to what we have seen out there, for example, in the U. S. With one of the operators in sourcing a portion of the business from rushing to Sprint and T Mobile. So I just wanted to get your thoughts on how do you think this in sourcing versus outsourcing dynamic plays out more broadly as it relates to managed services? What are the key considerations? And how would you think about it or help us think about it? And second question, if I may, there's also a lot of talk of longer term moving to virtual networks and open networks. From your perspective, what would that imply for the opportunity for managed services on a longer term view with more re RAN and open networks? Does that diminish it? Does that accelerate it? What are the puts and takes there, please? Thank you. Excellent question. And this is not sort of the macro trends in this slide. Outsourced telecom networks today is about 15% of the global telecom networks that are outsourced. So it's a quite limited amount. So we see that this is still a lot of business to take. And then you can say, will that increase or decrease? We believe that it will be continuous, that with the networks being more complex than less complex, that the operators then have a choice. Do we continue to do this? Do we invest in our own competence to manage these complex networks? We invest in the tools required. We invest in the AI and ML required or we let someone else do it or we find a hybrid solution where certain part we believe is core and others we believe it is not core. And in the U. S, specifically then, if you talked about the T Mobile and Sprint coming together, right? I mean, let's see how they play their cards there, right? But that will be, of course, and we are a big provider to them, both to T Mobile and to Sprint, right? So that will be interesting to see how they now do these strategic choices, right? But overall, we see that this is an opportunity going forward and that this will be more of this, not less of this. And that a little bit, we're going to see what's happening in the IT space, also happening in the telecom space. The level of outsourcing is significantly higher, as you know, in the IT space than in the telecom space. But we believe that it will be more. And also, as we see this being mission critical networks that you really and I think also COVID-nineteen also showed that, that it's no doubt how vital these networks are for society, then some will choose to do this in house, but also some will say that this is some where we need a partner to support us and also to be that second spine, so to speak, or that competence pool. And it might not be for all, but it could be for very critical parts that you take up the network, right? So yes, I hope that answered your question. Sorry, it was a long answer to a short question. And with regards to your thoughts on how open networks and virtual networks imply for managed services longer term? That's also very interesting. We get a lot of business now on operating and managing for example in the MHIT and ADM space, both managing older applications, asking for help to managing workloads into the cloud, both IT workloads primarily into the cloud. And then when it comes to virtualization of the core, it's a lot of complexity there to manage. So for our partners since M and 19 ADM, we have been able to add a lot of value on helping out in that virtualization from an operational point of view, then BDGS can help a lot from the solutions point of view, right? So I believe that's also an opportunity as we go forward because from an operational point of view, it's not necessarily so that it becomes less complex just because you use sort of off the shelf hardware or that you virtualize and so on. It could actually be quite complex. And that we also see in the system integration, and Jan Karlsson will probably talk more about that. That's also a complexity in that in itself, right? Okay. And just a quick clarification on the last answer, if I may. So if we move to open networks, you said the complexity exists, but because it's open networks, would that allow someone else to come and do managed services on top of equipment provided by Ericsson more effectively or less effectively? I don't really have an opinion on that. I don't think it will be I wouldn't have a view on that really. We also see that quite limited, right? I mean, this business is right now at least very much dominated by the large infrastructure vendors, right? And I don't see that changing even in an open situation because there will be so much telecom equipment in the end to end network anyway, right, even if you would have certain nodes being open, so to speak. Got it. Thank you very much. Yes. And we are getting close to the hour. So, Alexandra, could you please now let's have the final the last question. Thank you. And our last question is from Frank Mao from DNB. DNB. So if I may, just a question of how you manage customers' wishes for exclusivity in the markets, say, take the Telenor example. When Telenor has done this innovation journey with you in Asia, they might, of course, want to keep the gains in those countries and probably would like you not to work for their competition in those markets. How do you manage that? Do you grant exclusivity, if so, for a few years? Yes, some comments on that would be great. Thank you. Thank you for that. And thanks for all of you for listening. A great hour, and I hope you're as excited about this topic as we are. But when it comes to that exclusivity question specifically, we don't go into exclusivity like that. We it probably has been I can't recall, I mean, I've been in this business for a long time and it's probably been something like that in the past possibly. But as we have now, I don't see that we have exclusivity anywhere actually. That would be really a contra our philosophy of 1 to sort of one to many, so to speak. So no, we don't do that, not even on the local markets, right? And even so, if you take just the field side, right? I mean, the field operation that we do when we come to an end to end operation, it's great if we can serve multiple customers in one country like we do in many of the geographies. So no, we don't do exclusivity. And actually, I don't believe it's in the interest of the customer either. It's really taking that cost down for all of us with multi skilled labor, I think, is good. Okay. Thank you. All right. Okay. And with that, we then conclude today's call. Thank you very much, Peter, and I trust everyone on the call found it both productive and interesting. Until next time, which is Tuesday, June 9, when we have a session with Friedertek Jeling, Head of Networks and Jon Karlsson, Head of Digital Services. Stay healthy and well, and goodbye for now. Thank you.