Liberty Global Ltd. (LBTYA) Earnings Call Transcript & Summary

Hello everyone, and welcome again to the third edition of our CTO Symposium. Before we start, I would like to flag that for important disclosures, please see the Morgan Stanley research disclosure website at www. morganstanley.com/researchdisclosures.

Right. Our next session today is with Enrique Rodriguez, Executive Vice President and Chief Technology Officer at Liberty Global. Hi, Enrique.

Good afternoon. Thanks for having me back here.

It's a pleasure. Thank you. Enrique, for joining the CTO Symposium for the second year in a row, I really enjoyed our conversation last year, as I told you earlier. And yes, looking to get an update on a long list of topics. There's a lot of ground for us to cover today, and maybe a good way to start the conversation with a big picture. Could you talk about your top priorities at the moment?

Yes. I mean, as usual, with this kind of business, there are some priorities that continue year after year, and there's some new ones on board keeps the job interesting, if you will. I always have to start with our customer experience. And in the case of Liberty Global, as you know, we participate in the fixed and the mobile business. And so in general, the focus on customer experience tends to be about providing continuing to provide services that our consumers feel are best-in-class in the given market. When it comes to fixed, I think it's been already quite a few years where the way we really think about our business is not so much that we're providing fixed connectivity into a home. It's more about providing Wi-Fi in the home. And I think the consumer has gotten to be quite sophisticated on the way that they think about connectivity, also very demanding with respect to both the quality and the reliability of connectivity. So continued performance of Wi-Fi in the home, I think will remain a very important topic in this industry for a long time, and certainly a very important part of my objectives. Second part that I would highlight is the expansion of our networks, okay? And so we continue expanding our 5G coverage across the different countries. And probably more importantly, because it actually translates into higher levels of investment, is the evolution of our fixed network. As you probably know, we have a pretty aggressive evolution in a number of the countries in which we are they either translated in our complete network eventually into fiber or having a mixed environment where we can offer both fiber and HFC, given the conditions and what we feel is the best way to solve the requirements and the desires of our consumers. So fixed network expansion continues to be a very, very important subject for us. And then I think I'm guessing you will get this from just about every CTO and maybe even from every CXO is a new focus on AI, right? And trying to understand what's real, what's fiction. Trying to understand how to best prioritize the impact that AI will have in our business. How you make sure that we keep all the -- all the different concerns about AI and the different opportunities that we see in a properly balanced and trying to come up with an actual implementation plan against AI. So I would say those are the 3 big buckets when you look at my objectives.

That's very clear. Thank you very much. And effectively, we are very keen to focus on AI in this session, given the importance of the topic and evolutions that we are looking at. So we'll definitely come back to AI later in the session. You started by the customer experience, which I guess is sort of the endgame of this all. Could you discuss maybe what are the main network differentiators at the moment? What do customers value the most?

Well, I mean, I think as an industry, we come from a route where usually differentiation came down to speed, okay? And that was certainly true in the fixed networks. And it was also very true in the mobile networks. If you think about 3G, 3G it really was the first time that we could bring data onto mobile. You think about 4G, 4G really maintain a usable, right? I mean in a big sort of way, where you could expect every mobile experience was really powered by data. And we saw a significant increase in the utilization from a consumer point of view, a significant increase actually in the satisfaction as well, because people felt that the mobile really had gone out beyond a communications device, and it really became a personal information device inside the home for many, many years. It was about tracking the megabits and then eventually the gigabits and really making sure that you have the best speed in the market. Of course, quality, what's important, it always has been and it always will be. But the key differentiator was speed. And as you think about consumer speeds getting into the hundreds of megabits per second, getting into 500 megabit per second, getting into 1 gigabit per second. We have almost 95% coverage now in our fixed networks, where a consumer can choose a 1 gigabit per second service, okay? And so speed is not really the only differentiator anymore. More and more is the quality of that connection. People are starting to worry about things like latency. Things like, of course, return upload speed, upstream speed became quite important first during pandemic because we all started using video. And then, of course, as consumers, once we get used to a particular experience, we never want to go back, right? So it has become more thorough and it's less about just the raw speed. It's more about the complete experience that we can provide on that. On the mobile side, 5G of course, has got a lot of attention from a consumer point of view. But the reality is, as an industry, we're in a pretty tough situation, right? Because all of us have had to invest heavily on 5G. There's an expectation from the consumer that if you provide a best-in-class service, it must include 5G. Yet the actual killer applications on 5G, the ones that would make the consumer want to pay more for services, have really not appeared yet. And so I think we have a little bit of a conundrum there where we're all -- by we, I mean, all operators feel the pressure to really drive 5G penetration, especially in certain countries where we have not yet seen the benefits from our return to the capital, but we'll continue, obviously, investing on that side. But it's become a lot more subtle. I think it used to be a lot easier to explain to our customers why they should upgrade to our newest offering because speed was such a big determinant. I think now things are a little bit more subtle than that.

Right. When thinking of NPS scores, I guess, network differentiation is very, very important. Could you elaborate a little bit on this, maybe with some numbers, if you are able to share any? How important is the network differentiation or network quality contribution to the overall NPS scores? And in the case of Liberty Global, it would be also helpful to discuss how this contribution of the network quality and differentiation has been to the NPS scores has been evolving over time.

Great question. We track NPS of course, and we track the influence of different elements of our service onto those NPS scores. So maybe I've used a couple of examples that let me be more quantitative about the answer, if you will. When we introduce IP-only entertainment services, meaning entertainment services that only rely on IP, a small clear set-top box, very low power consumption, very quick start-up time, very fast interaction to the consumer, full integration of voice search, much deeper catalog of content, much better ability to search and select the content that you wanted to see, we saw an increment of over 15 points positive impact on NPS when we look at the complete satisfaction of the customer, right? So not just measuring entertainment, but looking at the complete satisfaction with us as a provider. We saw that initially in Poland. Of course, since then, as you know, we have sold our Poland operations, but we saw the same thing as we introduced the same services in Ireland, in Netherlands, in the U.K., where we consistently get better NPS from consumers that are able to enjoy these more modern experiences, right? So you don't have to deal with the older, bigger set-top boxes with recording in a hard drive. Now everything is in the cloud. It sounds pretty straightforward. But it actually has a pretty significant impact in ultimately how the consumer thinks about your services and of course, the NPS scores that come with that. For Wi-Fi, which is 1 of the fundamental services that we can provide to our customers, clearly, it's all about reliability. I think what we have learn as consumers is that Wi-Fi is expected to be everywhere, right? I mean you go to a coffee shop, and you feel frustrated not just about the quality of the coffee, but the quality of the Wi-Fi that you get while you're working over there. And so Wi-Fi has become something that we expect happens all the time, and it happens not just all the time, but it happens at a great quality all the time. And so a significant amount of focus and investment has gone into making sure that we can provide the best possible Wi-Fi. We have had now a multiyear, very successful collaboration with Plume, where we use both their software, their intelligence. By the way, it's 1 of the early use cases for AI, where Plume and us have been using AI for a long, long time to make sure that we can optimize these Wi-Fi connections, but through a very, let's call it, high-tech solution, both on the hardware and on the software side, we're able to significantly improve the Wi-Fi performance in the home. So it's -- I would say the battle for NPS, just like the battle for differentiation, it becomes more difficult every year because everybody's services continue to improve. But thanks to a combination of technology improvements like Plume as well as service improvements in the way that we interact and service our customers, I think we're able to keep up with that NPS war, if you will.

Very good. So I wanted to discuss the different network upgrade projects that you have when we talk evolutions, both in fixed and mobile, started by fixed -- and certain by DOCSIS 4 I recall last year, we discussed the benefits of DOCSIS 4 et cetera, it's evolution, time-to-market, prices, et cetera. So really keen to have an update on DOCSIS 4, a year after that. So perhaps starting with the basics in particular for the audience for investors who are not familiar with DOCSIS 4. Could you explain, please, what are the changes that DOCSIS 4 will bring?

Absolutely. So ultimately, it's about providing an improved level of service, both in terms of speed as well as in terms of latency and reliability of the network. Ultimately, everything is around that. Since you and I spoke a year ago, 1 of the additional elements we have introduced is the ability to take our existing DOCSIS 3.1 networks to a speed of 2.5 gigabit per second. So as I mentioned earlier on the discussion today, we basically have almost 100% coverage now of 1 gigabit networks. And in certain countries like Switzerland, we're going to take advantage of the ability to go over 2.2 or 2.5 gigabit per second even before we deploy DOCSIS 4 in the network. So it's giving us more and more flexibility, which ultimately, I consider 1 of the key contributions of the Liberty Tech organization is to make sure that we can make the best possible use of all these investments that we've been making over the years, right? So even before you get DOCSIS 4, we're going to see significant improvements to 2.5 gigabit per second. Now with DOCSIS 4, we have the ability to go all the way to 10 gigabit per second and provide a service, for example, 10 gigabit downstream, 1 gigabit, 2 gigabits, maybe even 4 gigabits upstream, so very significant improvements on the quality. Or in certain cases, we will be able to provide a symmetrical service where we have 8 gigabit per second, both downstream and upstream. And we will be making those decisions, let's say, market by market and case-by-case basis. But fundamentally, the #1 thing that DOCSIS 4 is going to let us do is to get to 10 gigabit per second without a significant upgrade of the underlying plan, right? We can still use the analog RF network that we have in place. In some cases, we may provide an upgrade, but that's kind of more of a natural evolution, if you will. And more importantly, we're going to be able to reuse the significant fiber deep network that we have been building on these HFC networks over quite a few years. So DOCSIS 4 is a significant step forward for the industry. The good news about the DOCSIS 4 development story is that it continues to be exactly as it was planned, right? So the dates that we have in mind today are the same dates that we had in mind a year ago. So we're quite happy to see the technology continue to develop. Maybe 1 of the things that we talked about last year, I remember we talked about the fact that there were 2 potential implementations of DOCSIS 4, something that is called FDX, which is basically providing a symmetrical up and down service level to the end consumer and then extended spectrum, or ESD, as we call it, which basically allows higher downstream rates compared to the upstream rates. The good news is that the technology has significantly unified, and now we're seeing chipset and we're seeing software stacks that are able to deal with both solutions at the same time. So we think that, in fact, the evolution of DOCSIS 4 looks more green today than it did even a year ago. So we're quite happy about that.

Well, that's good to hear. And I know that you've been conducting some trials in the Netherlands. Any key takeaways from these trials or anything interesting you would like to highlight from these trials?

The best news about the Netherlands trial is that there was no news. Okay? It actually worked exactly as expected. We trialed it in a live network. So this was not just a laboratory exercise. We trialed it with our technology partners, external partners and the technology side. So I'm happy to report everybody slept very well after the demo, no big problems. Of course, there's still a lot of work to do before we can say this is widely deployed. But from a technical point of view, I believe that the partners in the industry have done their job, have solved the big technical problems, and it's about leading the development work and then started more the deployment part.

Very good. And in terms of the cost of deployment, could you share a little bit of, or give us a range of how much should we think about in terms of the cost to deploy DOCSIS 4? And also, how this cost varies with density from a region to another or from even actually a country to another country. What are the key elements we should keep in mind when thinking of the cost of deployment?

So the main difference in deployment costs for DOCSIS 4 is when you actually have to go and touch the RF network or not. So in certain countries, given the topology and the investments that we made in prior years, the ability to overlay DOCSIS 4 is significantly better than other countries where we may be more restricted by the RF spectrum. We may be more restricted by the capabilities of the RF plan. So it's really difficult to give a number. And frankly, I've tried a couple of times and gotten in trouble with it because there's so many disclaimers onto any number that you use and the variation can be quite wide, right? So you hear reports where people talk about EUR 200, EUR 300 per house passed on DOCSIS 4. And of course, those are quite reachable. But you also hear situations where people are talking about EUR 800,000. And in certain conditions, that is also true. So I apologize, as it's not really an answer, but this 1 has been so difficult. So what we have done in the past, which in this case has been only on fiber, not yet on DOCSIS 4, is once we take commitment to deploy in a certain area, of course, now we understand the exact conditions of that area, and then we're able to provide some more external guidance as to what the cost is going to be, right? So as an example, when we talk about deployments of fiber in the U.K., we've been quite clear that we can deploy in new -- in greenfield, completely new deployments of XGS-PON in the order of GBP 600 to GBP 650 per house per home pass. And we can upgrade our HFC network, what we've said in the past and we still hold to those objectives, is that when you look at the complete upgrade of HFC network in the U.K., which will happen between now and 2028, we already started in 2023. When you look at that complete upgrade, the average cost on a home pass basis would be at or below GBP 100. And so again, once we've defined a particular program, it's a lot easier, but I wanted to give you those 2 examples just to compare GBP 100 pounds and GBP 600 pounds in the same country, getting to the same XGS-PON. Obviously, the big differences is starting condition. So extremely difficult to give you a number or even a range of numbers for DOCSIS 4 until we're ready to announce deployment in a particular country.

But that's helpful color. Thank you so much, Enrique. We have a question from the audience actually about DOCSIS 4 [indiscernible] in the conversation. So the question is about the energy efficiency or energy cost differential between DOCSIS 3.1 versus DOCSIS 4.0 versus fiber.

Yes. Well, I mean, the punchline is, fiber is significantly more cost efficient than both DOCSIS 3.1 and DOCSIS 4.0. We believe we're going to be able to deploy DOCSIS 4.0 without fundamentally changing the energy footprint, and we're working very hard. As you know, energy consumption is extremely important not only from a cost perspective, but also from our corporate responsibility point of view. And so that takes a lot of focus. But without a question, fiber can be materially more cost effective from an energy point of view than either flavor of HFC. Now having said that, 1 of the interesting aspects of AI is that we're seeing significant improvements when we can overlay AI into our power management. And we're seeing that across the board. We've seen that on mobile networks. We're seeing that on fixed networks. Of course, that's only POCs, proof of concept stage right now, but I'm actually quite optimistic as to what AI is going to be able to do in our ability to manage power into the networks. Extremely important subject both for the corporate responsibility offices in the company as well as the CFOs that are watching those energy bills significant increase over the last few years and want to get that managed back down. But without a question, fiber significantly more energy efficient than HFC.

Very good. Another question from the audience about DOCSIS 4 versus fiber. So DOCSIS 4 appears to be a good technological alternative to fiber. But how you overcome the perception issue in the customer mind? When do you think that fiber is better?

Well, first of all, the perception issue is real, and now we have significant more data to understand it in a more subtle way. What we have seen is that customers that receive good quality service on HFC, their tendency to move to fiber is really quite small. And that the reality is, what triggers a customer into looking at alternatives is a lack of satisfaction with the current provider, okay? So it sounds very obvious, but it's an important reminder when we start to see -- when we see the data that what we really need to focus on in those areas where we don't get upgrade to fiber is to make sure that we're providing a high-quality service over HFC. And we feel very comfortable that when you provide a high-quality service, when you provide the benefits of fixed mobile convergence, when you provide the improvements both from an economic perspective as well as from a total experience perspective, we're quite comfortable that we can compete, if you will, against a fiber provider in the same areas.

Thank you very much..

I'll give you -- sorry to interrupt. I'll give you just 1 anecdotal point on this, which is quite relevant to where the industry is today. If you look at the U.K., obviously, over the last few years, you have seen a big proliferation of so-called AltNets, people that were going to build fiber and have built fiber on top of the existing networks, right, whether those existing networks were HFC or copper or fiber. And what you have seen after quite a few years of construction is that their ability to get to homes passed is actually been generally okay. As long as you have the money, the funding, you can go and build. I mean that's not a big technological breakthrough. Where you see them fall significantly short from their expectations is in the actual penetration, getting customers to actually use the network, even though in many cases, they've gone in a very, very competitive price situation. So I think, again, it's not black and white. And again, I don't want to mislead here. If I could choose today, I would rather have a fiber network than an HFC network. However, with the HFC networks that we have deployed, the quality of the equipment that we have deployed the quality of service that we can offer, we feel quite confident that we can keep customers on these networks properly satisfied.

Okay. And time-wise, so we stay with the same time line then discussed previously, 2025? Is the year...

For what?

For DOCSIS 4?

Yes, definitely. We will start to see the initial deployments of DOCSIS 4 in 2025.

Okay. Very good. So as you were talking about DOCSIS 4 and fiber, maybe let's start talking about the different markets. But before we dig into each 1 of them, please, Enrique, could you give us an overview of the fixed network evolution strategy on the 4 main markets?

Of course, I'll start with Ireland. In Ireland, we've announced publicly that we will upgrade the complete network to fiber. The Virgin Media Ireland team has done a great job starting that program in 2023. It's up and running. It's, let me call it, the normal kinds of things that you find with a large scale program. And so we're quite happy with the performance of the team, the performance of the operate so far. We are basically meeting all our metrics when it comes to cost and deployment speed, so I'm very, very happy. It so happens, I'll be in Dublin tomorrow, going through a review with the team there, but I expect we will continue to see very good performance in the Irish market. When you look at the U.K., we have basically 2 upgrade paths to fiber. One is a company that we have set up, nexfibre, in which ourselves, Telefonica and InfraVia are joint investors, and we're looking at all the greenfield, the new build of fiber in the U.K. under this umbrella of nexfibre. Again, we're meeting all our metrics in terms of deployment speed and in terms of deployment costs. So frankly, I don't expect any significant issue with respect to that. We also have what we call [ fiber up, ] which is the progressive upgrade of our HFC network in the U.K. to a full fiber network. The only thing we have said about that upgrade is that we will complete it between now and 2028. We're quite comfortable with it. We're looking at the cost of the initial deployments, and we feel very, very good about meeting the overall metrics that we have communicated to the market before. So quite happy about that. If I look at Belgium, I happen to be sitting today in our Telenet offices. If we look at Belgium, we've announced the separation of Wyre. Now the company is called Wyre, our NetCo separation in partnership with Fluvius here in Belgium, separation from Telenet. We continue to execute against that. And as part of that execution, we are going to now use Wyre, this new company as the vehicle in which we're going to deploy fiber. As you know, in the Belgian market, and we had some announcements over the last few months. We're very fortunate to have a great wholesale customer, Orange is a big wholesale customer of ours. And so obviously, because of that, it provides us additional incentives to continue evolving the network in Belgium. So quite happy about that. If you look at Switzerland, at Switzerland, we had announced that we will use fiber that is available to us on the wholesale market, and we intend to continue doing that. In addition to that, we have plans now to deploy 2.5 gigabit per second service on our existing HFC 3.1 network, which will give us even more flexibility in the Swiss market. So we feel quite comfortable with that. And then finally, in the Netherlands market, as we've said before, we believe DOCSIS 4 is going to be 1 of the key elements on how we will approach the Dutch market. However, we have not announced any specific deployment plans or timetables. We're still working with our partners at Vodafone. And of course, with VodafoneZiggo, to make sure that we have the most robust plan that we can build in terms of upgrading the Netherlands plan. So I feel quite good about it. I think that I remember talking a couple of years ago and saying that our solution from a Liberty Global perspective is actually going to be a country-by-country solution. In reality, it's not only country by country, but in some cases, like Belgium is literally a region-by-region solution, right? Because you're trying to optimize what these programs that are very, very capital-intensive. And you find that not a single solution is going to be applicable in every single market.

Excellence. Which is a very good transition, actually, to my next question about co-investment. To optimize CapEx, would you consider co-investing with competitors on FTTH to enhance the returns? And if so, what should be your thinking or framework? What are the criteria's you will be looking at to take that decision?

Well, first of all, I think 1 of the differentiators of Liberty Global over many years, has been our ability to think about investments in a completely pragmatic way, right? I mean whenever we've seen an opportunity to increase value to our shareholders, by buying, by selling, by co-investing, by combining, I think we have every 1 of those plays have been exercised in the last 5 years. So I don't think it's just a headline when we say we're open to look at every opportunity, but we are open to look at every opportunity. I'll use the U.K. as an example. We have co-invested with Telefonica when we merged Virgin Media O2, then we further co-invested with InfraVia when we decided to create a separate company to drive the greenfield deployment of fiber, and that's going fantastic. I think we might be the only ones that have pulled anything similar to that over the last 12 months, 18 months. That's been quite important. And I think your point -- sorry, and we announced 10 days ago that we acquired 1 of these AltNets, the first 1 that we have acquired in the market. So I think that we continue to show a willingness and ability to look at every transaction and consider it on the value that it can bring to our shareholders. So we are open for business in just about every option, including co-investment opportunities.

And when thinking of investing with competitors, like, for instance, as you are in Belgium, that could be an example. Co-investing in Telenet, [indiscernible]. So the question is, which comes a lot in investors' conversation, but also from the management teams, is about whether it makes sense to avoid duplications and overbuild it at some point. So how are you to that conversation?

I will use a very specific example, not Belgium in this case, but Sunrise in Switzerland, where we have a very successful collaboration with Salt, which is a competitor in the market, where we co-invest in certain -- we share is better than co-invest in this particular case. We share certain infrastructure on the mobile side to make sure that both of us benefit from investment that each of us is making in actually improving that customer experience that I was talking about at the beginning of this discussion. So we have a very open view as to what it means to deploy capital and to sharpen the definition of that capital deployment to the needs of that particular investment. So I think we have demonstrated over the years, maybe the only company in the industry that has actually demonstrated over the years that we're open to any one of these scenarios.

Very good. And specifically on fiber co-investment, are there any markets where you see more opportunities than others?

Well, I think -- so first of all, let me be super clear, I'm not talking about any 1 market specifically. But the higher the capital requirements, the more benefits you see to co-investment. I mean it seems pretty obvious, but you really see that, right? So for example, in the case of the U.K., when you were looking at multibillion pounds of investment in building fiber in greenfields, then it's a natural place for co-investment. If you look at something like the upgrade in the U.K. where the capital intensity was not nearly as high, then it's a little easier to think about doing it on your own. But once again, it's a case-by-case basis. Unfortunately, there's no like -- very precisely defined formula that says as long as these 4 attributes happen and then co-investment makes sense. We really look at it on a case-by-case basis.

Very good. I wanted to follow up on Switzerland. So the access to fiber today, to FTTH is addressed with the wholesale agreement that Sunrise had with Swisscom. So what is the long-term plan for the network? Previously in our conversations, we talked about DOCSIS 4. Is it a mix of DOCSIS 4 on the Sunrise/UPC network and then complemented with FTTH through wholesale? Is it something else? I'm very keen, ready to project ourselves 10 years from now and to understand how the network picture will look like in Switzerland.

Well, first of all, within the next 5 to 10 years, you're going to see deployment of 2.5 gigabit per second on the DOCSIS 3.1 network. So we had already a mixed situation. I think we're making the mix even more complicated or have more options, if you will, by introducing a 2.5 gigabit per second, which brings a significant improvement in customer experience and service without the level of investment that a DOCSIS 4 upgrade would require. Does not mean that we're not going to deploy DOCSIS 4, but we're going to be quite surgical about where are the places where it makes sense. It makes sense both from an investment perspective and from a user experience perspective to deploy DOCSIS 4. And of course, we intend to make use of the agreement that we have with Swisscom that makes wholesale buy of fiber available to our Sunrise business. So it's really hard to predict what the mix is going to be. Switzerland, to some degree, may turn out to be the most dynamic market where we would really look at it at a small region-by-region basis and deciding what is the right investment strategy for that particular zone.

So while thinking of the next 10 years, again, I wonder whether we would see DOCSIS 4 sooner and bigger in the Netherlands versus Switzerland. Is this a fair picture? Or...

I wish I actually knew the answer, to be honest with you. I think that we're going to be defining that answer over the next 12 to 24 months. And a lot will depend on the evolution of the Dutch market. A lot will depend on the evolution of the Swiss market. We -- I don't spend any time comparing the speed of deployment between 1 country and another country. What we spend a lot of time is to make sure that the learnings of deployment in every country benefit all the OpCos in the Liberty Global Group. So as an example, the fact that we're deploying fiber so quickly in the U.K. provides benefits to deployments of fiber in Belgium. But we don't look at -- there's no like secret metric where we look at the speed of deployment of 1 country versus another one. We really take a very pragmatic approach and try to look at the results in that particular country.

That makes sense. And moving to the U.K. So we discussed nexfibre earlier. Could you give us update of the fiber rollout, not only on the greenfield areas for nexfibre, but also in the brownfield area where you are overall laying the DOCSIS network of Virgin Media O2 with fiber. And so that's the first part. If you can give us a bit of update, please? And then the second part is, you mentioned earlier, buying fiber asset. How do you decide between building and buying today in the U.K.?

Okay. To your first part of the question, the project we used to call Mustang, now it's called Fiber Up. That's our -- Mustang was our code name, if you will. Fiber Up is our real name. And I can tell you, our early metrics on Fiber Up exactly as planned. What we're seeing is that in most cases, the reutilization of ducts which is what allows us to make this upgrade in a very economical way compared to building greenfield, if you will, the utilization is actually what we expected, as a general statement, and I think that's normally what you find with engineering and operations teams. We are seeing the conditions of the existing plan to be slightly better than we expected. Okay? Now it's very early in the game. We still get probably a good 5 years to go in this upgrade, but the early indications are actually quite good. We're getting, with the early deployments, we're able to optimize, for example, what split ratios you're going to use, right? So when you go and touch the network, you have to assume certain utilization. So what utilization you use is very important decision because you could end up assuming very fast utilization and spend a lot in the network that you may or may not use later. Or you may assume very low utilization, and it sounds great because you're spending less money, but then very quickly have to come back and touch up the network. So this optimization and once again is a great example where AI is helping us to look at it in a much more granular level to use actual deployment data to then feed back into the planning process. So the planning process today on this upgrade is significantly smarter. And in this case, it's not just a little catchy word, but you're actually putting intelligence into selecting where you're going to upgrade first, where you're going to go and market it and the pace at which you're going to be doing that. So every indicator, frankly, is great. Jeanie and her team are doing a fantastic job at VMO2. The management team is very focused on making sure that we continue not only deploying but selling against this deployment. So it's been really, really, really quite good.

So moving now to mobile networks because we have -- you have a lot of mobile assets across all countries now. When talking about 5G with CTOs this year, also to a certain extent last year, we feel that the industry is going to 5G with a sort of pragmatic view on the investments versus really, the importance of the monetization. The killer apps are not clear at the moment. Also, the monetization is an important element. So I would love to have your views on 5G overall. What are your expectations in terms of a killer app that you see or predicts, et cetera. Internal, but also more specifically to Liberty Global, if you could discuss what is Liberty Global's 5G strategy today?

Well, I mean, first of all, we don't foresee a killer app from a consumer point of view to come up and save the day of 5G. 5G has been a tough story for the industry. 5G has represented significant investments up to now, and will continue to represent significant investments ahead of us until we get to a point where 5G is deployed everywhere. When you look at the practical aspects of deploying 5G, what we're finding now is that in some cases, deploying 5G is more about serving the demands of bandwidth and capacity than it is necessarily about 5G, per se. And in some cases, we are finding it is more economical for us to move traffic to 5G and provide better service to the consumer instead of upgrading the 4G network and just getting more deployment capacity. So I think we're getting to that point where 5G is a little bit more normal, if you will, from a technical deployment point of view and it's not waiting for that amazing killer app that nobody seems to know where it's going to come from. Now having said that, I'm very optimistic in some of the B2B cases that 5G enables. Specifically deployment of 5G stand-alone, deployment of things like mobile, private networks. That opens significant avenues of new revenue. Not yet here, but I'm quite optimistic as to what we're seeing. I'll mention 2 examples. In Sunrise, we basically ran a proof of concept where we're using mobile private networks and using what's called NEC, multi network edge compute, to bring significant processing capabilities, processing power into the given location and supplementing that with fantastic connectivity, we get some use cases. In the case of Switzerland, we have basically a store where you can walk in, pick up any item. You have the network recognizes who you are, obviously, taking care of all your privacy regulations that everybody expects to meet and you're able to use that computing power in the edge to significantly improve both the end consumer experience and the efficiency at which you can run a retail store. We have another proof of concept here in Belgium with the Port of Antwerp, where we're using mobile private networks to provide a higher security, higher reliability network for very high-value operations such as bringing ships in and out of the port. So I'm starting to see some optimistic views on how B2B our B2B business is going to be improved by more deployment of 5G. But I'm still cautious on that. It's difficult to see the time line. It's difficult to see the true magnitude of what those business opportunities are going to be.

Okay. So moving to AI , which is an important topic for the networks. AI is not new, obviously. It's already used in networks in a number of ways. But it looks like the future could be completely different from today in terms of its use cases and it can be a big transformation for the network. So really keen to understand how AI can transform telco networks today. But also, if you could help us understand the cost opportunity in here in terms of energy costs, in terms of network cost overall, whether we're looking at significant savings potentially from AI or not?

Well, as you say, AI is not new. However, if you read a newspaper, you think that AI got invented with ChatGPT last year, right? So yes, I think the way we're thinking about AI is 3 or 4 basic categories where AI is going to impact Liberty Global's business, okay? All of them are important. But I'll start with the 1 that has the most sex appeal in the market today, which is the customer-facing AI. Generally, that's the place where you're going to see GenAI, the capabilities of these new large language models to significantly improve the ability, the quality and the efficiency with which we interact with end consumers. And that could be at call centers. That could be through digital experiences, online experiences, through apps. And the fundamental difference is that we'll be able to use, in a very cost-effective way, significant amount of data that today exists already in our networks, exists already in our customer databases. But today, they rely on humans being able to connect A and B, whereas with GenAI, we're going to be able to rely on the systems in optimizing that for every single interaction for every single consumer. So we're quite excited about the capabilities there. But obviously, that is the part of AI that is most new, okay? It also is a part of AI that comes with some risk. And we're all going to be cautious about deploying it. Every time that you deploy something that you're using to communicate to end consumers or to interact with end consumers, obviously, any concerns about data privacy, any concerns about security, any concerns about everything that has to do with ethics become even more increased. So we'll be cautious, but we're making some progress in defining how we're going to take advantage of that. The promised land there is not only you're going to save money, but you're going to significantly increase the satisfaction of consumers because you will be speaking to consumers in the language they want, in the tone they want, at the moment they want with the level of specificity that they want. Sort of significant promise there, but I think it's going to be a longer haul then it would appear from the news articles. The second category which you mentioned as well is the use of AI in improving our network operations. And here, it goes everything from the way we write code and we analyze code and we validate code today, all the way to the way we use power in things like our brand networks and mobile brand networks. And here, frankly, the opportunities are amazing because what we are seeing is that very quickly, you can see improvements on an operating basis, and you're doing that, let's say, within the walls of your own enterprise. So your ability to control it, your ability to control the deployment, your ability to really understand the payback is significantly better than any time that you're going outside the enterprise. So I'm quite excited about the second category. The third category is AI to improve our employee experience and our employee productivity. And so that goes everything from AI into your traditional office applications. I'm sure that you're familiar with Microsoft advances, for example, in bringing AI onto that. But we feel that there's an opportunity here not only to make our employees actually more productive, but to do so while also improving the experience they feel as being employees, right? And to me, the fundamental here is to what extent can you decrease or even eliminate the repetitive, mundane work that we all have to do in order to get to the higher level, more value-add work that we all wish we would do. I think that balance, AI is going to play a significant role in resetting that balance in the expectations of us as a corporation and our employees as individuals. And then the final category for Liberty Global, which is always quite important is, as you know, we're also a quite active portfolio investor. Just to put some numbers, we have over USD 900 million of value today in our Tech Ventures portfolio. And as you can imagine, investing on AI is the hot area and it's also quite important for us. So we're trying to take an aggressive, innovative, yet guarded approach as to how do we think about AI. Probably, I would say, the biggest challenge that we have, which is probably quite common in the industry, is to prioritize where we're going to use AI, okay? Because AI today is still fussy enough that it looks like it could help, absolutely, right? Is it more effective for me to summarize this conference call, then to go and do task A, B or C? I think that a great deal of our challenge is to figure out where we're going to apply AI first, and then to go and select the right partners against that and then actually go and implement it.

That's fascinating. So looking forward to hear more. I'm sure that's a topic that will stay with us for a number of years. Moving to the fair share debate, which is a very important debate in Europe. The European Commission started a consultation and looking at this and whether Big Tech should pay something to telcos for the usage of the network. But it looks like the industry is struggling to have a consensus on this. So we hear some operators are afraid, some operators are less keen. And also at the countries level, it feels that also -- enjoys some different conversations on the topic. I'm going to get your views on the topic. What do you think is the right path for the industry to improve the return on the traffic which is generated by Big Tech. Is it the fair share, the proposition of the European Commission, the European Commission is looking at? Is it something else? Really keen to hear your views on the topic.

Well, I'll start by saying I am not the expert in the area. And my role is mostly to inform the experts in our company that are dealing with this. But as we have discussed throughout the whole conversation today, developing and deploying the significant improvements on networks and network performance and network services to end consumers is a very capital-intensive process. And I think the nature of fair share, which is can we find avenues in which the burden of building these networks is better shared across the industry? Obviously, we're a big supporter of that. Exactly how you do it, in which country, in mobile versus -- versus fixed versus network-as-a-service, even our network service is its own separate subject, is yet another way to think about how you monetize these significant investments and how do you bring other industries to participate not only on the cost sharing, but also the value creation that comes with that. So at the headline, we do believe that it's a very embedded dialogue. We do believe that the concept of better sharing the investment required to go and build these networks and better sharing the benefits that the business benefits that come afterwards is a valid concern and something that we want to see move forward, we want to see get executed. How exactly it happens, I'm sorry, I wish I could tell you I got a perfect view of this. I don't, but we certainly believe that there's room for something like fair share and network-as-a-service to happen and happen at scale in the industry.

Very good. Well, we're coming to the end of our session today. It was a very informative session. Again, Enrique, thank you so much.

Thank you. Thank you so much.

For your time. Thank you for -- to the audience for listening, and have a lovely day, everyone. Thank you.

Thank you so much. We'll see you hopefully not in a year.