Hello. Good afternoon. Welcome, everyone, to Cellnex Telecom's first half 2026 results presentation. I'm Maria Carrapato, and it's a pleasure to be with you today again. Before we begin, as usual, I'd like to remind you that this presentation contains forward-looking statements. Please refer to the disclaimer included in the appendix in the slide deck.
Marco will open the presentation with the main highlights. Raimon will take you through the financial performance, then Marco and Simone Battiferri, our COO, will close with some considerations on other industrial and strategic topics that are often raised by the market. With that, let me hand you over to Marco.
Thank you. Thank you, Maria. Good afternoon, everyone. Before going into the number, let me frame the first half in one sentence. Cellnex is delivering exactly on the model we described to the market: predictable organic growth, expanding margins, accelerating free cash flow, and tangible shareholder remuneration. Let me highlight the four key messages of the first half. First is that operational momentum remains strong.
Organic points of presence growth reached +4.9% year-on-year, confirming sustained demand from customers across the entire portfolio. This translated into solid financial performance. Revenues grew by 5%, Adjusted EBITDA by 6.4%, EBITDA after leases by 7.7%, recurrent level free cash flow by 11%, and recurrent level free cash flow per share by 18.1%. From a margin perspective, both EBITDA and EBITDA after leases margins increased by nearly 200 basis points, reflecting continuing operational efficiencies and land management actions.
The second is that free cash flow has entered a new phase. From EUR 19 million in first half of 2025 to approximately EUR 301 million in the first half of the year. This allows us to reiterate our guidance. Third, we continue to deepen our customer relationship. Sunrise in Switzerland, Vodafone España, and Telefónica in Spain are all examples of Cellnex being selected as a trusted infrastructure partner for network expansion, contract renewal, and network resilience.
Fourth, shareholder remuneration, which is delivered as committed. The EUR 500 million dividend has been paid, the EUR 300 million share buyback has been completed, and 11.3 million shares are expected to be canceled during the second part of the year, improving the per-share metrics. I'm pleased to share that today our board of directors has approved an additional share buyback program of EUR 200 million to be completed until the end of 2026.
Overall, the message is clear. Our industrial position is translating into predictable growth, stronger cash, and direct shareholder value creation as promised. With the announcement we are making today of an additional EUR 200 million in share buybacks, the total shareholder remuneration in 2026 will reach EUR 1 billion, combining the EUR 500 million of dividends already paid, the EUR 300 million of SBB executed in the first half of the year, and the additional EUR 200 million of share buyback just announced.
Between 2025 and 2026, we will have returned a total of EUR 2 billion to shareholders between dividends, share buybacks, representing 11% of our current market capitalization. We have no doubt that the share buyback we're announcing is highly accretive capital allocation decision, driving significant growth in per-share returns and long-term value creation for our shareholders. With that, let me hand over to Raimon, who will take you through the financial and operational performance in more detail. Raimon, floor is yours.
Thank you, Marco. Good afternoon, everyone. Let me start with the slide seven, which summarizes the financial performance of the first half on a pro forma organic basis. The key message is simple. Cellnex continues to convert predictable top-line growth into higher cash generation per share. Revenue growth was +5%, Adjusted EBITDA grew by 6.4%, and EBITDA after leases grew by 7.7%.
At the cash flow level, recurrent lever free cash flow increased by 11%, while the metric per share grew by 18.1%, reflecting both operational execution and disciplined capital allocation. On the next slide, we show the bridge from reported revenues to organic revenue growth. Starting from the first half 2025 revenue base, perimeter adjustments bring us to a comparable pro forma base.
This results in EUR 2 billion of organic revenues in the first half 2026, representing 5% growth year-over-year, supported by price escalators and continued demand from customers. Moving forward, you can see points of presence continue to show healthy commercial momentum. In absolute terms, second quarter showed strong performance with more than 2,000 new net PoPs and positive contributions across our main regions.
In the first half, gross PoP growth reached 5.7%, while net PoP growth was 4.9%. Importantly, this growth comes despite consolidation trends in some markets. Consolidation does not eliminate investment. Healthier operators continues to deploy more capacity, coverage, and network quality.
This is one of the most important messages from the first half. The need for densification remains strong, and Cellnex continue to capture that demand through both colocation and build-to-suit programs. The operational momentum translates directly into tower revenues, as you can see on the current slide. On a pro forma basis, excluding Ireland, tower revenues grew organically by 5.2%.
Tower revenues remain the core growth engine of the group, driven by contracted price escalators, colocation, and build-to-suit activity. This is the essence of our business model: growing coverage, improving densification, and leveraging our existing asset base. Moving to slide 11, our other business lines also continue to provide growth upside.
Fiber connectivity and housing services grew organically by 7.8%, adjusted for the French data center disposal and supported by the continued rollout of the Next Loop project in France. DAS, small cells, and RAN as a Service grew by 4.5% organically, supported by high demand in high-traffic locations, venues, and complex indoor environments.
Finally, broadcasting remained stable, growing by 0.5% organically, continuing to provide a steady and predictable revenue stream. We are providing you practical examples of how our industrial strategy translates into real commercial activity.
First, in Switzerland, Sunrise and Cellnex have extended their long-term strategic partnership through an expanded build-to-suit program covering 300 additional sites. This supports the next phase of Switzerland mobile network evolution and reinforces Cellnex roles as a scalable infrastructure provider.
Second, in Spain, our legacy Vodafone framework agreement has been renewed for 10 years, covering approximately 2,000 existing POPs. Importantly, the renewal has been signed on unchanged technical and financial terms, and Cellnex will also host a limited number of additional POPs on existing infrastructure.
Third, Telefónica and Cellnex have extended the backup battery partnership to a total of 3,800 sites, reinforcing network resilience and energy security after the recent blackouts. This reinforces our role as a trusted infrastructure partner and show how resilience, energy security, and network availability are becoming increasingly relevant customer priorities.
Beyond our traditional tower business, we are also expanding our presence in DAS and neutral host solutions. For example, Cellnex is deploying a multi-operator DAS at the new Valencia football stadium in Spain and expanding neutral host mobile connectivity along the Brighton mainline in the U.K. These examples show how Cellnex is actively shaping infrastructure solutions for our customers' needs and their ongoing network investments.
Turning to slide 13, operational efficiency continues to be a key lever of value creation for Cellnex. On a pro forma basis, cost per tower decreased by 3.3% year-on-year, maintaining a high level of operational quality across the portfolio. As a result, we continue to expand margins, with EBITDA margin reaching 84.6% and EBITDAaL margin increasing to 61.8%, the highest level achieved in recent years.
This reflects the operating leverage embedded in our business model, the benefits from our efficiency initiatives, and the continued progress of our land management program. In short, we are not only growing revenues and cash flow, we are doing so with greater industrial efficiency, supporting sustainable margin expansion and value creation for shareholders. Next slide show the cash flow bridge from the first half 2026.
Starting from EBITDA after leases, we reached recurrent lever free cash flow of EUR 908 million, and free cash flow after expansion and build-to-suit CapEx reaches approximately EUR 301 million.
The three key drivers behind the result are the solid operating performance, an efficient capital and tax structure supported by optimized cost of debt, and lower build-to-suit CapEx intensity as the build-to-suit cycle normalizes. On slide 15, shows the free cash flow inflection point where it's clearly visible.
Pro forma organic recurrent lever free cash flow increased by 11%, and recurrent lever free cash flow per share increased by 18.1%. The share buyback program is enhancing per-share value creation, while the business itself continue to generate stronger underlying cash flows. At the same time, free cash flow increased from approximately EUR 19 million to EUR 301 million in the first half 2026, an increase of approximately EUR 282 million year-on-year.
Free cash flow generation is no longer a future promise. It is happening now, and it's accelerating. Our liquidity and funding position remain very strong. At the end of the first half 2026, liquidity stood at approximately EUR 5.3 billion, including circa EUR 2 billion of cash and EUR 3.3 billion of undrawn committed credit lines.
As such, our 2026 and to 2028 maturities are largely funded, giving us flexibility to navigate market windows. With that, let me hand back to Marco to discuss some broader industry dynamics and why they reinforce our confidence in the long-term investment case. Marco, over to you.
Thank you, Raimon. I would like now to step back from the financial result for a moment and discuss the broader industry backdrop. This matters because our equity story is also about why demand for our infrastructure will remain strong for many years. The next slides address some of the topic investor raised most often. Traffic growth, direct-to-device satellite, Europe's competitiveness and digital sovereignty, and MNO consolidation in France.
For us, the conclusion across all four of the topics is consistent. Europe needs more infrastructure investment, and Cellnex is one of the best-positioned platforms to capture this investment cycle.
The topic is very technical, so my colleague, Simone Battiferri, our Chief Operating Officer, is also joining us today, and he will walk you through trends in mobile data growth and explain some fundamental concepts on direct-to-device satellite connectivity. Simone, make it simple, please, and drive us through the mystery of the technical stuff.
Thank you, Marco, and good afternoon, everyone. Well, looking at slide 18, the key message is that we see a clear positive inflection point in mobile data growth. In fact, it is accelerating again. Global mobile networks data traffic increased by 22% between first quarter 2025 and first quarter 2026, confirming that demand for mobile capacity continued to expand at a sustained pace.
Importantly and curiously, this acceleration is not yet AI-driven. Latest industry traffic numbers make the point clear. AI applications are still a very small portion of total traffic, let's say low single digit today.
There has not yet been a visible AI-led inflection in mobile uplink trend. The main growth drivers remain video streaming, the expansion of FWA, and the increasing penetration of 5G devices. This matters because Europe is still in the middle of the 5G adoption cycle.
5G-enabled devices represent only half of the mobile connection in Western Europe today, and adoption is expected to move close to 95% by the end of this decade. Well, the implication is clear. 5G subscribers already consume around 3 times more than a 4G user on average. 5G traffic per user is expected to further double in the next few years. The maths are simple.
As the customer base continues to migrate to 5G in the coming years, overall traffic will increase significantly. Underlying consumer behavior is therefore still pointing to annual data growth above 20%, and more network capacity will be required just only to avoid congestion and preserve service quality. At the same time, the nature of traffic is changing too. The network is not only being asked to carry more data, but to carry more demanding data.
AI-enabled applications, real-time collaboration, industrial automation, connected devices, future wearables will require networks that are more responsive, more reliable, and more available indoor. They will also increase the relevance of uplink capacity, making networks demand progressively more symmetrical than in the past. Mobile networks are entering a new phase.
The question is shifting from, do I have coverage? To, do I have guaranteed quality? Quality means higher capacity, stronger indoor performance, more predictable service levels, better reliability, higher uplink capacity, and lower latency. The physics of the networks at the end are simple. Carrying more data requires more capacity.
More capacity generally means using more spectrum and adding higher frequency. Higher frequencies deliver more bandwidth, but at the same time, they propagate over shorter distances, requiring a denser grid of cells. Furthermore, they penetrate buildings less effectively, which makes dedicated indoor coverage increasingly important.
The only way to solve that equation is to bring the network physically closer to the user. That means densification, particularly in urban areas, that translates in more indoor systems and a larger and more capable infrastructure footprint. If we go to slide 19, let me address another topic we are increasingly asked about by our investors.
That is the direct-to-device satellite connectivity. Let me start by saying clearly that satellite is a powerful and valuable technology. D2D can play an important role in the broader connectivity ecosystem, especially in remote or underserviced area where terrestrial coverage is technically difficult or impossible, or even economically inefficient to deploy.
Satellite and terrestrial networks are designed to solve very different problems. Terrestrial networks are optimized for capacity, latency, reliability, and indoor performance. Satellite networks are optimized for extending coverage over very large areas.
The reason is not only technological, it is physical. A terrestrial network can reuse spectrum every few hundred meters, serving a very large number of users simultaneously with high capacity. A satellite beam, on the opposite, covers a much larger area, shares capacity across many more users, and has far less ability to reuse spectrum effectively.
This is why the average downlink capacity of a satellite D2D connection is today less than 0.1% of a terrestrial mobile network in urban areas and less than 5% in remote location. Said in practical terms, even when constellations are fully deployed, the service will remain much closer to a 2G-like experience, valuable for essential connectivity, messaging and potentially basic voice, but not comparable at all to mobile broadband. Distance is another fundamental constraint.
A low Earth orbit satellite is typical hundreds of kilometers away from an handheld device and sends a vertical signal that, in urban and suburban environments, must pass through multiple physical obstacles before reaching the user. By contrast, a mobile antenna on a tower is usually a few hundred meters to a few kilometers away and projects a horizontal signal designed to serve user with much higher capacity and better penetration into buildings.
This is particularly important because most mobile traffic is generated indoors. We are talking about two-third, where satellite signals face a structural limitation. The differences define the role that direct to device can play, an excellent complement for coverage, but not a total substitute for terrestrial capacity. If you go to summarize the key takeaways in slide 20, first, satellite D2D is fundamentally a coverage solution.
Its strongest use case is sustaining basic connectivity to remote or hard-to-reach locations. A clear evidence of these days is the emergency connectivity provided by satellite during the wildfires raging across Spain and France. Second, the capacity gap versus terrestrial network is structural and rooted by physics itself. Satellites are much farther away.
Beams cover much larger area. Spectrum is reused far less efficiently, and signals face more difficult propagation and penetration condition. Even if there is a major future technology breakthrough in direct to device, the satellite capacity will remain multiple times lower than mobile terrestrial networks.
Third, indoor performance remains a fundamental limitation, particularly because most mobile traffic originates indoor. Fourth, handset battery life and antenna constraint create additional challenges for uplink capacity and user experience. Let me say that the conclusion is clear.
Satellite direct to device does not replace the need for micro towers, network densification or dedicated indoor solution. As mobile traffic growth and quality requirements increase, especially in dense areas, investment in terrestrial mobile infrastructure will remain essential. Well, back to you, Marco.
Thank you, Simone. Thank you for this jump into the technology. I think you made it simple and clear. Really thank you for your effort. Europe faces a clear challenge in terms of digital competitiveness. Average mobile download speed in Europe is around 86 megabit per second, materially below North America, China, Korea, Japan.
This gap is increasingly recognized as a strategic issue for competitiveness, resilience, and security of our continent. GSMA estimates total mobile investment needs in Europe at around EUR 475 billion over the next decade.
EUR 270 billion just to maintain the ordinary technology cycle, which explicitly includes densification to improve quality, coverage, and performance, plus EUR 200 billion more if Europe wants to regain connectivity leadership.
I would like to call your attention to yesterday's announcement by AGCOM of the spectral renewal public consultation process in Italy and the network performance metric expected from the MNO. Please consider that similar regulatory processes are advancing in Portugal and France.
The proposed obligations include investment in transport routes, 5G Standalone coverage, network resilience, and AI-ready networks, all areas that are infrastructure-intensive. We haven't even begun to talk about 6G, which is expected to kick off in 2031 or 2032. This is the core message for investors.
Europe cannot close the gap with spectrum alone or software alone. It needs physical infrastructure, more densification, better resilience, higher quality network. Cellnex is uniquely positioned to provide efficient, shared, and sustainable platform to deploy those networks. Let me finish with France, which is one of the topics that animate quite the discussion around Cellnex.
All of you remain constructive. We believe consolidation has the potential to create stronger operators with greater capacity to invest in network quality and infrastructure. As shown on the slide, our direct exposure remains manageable while the process itself is expected to be lengthy, involving regulatory review, approvals, and a very long transition period. Timing is important.
Operators are unlikely to make long-term decision based solely on the network requirements that we see today. By the time the consolidation process is fully completed, traffic volumes, 5G penetration, digital user patterns will be significantly different from where they are today, as we already discussed with Simone.
In that environment, operators will need better networks, not smaller networks. Over the past decade, France has been adding around 5,000 tops per year, reflecting the industry's ongoing need to invest in network capacity and quality.
Despite that level of investment, France still ranks only 21st in Europe and 34th globally in mobile network experience, highlighting the significant room for improvement that still exists. That is why we believe it is important to look beyond the analysis of the respective networks today and focus on the network that operators will need five years from now.
Stronger operators with healthier balance sheet will be better positioned to continue investing, densifying, and improving network performance. Our objective is therefore to remain constructive and proactive, working alongside our customers to support that next phase of investment.
If we step back from all these industry trends, the message is remarkably clear. Mobile traffic continues to grow. Europe needs to improve its digital competitiveness. Satellite enhances connectivity but does not replace terrestrial capacity. Market consolidation can create stronger operators with greater ability to invest.
When we look across traffic growth, digitalization, AI adoption, network quality requirements, and market consolidation, we arrive at the same conclusion. Europe will need more network capacity, more densification, and continued infrastructure investment.
In short, the future require better networks, and better networks require more infrastructure. Maria, it has been a little bit longer than usual. Please, I will like to excuse with our investors and analysts and back to you.
Thank you, Marco, Raimon, and Simone for very clear explanations. We're now open to take calls. The first question comes from Ondrej Cabejsek from UBS.
Hi, can you hear me now?
Yes, we can.
Good. Thank you. Thank you for the presentation and all the very helpful additional color. I had two questions, please. Both topics that where I touched upon. The first one was on the French deal, just from a regulatory perspective. We now know that this will be analyzed locally, and I was wondering whether from your perspective there might be any differences to the kind of remedies that you would expect around, say, investment obligations.
Obviously the reason I am asking is that while we have some commentary from the French authorities around the need to improve networks, it is perhaps not as explicit as a policy as it is from the European Commission with the Draghi report. Any implications of the local jurisdiction versus EC in France would be very helpful.
The second question, Marco, you touched upon this, and Simone I guess as well, but the AGCOM proposal yesterday where obviously the network obligations are, I guess, very conducive for someone like yourselves. I guess there are two dimensions I would appreciate if you could give us color on.
One are the coverage obligations and looking at the targets that they set out or AGCOM sets out, I was wondering what kind of increase in the number of sites and the physical infrastructure that you were talking about do you reckon is needed to comply with these numbers?
I believe Inwit suggests something like 20%-25% to reach some kind of good coverage in Italy or up to 12,000 sites. Is this roughly your estimate as well that Italy should see an increase in the number of towers similar to that?
Secondly, and perhaps more technically, the regulator says there will be a minimum download speed requirement of 150 megabits per second. You make the point that data traffic continues to grow and Ookla reported median speeds in Italy at 1Q exactly half of this proposed floor.
On that one, I was wondering how these higher speed requirements could translate into more business for you. Again, any estimate of PoP increases to get there would be very helpful. Thank you.
Sure. On the regulatory France versus Europe, I think that it has been a good decision because I expect that the French regulator will be closer to the market, closer to the interest of the French. At the end, when you make a new market design, you're asking to a country to renounce to something, which is the competition on the market.
You're making something that is very important and so the citizens of the nation, the community needs to receive something in exchange. It's good that the regulator is French, but I'm 100% sure that there will be a super strong bonding and connection between the French regulator and the EU authority. Ultimately, it will be more or less the same with a, let me say, an easier accessibility of all the parties that are involved vis-a-vis the regulating authority.
Remedies are expected to be the usual ones, the ones that we expect. It's not what, it's the size probably that will be decided by the authority. This brings me to your second question. Your second question is, 12,000 new sites is mid-high in my expectation. I would have said 10 to 15. 12, you are in the mid-high part of Italy, of the Italian need.
How much coverage, how much densification? I think that the coverage has a permitting process way easier, but it has to be thought, I would say with a business model that has to be more convenient for the operator. I would strongly suggest to have more densification, more sharing, more collocation. This is what is needed. Now, your technical part on the speed.
You know that when you move from the current 5G to the 5G Standalone, you're not touching just the transmission. Because the transmission, at the end, is limited by the core network. Good part of the problems in speed latency and responsiveness of the network depends also on the fact that the core networks of the operators are not designed for the 5G Standalone.
This is something that most of the MNO make very clear every time. My old friend Pietro Labriola makes clear every time. He has to invest a lot on the core network. Core networks are expensive.
Yes, core networks are relatively expensive, not dramatically expensive, a few hundred million EUR. The problem is that it's a few hundred million EUR each. It's very difficult to share a core network. If you can be efficient in towers and in transmission, in transport is less obvious. If you put in your model something on core network, you don't make a mistake. Hope I answered, Ondrej.
Very interesting. Thank you.
Okay, the next question comes from Andrew Lee at Goldman Sachs.
Good evening, everyone. I had two questions. Firstly, I just want to say thanks for the satellite technology articulation around the debate. Obviously, a lot of misunderstanding or lack of understanding there, and I think articulating it is really helpful, especially given that I think operators have really struggled to do that, which has not helped the conversation.
Moving on, I had two questions. Firstly, just wanted to ask, have you seen any signs of post-consolidation densification acceleration by operators in Spain and the U.K.? Obviously, it's a key area of confidence building in the consolidation debate.
Do you have any visibility on when this will begin? Given it doesn't look like it's started yet. Then second question, there was a press article a week ago suggesting that Cellnex had been examining strategic options, including buyouts, large-scale mergers.
I just wanted to ask, is there any truth to this? I'm not expecting you to comment on specific examples, but do you think there is a material strategic option available? In the context of this, just wonder if you could comment on why you chose to buy back shares rather than pay down debt today. Thank you.
Okay, definitely your two questions are three.
You got me.
Well, Andrew, let's start from consolidation and Spain and U.K. In Spain, we start to see something happening. MasOrange is well advanced in their integration phase. You saw that last year we had a big bulk of their consolidation. In the second part of this year, we will have a second step that we agreed with them, but as before.
Now, they are working on two areas. They are working very seriously on transportation corridors. Transportation corridors, MasOrange is making this as a strategic investment area. They are starting using small cell way more than what was done in the past, especially in problematic dense urban areas. This is good because we are working very strictly with them.
It's not particularly known, but we are the largest operator in Europe in DAS and small cells, so we have a big know-how, and in particular, our Spanish chapter is possibly the most advanced that we have in the group. U.K., unfortunately it's taking longer than what we expected.
We don't change our view. The fact that something will happen, I'm totally sure that something will happen, that there will be a need of some thousands of sites that have to be built, both urban and non-urban.
We're talking about several thousand sites that we, in our network simulations, that we see. For the time being, different from other markets, the carrier neutral model for, again, transport lines. The Brighton line is something that possibly some of you use every day, where the coverage is provided by us.
This is something that, again, we see, but unfortunately not in the order of magnitude that we expect. Is it something I do expect for 2026? Honestly, doubtful. Is it something that I expect for 2027? Yes, definitely. On strategic option, I read what my lawyer wrote me. He's here in the room looking at me.
As a leading European player, it's natural that the company may attract investor interest at current valuation levels. However, we're not familiar with the conversation referred to the article.
Apart of being very well written, you get the sense, yes, it's convenient this price for Cellnex is the reason why we're making the share buyback. Your third question is why share buyback and not debt repayment. Well, our cash generation is doing well. We're convinced that we can do well for the entire year.
This means that we, as I told, entered in a different phase in which cash is going to be very evident. We're not changing our overall targets for capital structure. Today the share price does not reflect our vision on the intrinsic value of the company. Headwinds are, in my view, overestimated in our price, and tailwinds that Simone was saying are not included.
The board is convinced that allocating EUR 200 million to share buyback in this moment generates value to the shareholders, not only in the short term, but most importantly in the long term. This is why the moment is correct for doing it now.
Thank you. Very clear.
The next question comes from Akhil Dattani at JPMorgan .
Hi, good afternoon. Marco, maybe I can start with the shareholder return comments you just made and maybe just ask for a bit more color. I understand, as you mentioned, it's a reaction to seeing value, but I guess I'd love to understand the general framework you're using in terms of thinking about what you want to do.
Because over the last couple of years, we've had a few add-on buyback decisions that you've taken as you felt it was appropriate. If we try and step back and think about the journey going forward, can you sort of help us frame how you're likely to approach your decision around buybacks? Is it going to be opportunistic based on share price? Is it going to be based on more framework-driven decision-making?
If you could just elaborate generally how you're thinking about the philosophy of what you're likely to do, that'd be super helpful. Then the second one was the topic you've mentioned around tailwinds. You talked a lot about various opportunities. One opportunity you didn't mention, which your U.S. peer has talked about a lot, is edge computing.
I'd love to understand what your general thoughts are. I appreciate it's a long-term topic, but the U.S. tower cos are already starting to make investments in this space. Do you see it a little bit like small cells where the U.S. pushed hard and ultimately didn't amount to a huge opportunity, and I guess you didn't pursue that at the time? Or is it different this time? Is this something that you similarly also see as being an interesting opportunity for towers midterm?
Cool. Well, on the philosophy, you remember we made very clear that we have a dividend policy that we're not changing. Our dividend policy is EUR 500 million growth, 7.5%. You remember we are delivering. We paid the EUR 500 million, you can bet that next year we will do EUR 500 million plus 7.5%.
Second, we said the minimum we're going to do is EUR 800 million, the delta between dividends and EUR 800 million is depending on the value creation. The value creation is, in this moment, clearly coming from share buyback.
I wouldn't describe this as opportunistic. I would describe it as logic or, if you want, fundamental analysis. There is a big difference between yesterday, today, and tomorrow. Yesterday, we made share buyback because we had some extraordinary disposal, and which made available some extraordinary money that we used.
Today, we are making a decision that is based on our capacity to generate cash flow, which going forward will remain generous and abundant. We said that we will allocate this extra capacity in the way that will generate more value to our shareholders. As of today, the board made the decision that share buyback was the way.
I would say that this is the philosophy. Let's consider what is the value creation, where the value creation comes from. Let's see what are the resources available, structurally available, and let's do it.
On your second question, it's very interesting because Simone and I, we have been discussing this topic not less than five times in the last two weeks. If I look what has been done today by, look, for example, NVIDIA and Nokia, it's not really an edge computing on a tower.
It's an AI-integrated equipment with a sort of self-configured or AI-driven configuration of the equipment. It's not really edge computing on the tower. I'm convinced that the more we enter into distributed AI, the more a system that brings everything to the center is tremendously inefficient in terms of traffic load.
Every time you have to transport a lot of data that most of the time are useless. Imagine that you need some AI for self-driving vehicles. Do you really need to have data going to U.S. and coming back, or it's better to have some maps on a tower that is 300 meters from you? I think that this is something that can happen, but it's really, at the moment, a bit unclear how it will happen. Simone is working very actively.
Simone has been in the IT and in the semiconductors for a good part of his career. We are activating our contact. Just stage 1, I think, is better understanding. I see an opportunity, but as you said, it's a midterm opportunity more than a short-term one. I hope I answered.
Yes, that's great. Thank you.
Thank you, Akhil. See you soon.
Okay, the next question comes from Rohit Modi at Citi.
Hi. Thank you for taking my questions. I have two, please, as well. One is the follow-up on Andrew's question around the article. I understand you can only talk to some extent on it. In general, your discussions with private players, what are the key constraints that you see in terms of valuation of towers?
I mean, apart from rates, is what is going on in Italy and partly in Spain as kind of rippling effect in the way private players see now towerco valuations with all the renewal risk. Any color around that.
Secondly, you mentioned about the cash flow generation was pretty strong in the 1H. Looking at the 1H number and the phasing you had in last 2 years on your free cash flow and recurring levered free cash flow, you are heading towards the upper end of your guidance. Is that the kind of base we should look for the second half and full year? Should that be the base for the next year in terms of when you look at the guidance range? Thank you.
Good. I'm looking my lawyer for a note. I can't add very much. What I can tell you, we trade at 14 times. We trade at 14 times. It's 14 times we trade at more than 10% recurring levered free cash flow per share yield. I think that there are numbers that speak for themselves. We are large. We are diversified. What I can tell you, honestly, I have very little to add. Cash, Raimon.
Hi, Rohit. Look, during the year 2026, as you have seen, we have massively increased the free cash flow. There is a changing point to a situation where we're going to be this year between EUR 600 million and EUR 700 million that we gave as a guidance. That will grow next year to a level that is from EUR 975 million -EUR 1,075 million.
We are reiterating our guidance. We are not expecting any change on that. This year, we are halfway on the free cash flow. Second part of the year, we will see as the first half, some build to suit still coming.
We will still have some of the growth coming from the colocation. All of it will help us achieve the guidance that we have given to the market. We are not giving any short-term guidance or anything similar because we are just expecting to be, as promised, between EUR 600 million and EUR 700 million.
Got it. Thank you.
Okay. Now moving to the next question. We have Roshan Ranjit at Deutsche Bank.
Great. Evening, everyone. Thank you for the questions. I've got two operational ones actually. Firstly, turning to Spain and the renewal of the framework agreement with Vodafone Spain. I think the first part very clear, renewal of existing PoPs on the same conditions. You've also added additional new PoPs.
I just wanted to get a sense of the kind of, I guess, level of discipline in that market, because clearly, one operator's been very strong about wanting to move. Is there scope for those additional PoPs to go higher that you could offer?
I know, Marco, you've been very clear on the kind of degree of overlap in that market. Is that still a big barrier for any kind of operator to switch, please? The second question, again, on the operational side in France, we've seen a pickup in the BTS deployment.
It seems quite evenly split between the Bouygues and the SFR build suits. How should we think about that going forward in the context of the kind of ongoing regulatory review? I know previously there has been talk about synergies from potentially combining build suits, or is it kind of business as usual in terms of the deployment until we get a bit of news flow through the year or perhaps next year? Thank you.
Thank you, Roshan. Spain. Our goal number one was we had 2,000 PoPs, which were secondary PoPs that were expiring, we wanted to renew. Our Spanish team has been able to renew at the same terms and condition we had before. Tick the box that we made another renewal without suffering, which is one of the many headwinds that time to time we have to face. We continue to renew ordinary course of business.
The second part of your question, we have been asked, making an analysis, if some of our towers could be eligible for hosting antennas from Vodafone. We made a technical analysis. Some towers, the answer was yes, some towers, the answer was no. We applied the usual price list. We did not make any special favor. We have been asked to host some 100 antenna, which is good.
To your point, is it something, it's new deployment or Sorry, I'm not the CTO of Vodafone. Is it densification? Is it coverage? Is it To say the truth, I don't know. Possibly it's densification. Second, France. Did something change in our build to suit program due to the SFR split? Of course, yes. It's obvious. What we are doing is there are areas that are not under discussion.
All the zone creuse, let me say, all the non-dense urban zone is coverage, and coverage is coverage. If there is not enough network, we build the network and it's good. By the way, we continue to insist to the concept of co-location to suit. Every time we build a tower, we strongly insist to have more than one operator in order to make those network more efficient.
First of all, for them, in order to avoid that those networks becomes way too expensive. In the urban areas, of course, we are working more prudently. We have to avoid to generate new overlaps, even though the consolidation takes time.
There is some business as usual, but business as usual with good common sense. Let's avoid to create today the problem of tomorrow. I think that in this, everybody is well aware. The attitude is constructive. We have several years of experience with all of them, so we're working well with them.
That's great. Thank you.
Thank you.
Okay, now moving on. The next question comes from Ulrich Rath from Bernstein.
Yeah, thank you very much. I have two questions, please. The first one is on BTS. Marco, on the recent interview that sort of popped up on YouTube, you talked about legacy BTS terms that need to be adapted for future BTS. Can you comment on how these terms are changing, such as the one that you're announcing now for Switzerland?
If it's not the numbers, it's sort of which elements of the BTS are you touching? My second question is could you provide an update on the land management program in terms of how far you are and how it's going? Thank you.
Okay. I answer the first. I leave the land to Raimon. Possibly the Swiss case is not the best example of an innovative contract. The Swiss case is a bit more of the same. We had a program with our client. We expanded the program. Please keep in mind that building new sites in Switzerland is quite complex because of permitting.
It's one of the country with most severe legislation, both on permitting and in electromagnetic emission limits. It is a bit more of the same, even though please remember that the Swiss market is really a very solid one because of its structure. When I say that going forward, we should imagine something different is yesterday, the build-to-suit were a sort of a forward execution of an M&A.
You were taking the M&A. I buy a part of the portfolio, which is an existing portfolio. I buy a part of the portfolio, which is a forward delivery at the same conditions. Which means that the same conditions were the conditions of a world that does not exist any longer. Rates are different, conditions are different, et cetera.
What I assume, I assume first that towers should be built by design multi-tenant. Every time we go somewhere, we have to make the question day one, how can we make it multi-tenant? This can allow to have two semi-anchor fee.
You should imagine something between a full anchor fee and a full second tenant fee, which will be convenient for both at the end. It will be convenient because you build day one the tower optimized for multiple tenants.
With the structural exercise, the structural engineering that is okay, with the energy, which is good. You have not to go there twice, so you save a lot of money. Possibly, if I make something like this by design, I can buy the land by design. If I buy the land by design, possibly we can share part of the benefit because this is something that's.
All in all, I think that the European case suffered a bit of over-financial engineering, and today we are telecom engineers. Telecom engineers work a little bit different, work more on the cost and the value that we are transferring to our clients.
Proximity to the clients, making the network not too expensive, transfer, sharing the synergy, sharing the savings that we make. We are making incredible work on AI applications for infrastructure. Of course, we are specialized infrastructure, so we invest. That's it.
On the land, Ulrich. As you know, we launched in the year 2024 the concept of Celland. It was an entity to be able to accelerate the acquisition of land, but as well, the cash advances on mainly rooftops in order to achieve an improvement of the efficiencies. This year, our cost per tower has been improving month after month.
The efficiencies that we have achieved so far this year offset and are a bit higher even than the increase on CPI. That is our target, always trying to offset the increase on CPI. Although you will have seen in the numbers in the free cash flow that the efficiency CapEx and the land acquisition CapEx is a bit below last year. It remains more or less in line, and we're expecting like last year, a bit of acceleration in the second half of the year.
We have already acquired more than 700 sites this year, more than 1,000 sites where we have done cash advances, and we continue with the same rhythm. The returns that we're getting are very much in line with what we had last year. The only thing is that, as Marco has mentioned before, we are being a bit more careful on some places.
For example, in France today, we're looking at the consolidation potential effects to make sure that we buy the sites that make sense buying, and we avoid buying sites that can generate a problem for tomorrow. So far, the program continues working extremely well, and we are not expecting any change rather than accelerating on the second half.
That's great. Thank you very much.
Okay, now over to James Ratzer from New Street Research.
Yes. Thank you very much indeed. Good evening. Two questions, please. The first one, Marco, is we've kind of talked a lot through the presentation about kind of technological drivers helping to support your growth, whether it's kind of 6G or FWA, more transport connectivity, just kind of growing usage.
Obviously today you've announced the deal with Sunrise, and we had the AGCOM announcement yesterday and other drivers as well. When I take that all together, how do you then actually think about what your organic tenancy growth will do over the next, let's say, kind of three, five, eight years? You're currently growing your organic PoPs at around 5% year-on-year. Do you think that rate of growth is sustainable at that level for the foreseeable future?
I just love to get your thoughts on putting all these drivers together, what it means for overall PoP growth. The second question I had was just, would love it if you could just dig in a bit further on the answer you gave earlier around Spain, to make sure I understand this correctly.
It sounds like on the new tenancies you've signed with Vodafone, were they approaching you to ask you for a much bigger potential portfolio of additions, but your pricing was suitably high that you only managed to agree on a few hundred? I just would love you to expand on that answer you gave a bit earlier to understand the process by which those few hundred new sites were agreed on in the Vodafone España contract. Thank you.
On the technological drivers, you have to split it in two. One is increased colocation. Increased colocation, if you want to really to understand increased colocation, first of all, you have to split between towers and rooftops, and when you split between towers and rooftop, you have to split once again between urban and non-urban.
On a tower, the tenancy can be way more than two. Because you can go with two tenants, plus an FWA, plus some other dishes. You can put a lot of things. When you are in a dense urban areas and you have a rooftop, going above one, it really depends where you are. If you are close to the center of Paris, it's possible that the mayor does not give you the authorization.
All in all, our portfolio, if you take mature countries, which are the countries in which we've been able to grow since more time, Spain, Italy, et cetera. The overall blended goes in the direction of slightly above two. If you take Inwit, it is above two. If you take mature tower operator, they tend to go above two.
Which means that on tower, you are well above two, and on a rooftop, you are in 1.5 or 1.6, et cetera. What makes the difference is the price mix. Today we have a price mix in which there is a big difference between an anchor, a second, and a rent sharing. Going forward, the densification doesn't bring the same effect of new network creation. The economic impact is the economic impact of a second or a rent sharing.
You have the future need of further network creation. There will be further network creation. As I was saying one second ago, the business model possibly will be different, but there will be still to build more towers. I would say that possibly another exercise will be proactively dismantle some towers.
This is going to be an exercise that I see coming. If I can proactively make some network rationalization, this can drive efficiency that can be shared between the tower operator and the MNO, which once again, reduce the appetite for making something bold, because we feed them with the savings that we can make for them. Is it a 5% growth sustainable eight years from now? Only God knows.
I would say that a 5% PoP growth possibly is, at eight years from now, possibly is a bit generous, but it will very much depend on how much network creation we will have. I don't think it's going to be zero. This is the big mistake, is that people believes that we have already too many towers, and this is wrong. Spain.
Let me try to put a little bit of order. Did we make special prices, or eventually higher prices, for new location? Absolutely not. We have a price scheme that is the same that we apply for second tenant. By the way, we apply to Vodafone the same as we apply to others. We don't privilege and we don't penalize.
Of course, what make the difference is that if, in order to host a new antenna, I have to rebuild the tower because I have to make so much CapEx to strengthen the tower that is an absurd, it's a little bit difficult. The first exercise is if I can materially host you in the place where you're interested, because by the way, you're not interested in every tower I have in my portfolio.
We have been asked for a certain list of towers, and we answered which of those list of towers could be eligible, easy, not easy Are dramatic. Okay? Is it densification? As far as we understand, there is a bit of everything. This has been the process. We don't add the specific CapEx, which is not the tower reinforcement.
We don't pay for the antenna movement, in case there is a movement of an antenna. It's really business as usual. I think it's a good contract for our client. The price we made for them is a very good second tenant contract because unfortunately in Spain, prices for second tenancies are a little bit lower than what I would like to have.
That's clear. No, that's great. You wouldn't expect any further announcement with Vodafone España at all over the next year or two as they resolve their issue with Vantage? You think your agreement with them is now finished?
As far as we have interacted with them, this is what they told us. If they will approach us again for having more, happy to serve my clients.
Got it. That's clear. Thank you.
Thank you.
The next question comes from Arnaud Camus from Bestinver.
Good afternoon. Thank you for taking my questions. On network resilience, could you provide more details on the 15-20 years life protection solution offered to telecom operators, including the recurring revenue profile and unit economics? Is it fair to assume the opportunity is greater in the U.K. and France?
The second one, given the recent geopolitical context and the growing use of drones in modern warfare, how significant an opportunity could this become for Cellnex, particularly in markets close to Ukraine, such as Poland?
Should we expect deployments of anti-drones to be mainly driven by public sector contract, and how should we think about them within your reporting framework, if it's a new tenant, an additional collocation, or any other type of services? Thank you.
Okay. When you refer to network resilience services, I suppose we are referring to energy resilience services-
Yes, batteries
which is the batteries, to give the vulgar name. We buy the batteries at pan-European level. These allow us to have better prices and better conditions. Better conditions means that we have guaranteed the life of those batteries for a material number of years that we are extending. Now we are depending on the supplier between 12 and 15 years.
Is it good? It is super good because, of course, if we put the battery, you have to imagine that those batteries can stay indoor or outdoor. Of course, the life of a battery which stays indoor is longer than the life of a battery which stays outdoor because of the obvious conditions, even if they do not burn, in any case, there are weather, atmospheric elements that shorten the life.
This is important because when we make the agreement with the MNO, the agreement with the MNO has the duration of the underlying contract that we have for the tenancy. If I have a 15-year contract, I have to provide the batteries for 15 years.
It is important that we have guaranteed behind us that after seven years, I have not to remake another cycle of investment, because otherwise the business case basically does not work. Some economics, we make more or less with the total is a sort of EUR 5 million to EUR 6 million a year in this moment with our client.
If you look, it means that good for them, good for us. Of course, this includes also all the maintenance services that are on us. It is turnkey. The client has not to worry about nothing because we do everything. Your second, is this potential big?
Yes, I think it is. Of course, Spain started first. Why? They had the blackout. By the way, now Spain is, we created Iberia, and Portugal had the same problem as Spain when there was the blackout. People started to be sensitive. I think that this is becoming more a topic, a trend, a European trend, that resilience. I see two big trends. One is sovereignty and the other is resilience.
On sovereignty, I can do nothing, because honestly, the batteries, you can have all the fantasy you want, and then you end buying in the same place. On resilience, we can do a lot. Defense. Can the tower be used for defense? Well, defense are big budget. Most of the time when you talk about defense, the problem is not the budget, the problem is the solution. What is the solution?
We are not a company which make anti-drone system. We don't make anti-drone system. We have good IT, but our IT makes towers. It doesn't make anti-drone systems. Of course, a tower can host.
To your question, what is it? It's a colocation, and it's a colocation of a PoP. Which kind of a PoP is it, and who is the client? It's more or less the same animal of a client that is not an MNO. That's it. We don't do more than this, and our business model is not different from a usual PoP colocation.
Okay. We've got another few questions. I'm conscious of the time already, let's try and rush through now. It's Fabio Pavan now from Mediobanca. Or maybe not.
Yes. Hi.
Yeah. Yes, Fabio.
It's here. Can you hear me? It's tough.
Yes.
Well, thank you for the presentation. I think it is interesting to have it today. I was wondering if you can help me in reconciling what we just discussed about this need for densification to support all these data center plan, AI spending capacity, with the news we had today with Europe launching the EUR 30 billion plan for gigafactories, how we could think about digital network to be involved, because I think this is clearly needed, so wanted to have your view on this. Thank you.
Let me make two points. One is, one of our peer invested in a big data center, one of our U.S. peer. Of course, this has given them a very good growth, but is draining an enormous gigantic amount of CapEx. This is one piece of the answer.
The second piece of the answer is that requires a huge expertise and know-how, and I love my engineers, but they are telecom engineers and not data center engineers, and we don't have this kind of a know-how.
We decided that investing on scale on gigafactories, on data center, et cetera, is not for us. This is why we sold our data center in Spain. This is why our proxy of a data center in France have been sold, and this is why we're not going to put our fingers in big data center projects.
There are very good specialized companies, and we leave to them. Second is, somehow what was the question from Akhil. Where the data center ends. There is a portion of the data center which can stay at the foothold of a tower. Fabio, I'm listening this story since not less than five years.
I saw some of those, I don't know how to call them, remote data center, and at the end, never happened on scale today. The point that Akhil is making is sometime it's just too early, how sometime what is not working yesterday, it was just because it was too early. Is it tomorrow the day that this will happen?
We are super active in the technological space. You know me since ever, you know Simone, we are both tech geek, very curious. We monitor this and believe me that if there will be an opportunity, we'll be there.
Okay. Moving on. The next question comes from Fernando Abril-Martorell from Alantra, sorry.
Hi. Thank you for taking my questions. Two very quick ones. First, you targeted five to six times leverage back at the CMD. Where do you want leverage to be as things stand today? Where do you want leverage to be at the end of the decade? Low, upper end, somewhere else?
Linked to this, because obviously this is the other part of the equation, can we assume that you distributed EUR 1 billion last year, EUR 1 billion this year, should we assume EUR 1 billion as the shareholder return floor for the next years with obviously the committed dividends and on top, dividends or buybacks now depending on the share price. Should we assume EUR 1 billion as the new floor? Thank you.
Okay. leverage, we're going to be between five and six at the end of the decade. Sorry, it seems to be a little bit of a stupid answer, but it will really depend on the conditions on the market. If the market we see with structural tensions, let me underline structural, because what we see today is a crisis that is driven by energy because of a war that we all hope that will not last long.
Structurally, we don't see high interest rates long-term. We don't see the need to explore the bottom part of the range. If you ask me about end of the decade, I honestly don't know. We will be coherently in the range between five and six, depending on the structural conditions on the market. Very good question on the floor. It's a mixed answer.
Is it a new floor? No, it's not a new floor. It's the demonstration that when we said that everything that is made available will be managed properly is what we do. We said our floor is EUR 800, and if there is more, we would return to shareholder. There is more, we return to shareholder. What about next year? We have a floor, which is EUR 800. If there is more, we would return to shareholder.
Okay.
Thank you very much.
Last two questions. First coming from Abhilash Mohapatra at Exane BNP.
Hi, good evening. Thanks for taking my questions, and thank you obviously for all the detailed Q&A and thoughts. I just had a couple of questions on the Q2 results themselves. Firstly, just on slide nine where you show your net colocation growth, nice pick-up in Italy and Spain. If you look at the growth on an equivalent basis, it's about a third of the total numbers, or should we just assume that Spain growth is mainly Digi RAN sharing PoPs and Italy is IoT?
The second question, just on the cash flow this quarter. Quite a strong positive working capital contribution. Just be interested to hear your thoughts on how you expect that to evolve on a full year basis. Is it still a sort of neutral contribution, or do you expect this to remain positive? Thank you.
Italy is not IoT. Italy, there was some RAN sharing, and the rest was second tenant. The number of build-to-suit is relatively modest in Italy. In Spain is Digi. The majority is RAN sharing. We are moving towers for MasOrange. Net when I move from A to B still remains one. Sometimes moving one, you generate the need of another one.
We have some mild growth, but we have some growth. Spain is Digi plus MasOrange. Telefónica, not that much. Vodafone for the time being, no, but let's see going forward. Italy has been some RAN sharing.
Very interesting. Some RAN sharing of who? Some RAN sharing of Vodafone Fastweb. Vodafone Fastweb, who had a RAN sharing agreement with WIND, is still adding some RAN sharing PoP.
This tells you that Swisscom is used to a network quality whose KPI are not the same KPI of Vodafone Italy. They wanted to improve the network quality. That's super interesting. It's something that honestly a bit surprised us. I have next to me the king of the working capital. Raimon, please.
On the working capital, Abhilash, basically as you have seen, the second quarter we have been improving following a similar trend to what happened last year. We had a first quarter with a negative working capital, second quarter improving the working capital. There are a couple of things there.
First, there is some seasonality. We have some contracts that have different payment terms. Some of them get paid end of quarter, some get paid at the beginning of the next. That always plays. It's true that we have been improving working capital year after year.
We were having a working capital closer to the 8%-9% on sales. We are now closer to the 6%-7% on sales. We expect to keep on improving. We still have some room for improvement, both on the receivables, but also on the supplier side.
There is a continuous improvement plan, we expect that it will keep on improving. It's true that the more that we make it more efficient, the more difficult it is to keep on having a big improvement.
As you know, we always say that working capital tends more or less to zero. Trying to make sure that any growth that we have that has an impact on the working capital can be made more efficient and get back to zero.
That's clear. Thank you very much.
Okay, the last question from Fernando Cordero at Banco Santander
Hello, thanks for taking my only question. It is quite a follow-up on the former tenancy ratio question, with a different scope, more on the short and medium term. As we are approaching to the end of the build-to-suit programs, it's clearly impacting cash flow, also would like to understand the impact in organic growth.
In other words, at which extent the build-to-suit program fade away should be a positive driver for the pure colocation growth. Just to understand if there is any, let's say, link between the end of the build-to-suit programs, let's say, some increase on the current trends from pure colocation. Thank you.
If you look today, the 5%, you split the 5% between CPI colocation and build-to-suit, you have a sort of a 3% from CPI plus colocation, a little bit more, you have a sort of a 2% from build-to-suit. Big numbers. If you take this trend, you see that we are investing less in build-to-suit, so this contribution from build-to-suit is progressively reducing.
This is also the reason why we were growing at 6%, we're growing at 5%. The question is should we imagine a build-to-suit down to zero? The answer is no, we should not expect down to zero. Should we expect it down to the huge numbers we had in the past? No, it's not going to be like this. Possibly, of course, our effort will be to push our machine in order to make more colocation.
We have several countries in which we can do more, we have to make a little bit better if the contribution from build-to-suit will decrease. Some network creation, I think it will be more selective, it will be with a different contract, et cetera.
Some network creation will remain. You have not to factor that the three becomes five and the two becomes zero. The three will grow a little bit, and the two will decrease, but it's not going to go to zero. I hope I gave you help.
Absolutely. Many thanks, Marco.
Thank you. It was the last one. Before passing to Maria, please let me thank all of you for your time, for the participation, and have a super good and joyful summer vacation.
Thank you, Marco. As usual, if you'd like any follow-up with any questions, you know where we are in the IR team. We've also purposely left a slide being projected, which gives you a link to many documents that we've been posting on the website. Please take a look, because there's quite a lot of information which could help further understand the equity story. I reiterate Marco's words, have a lovely summer holiday.