Good afternoon, ladies and gentlemen. Thank you for joining Merlin Properties' 2024 results presentation. You can find all the materials that will be covered in today's call available on our website. I will please ask you to abide by the disclaimer contained in it. Our CEO, Ismael Clemente, and our two directors, Inés Arellano and Francisco Rivas, will walk you through the main highlights of 2024. We'll thereafter open the line for Q&A, where you will have to press star five. With no further delay, I pass the floor to Ismael.
Thank you, Teresa. Good afternoon and welcome to Merlin Financial Properties 2024 results presentation. As commented by Teresa, the call today will be joined by my colleagues, Inés Arellano and Francisco Rivas, in order to avoid a monologue exposition from my side. Our CFO, Miguel Ollero, cannot join today. He recently underwent surgery and is still at the hospital. So it will be Fran, Inés, and myself taking the call today and the questions afterwards. Regarding the operating performance of the company, the operating momentum during 2024 was exceptionally strong, with very good rental growth combined with record occupancy, which for the overall portfolio now reaches 96.7%. It's very, very successful. The offices continue to perform very, very well, with a positive like-for-like of 3.9% and a re-leasing spread of 2.3%, which is affected by some little weakness, particularly in Barcelona, because otherwise it will be even stronger.
Madrid is performing very well. We have reached full occupancy in the real sense of the term in logistics, with 99.4%, historical record too, with a positive organic growth of 2.8 like-for-like and good performance in terms of pre-let of existing development pipeline. Shopping centers delivered an incredible year with a very good like-for-like of 2.7, but more importantly, with strong sales and footfall reflected in a very affordable level of occupancy cost ratio of 11.2%, which again is our historical minimum. So our tenants are really doing very well. And what we hear from them regarding 2025 is positive too. I mean, they have started the year with a very positive tone. The Reyes and sales campaign in January have gone very well, and they are all operating under the assumption that the year is going to be very positive as compared to 2024.
In terms of financial performance, the company generated a very solid FFO. Clearly, there was also some boost by the capital increase carried out in mid-year because that increased our financial balances at banks. And that interest income, of course, helped the performance of our FFO, but a very significant part of it is purely operational. The valuations remained broadly stable. I mean, we continued depreciating a little bit logistics, shopping centers, and offices in order to be as close as possible to what we believe are reasonable passing yields in order to stabilize the valuation of those assets. But that devaluation has been offset by a solid gain in the appraisal of data centers. Our financial situation following the capital increase, of course, is super strong with a 28% LTV, everything at fixed rate, and no additional maturities till November 26.
We are repaying with existing cash, the bond maturing in May, which is yielding 1.75%. So, of course, this will become a recurrent, I guess, theme over the next years. We are facing a steady increase in our financing costs, and that will continue through 2025 and 2026. And I guess we'll reach stabilization towards the end of 2026, 2027. Both Moody's and S&P have upgraded our company. The interesting thing about that upgrading, that rating is not only financial standing, it's also a testimony of our ability to continue generating or to grow our cash flow profile in the coming years through our data center activity. And, well, as you all know, we carried out a successful capital increase mid-year in order to fund the Phase II of our data center development program.
Regarding value engineering, we sold EUR 73 million of non-core and left some EUR 53 million signed for this year. So our objective of EUR 100 million for the year was missed just because of the year-end cut-off date. But in general terms, we exceeded that objective. In our Landmark Plan, we are basically finished. I mean, we have delivered Plaza Ruiz Picasso 11 with incredible tenants. And for those of you who are familiar with Madrid, if you have visited the asset, you will see that it has become a landmark of quality in the Azca area. And not only in Azca, I believe in all central Madrid is a building which now is praised by everybody. Regarding the plans, Best II and Best III for logistics, we have finished our capacity to develop in Portugal by leasing close to 135,000 sq m to a Portuguese end client.
Those square meters will be delivered in several phases, mainly during 2025 and 2026, although there will be one shed that will be delivered end of 2027, beginning of 2028. With that, we have basically finished with our land development capacity in the Lisbon Logistics Park, taking into account that we have reserved, we have booked one of the pieces of land there for expansion of our Lisbon VFX data center in the future. Regarding data centers, there will be a specific explanation by Fran during the presentation. The main highlights of the year is that we signed a block of 15 MW IT in Barcelona at the very end of the year that started cashing in January 2025. We are in very advanced negotiations for the lease of another 18 MW block in the Bilbao- Arasur data center with more than one client.
We are now at the point of exchanging drafts of contract and helping the clients with the prices of electricity and the construction of a feasible PPA in order to ensure competitive cost of electricity over the coming years. Regarding the results on page six of the presentation, you will see that gross rents jumped by approximately EUR 25 million . That 25 difference remained true after incentives. Regarding EBITDA, we lost a little bit of margin as compared to last year. In FFO, you can see that we recovered through the exceptional performance of the cash remuneration at banks following the capital increase. It's been a lucky coincidence that we had a period with track a period of high interest rates for the last second part of the year.
In terms of magnitudes per share, our EUR 0.61 of last year converted in EUR 0.55 this year as a consequence mainly of the capital increase. But you might remember that the capital increase implied the issuance of 20% new shares. So post-money dilution was 16.7%. So that means we have recovered approximately half of the dilution caused by the capital increase by organic growth during the year. And also importantly, because I know some of you asked us to specifically stress the idea, we count the number of shares as number of shares at present. If we were to count the number of shares as weighted number of shares during the year, then the cash flow will be EUR 0.61. So absolutely flat compared to the year before.
In terms of NTA, the 15.08 moved to 14.32, again as a consequence of the dilution caused by the capital increase, corrected partially by the increase in value created by the data centers during the year. For 2025, you will immediately notice, I mean, as it will be mentioned, I guess, during the Q&A session that we are giving a relatively flattish number. There are two reasons why we are doing that. One is because we still do not know which months of the year will be impacted by the new lease in the Basque Country, and second, because we continue to onboard more financial expenses. So particularly after repaying the main maturity of the bond, we will continue increasing our cost of finance. We hit a minimum of 198 in 2022 that has now moved to 246.
We believe we are going to finish the year at around 2.6, and that number will move to around 2.9 by 2026. So as you can imagine, I mean, we are onboarding one full point of extra interest cost, which is around EUR 50 million of cash flow that we need to generate in order to simply kill the increase in interest expenses. There is another reason why our guidance for 2025 cash flow may look like weak for many of you, which is that we don't capitalize interest. We have had some internal discussion regarding that. I know there are different schools of thought. We know that some of our peers do it.
At some point, if we are obliged by either the regulator or the auditor, as it happened with the linearization of the fit-out contributions, at that point, we will, of course, abide by whatever rule is imposed on us. But for the moment, we believe it doesn't reflect a good practice to pro forma your cash flow by kind of capitalizing your interest expenses because in reality, you are showing a stronger image of yourself today at the cost of prejudicing your future cash flow machine. So we prefer not to do it. I mean, we told you in the last call that we will do an exercise in order to give you the pro forma. The pro forma is basically EUR 0.05 more. So instead of EUR 0.54, we will be at EUR 0.59-EUR 0.60 for the year.
But we prefer to do it that way because this is the cash that we actually have at banks as of end of the year. The rest is simply a linearization or a pro forma, which is good, but it's not a true image of your accounting. So we prefer to stay that way and continue giving you the actual FFO, including the interest expense incurred during the year in the development of our CapEx in data centers. And what data centers and some other business lines, but mainly data centers. On page seven, you will see the GRI bridge with the performance of offices, logistics, and shopping centers. On page eight, you will see the evolution of occupancy, including also different asset classes, reflecting what I commented, which is that we have reached a new maximum of 96.7%, increasing close to 60 basis points during the year.
As you all will immediately notice, I mean, these figures will be very difficult to exceed in the coming years. I mean, we can add a little bit of occupancy in offices. We believe the growth in offices will come more in rent, but we can add a little bit of occupancy in offices. But there is very little we can do now in shopping centers or logistics because in shopping centers, as you know, those of you who are familiar with the business, there is a significant rotational vacancy, which always strikes you at every month's end, and December is not an exception.
Regarding offices, as commented, a little weakness in Barcelona reflected in the loss of more than 1 percentage point in occupancy, but a very reasonable re-leasing spread of 2.7%, which is remarkable taking into account that in the past year, we have been eating significantly in our reversionary potential by applying inflation. But the market seems to be digesting that, and the market rents are growing, and we are growing together with the markets. In Madrid, with a very, very active year with more than 130,000 sq m contracted, we have a re-leasing spread of 1.9%. It's a market which is much more stable because it's significantly bigger than Barcelona or Lisbon. Regarding LOOM, on page 12, LOOM is now reaching maturity with 35,000 sq m now in operation and an 82% occupancy in 14 spaces.
The ADR that we are charging at present is allowing us, for the first time this year, to obtain positive EBITDA at all levels, including the overhead of the, let's say, management of the division. So LOOM has been a clear contributor to Merlin since its initial creation by providing the company with a service that our clients demand. Then it moved also into a good contributor in terms of rents, although the LOOM rents represent only around 2% of the total rents of Merlin in the office division. But recently, we are now making money with the business. I mean, besides paying market rents for the properties we use, the company or the division is making money on its own and attracting very interesting clients. I mean, this year, Microsoft, Sanofi, OneTrust, I mean, those are incredible brands that have entered the Merlin universe through LOOM.
It's very interesting because in many cases, those companies then end up expanding into other assets of the portfolio in regular lease contracts during their business lives in Spain and Portugal. On page 13, we simply wanted to bring your attention to the fact that that little trend, which is just emerging of transforming office into residential, it's clearly making us hopeful that some tension will be created in the rental market in the coming years because that destruction of supply that is happening, particularly happened in 2024 in a very clear way. I mean, depending on the source you follow, I mean, whether it's Savills, JLL, CBRE, or Colliers, between 200 and 300 thousand sq m have been destroyed of offices in 2024 and moved into residential.
For those of us who have been working in Portugal for many years, this is a very similar wave to the one we saw in Portugal following the impatriation law. When the Portuguese government enacted the high-value-added professional impatriation law, the residential prices in Lisbon went through the roof. And as a consequence, the highest and best use of many buildings was residential. And many buildings in the Baixa and in the Chiado and in Rossio were moved into residential. And as a consequence of that, in only three years, the little stock, because Lisbon is a relatively small market of less than 5 million sq m counted by CBRE, because probably the residential use as offices can take that number to grow to 10 million, that the offices looking as offices, which is 5 million, all of a sudden went down by 600,000. I remember that perfectly.
The immediate consequence of that is that today we are trading in Portugal at rents of EUR 28 and above. We believe that in our Liberdade building, we can reach higher rents than that. We are probably going to hit the mark of EUR 30. In the peak before the Great Financial Depression in 2007, the maximum rents in Lisbon were in the region of EUR 22. I mean, that was what McKinsey paid for Marqués de Pombal III, which currently is a building of ours. If you compare that to Madrid, you will see that the rents that we attained in the 2007 maximum were EUR 46-48 in Castellana 64, Japan Imperial Tobacco, and Goldman Sachs and UBS. Those rents have evolved 17 or 18 years later to only EUR 40-41 at present in true terms.
I mean, a very different thing is the face rent that some people are bragging about. But then you need to do the fit-out contributions, the free rent periods, etc. But when you calculate the real rents, we are at present at something between EUR 40 and EUR 41 in true terms in Madrid. And we believe that given the effort rate of our clients, which is very low in offices, there is some room for future growth in the rents of city center in Madrid, particularly if the current, let's say, hyper-development of the city continues. I mean, Madrid is now a 7 million people city, including metropolitan area, and is set to continue growing and reach very close to 8 million people by 2030. Madrid is at present economically, is clearly a rocket. It's an island in Spain.
So I simply wanted to bring your attention to this because I know there has been a lot of bad mouthing about offices, and many of you are influenced by the problems in San Francisco and the U.S., etc. But for the moment, I think the office market in Madrid has coped very, very well with work from home, which has been negligible in terms of real effects in occupancy, and also with the densification wave, which, by the way, now is returning. It's coming the other way around. I mean, the pendulum has clearly shifted, and many companies now that they are a little less concerned about ESG and a little bit more concerned about the quality of life of their own employees. So we are seeing some of our clients now redensifying for lower density rather than for higher density within our portfolio.
Logistics, page 15, a very good year, I mean, with good like-for-like. A little drop in Barcelona, which is not a trend. It's simply a consequence of two modules in the ZAL Port Zona Franca that couldn't be relet on time for the year-end cut-off date, but very interesting year in logistics, resulting in that 99.4% occupancy of the portfolio, which is absolutely remarkable. On page 17, you will see that our investment in ZAL Port continues to perform very satisfactorily, although it is clear that it has become now a completely mature and stable business following the development of more than 370,000 sq m since we joined that were vacant. I mean, we're in land status.
We developed a lot in that area, and now we have reached full occupancy and full development, and there is very little we can do in order to continue growing ZAL in its current framework. Shopping centers, page 19, again, super strong year, which is continuing through 2025. Good evolution of footfall, despite, of course, the drag that we have had since 2019, which is the cinema industry, which, by the way, now seems to be resuscitating a little bit. But despite the drag of cinemas, very good footfall evolution and very remarkable rents and sales evolution. It's incredible how strong consumer spending is. Those of you who know me, you know my opinion.
We are a little bit doped by a lot of public spending, but one way or another, this is going into consumption, and at the end, it's reflecting in the performance of our shopping center portfolio. And then I will pass the floor to Inés, who will talk to you about valuation and the position of the company.
Thank you, Ismael. Good afternoon, everyone. So as commented with Ismael, we go to slide 22. Valuations are pretty much in line with 2023, which stands at EUR 11.5 billion, reaching a 5.3% passing yield for the entire portfolio coming from a 5.1%, which derives or it implies a 4.5% net initial yield. In terms of how we've come to this flat valuation, you see that in the existing portfolio, we've had a 21 basis points expansion in yield with. That's on a gross basis.
But on the net basis, it's been absolutely negligible for office and for shopping centers, slightly more impact for logistics. But this says about EUR 16 million of impairment has been netted off by the 17% uplift in data centers. As you know, data centers, we carry Phase II and everything that we have in land bank at cost. So this is just part of the valuation that will be created, and valuers are taking a conservative approach in these new asset classes for us. Moving to page 24. Well, as you know, we have a very sound financial structure. Thanks again for the capital increase that we carried out in July. That was a great achievement. And obviously, that was very well supported by you, by our own shareholders.
But as well, this position has been strengthened, or there's been a big achievement, as Ismael commented before, having an upgrade from S&P and Moody's on the basis of sustainable lower leverage and expanding cash flow generation. We stand at EUR 4.9 billion gross debt coming from EUR 4.5 billion, as you know, and we've commented this many times in the year because this happened throughout the first half of the year. We raised almost EUR 284 million of mortgage debt, and we also did a 100 + on the September 29 bond. Those are the main actions that have taken place on the debt side. And the implied LTV, the 28.3%, it's about less than 10 x net debt to EBITDA. We are at 8.8 x net debt to EBITDA, which is a ratio that more and more creditors and shareholders care about.
We've also commented about the average cost and that this average cost is going to increase definitely because we are refinancing our cheapest bonds. Sorry. 14% of our debt, and we are in page 25, as you can see, is mortgage-based, but we remain with more than 85% being corporate. And again, our maturity for on the 26th of May, the bond, the EUR 600 million bond maturity, that's done. We did the homework back in 2023, and the next maturity only comes in November 26th. So we still have a lot of time ahead of us. We'll probably start to tackle it at the end of this year and on an opportunistic basis. Moving to investments and divestments. So I'm going on slide 27. Well, again, you see there's been a modest activity, both in investment and non-core divestment in the traditional asset classes.
A little bit more active in data centers, but Francisco will comment on that in a minute. And in terms of what we've done, it's basically consolidated the ownership of the adjacent building of our landmark Plaza Ruiz Picasso office building in the prime Azca area of Madrid. So that's only 4,500 sq m building. And regarding divestments, again, EUR 73 million disposals are double-digit premium, 10.8% to GRI. That's below our EUR 100 million target just because some of those transactions have moved to 2025, sorry. We've been very focused on reconversion opportunities to residential, either selling land, office land, or non-core office buildings to be reconverted by next buyer. And we're talking about 33,000 sq m in 2024, and we have identified another 15,000 sq m in our own portfolio that could be reconverted in the future.
In terms of value creation for logistics, as you all know, our pillars of growth come from data centers, mainly, and from logistics as well. Data centers, again, will be explained by Francisco, but I'm going to walk you through very, very briefly through the logistics. So our total land bank amounts to 511,000 sq m + we have refurbishment in the Vitoria-Jundiz Warehouse. It's 73,000 sq m , for which we are already in negotiations, advanced negotiations for pre-let. So overall, we have 40% of our portfolio with either pre-let or head of terms. We ran out of land in Lisbon, as Ismael commented before, and out of the committed pipeline, we are expecting to deliver all these warehouses, except for 68,000 sq m warehouse in Lisbon.
We're expecting to deliver this on the first half of 2026, and the rent that will be generated, they're around EUR 20 million, EUR 19.6 million, which entails a 7.5% yield on cost and obviously a double-digit yield on CapEx by investing EUR 171 million. For the future, we have another 244,000 sq m of land that could generate around EUR 13.8 million, also implying a 7% yield on cost or an 11% yield on CapEx. With no further delay, I'll pass the floor to the most interesting part of the presentation that will be carried out by Francisco. Thank you.
Many thanks, Inés, and good afternoon to everyone. I'm glad to cover now our Project Mega, focusing on the achievements completed over 2024, but also giving some guidance on our expectations from an operating point of view for 2025 and following years.
So as you can see in page 31 and 32, we have summarized the current positions of our data center division that we then call Merlin Edged for Iberian Peninsula, and it shows both the original two first phases that we have, that we named Phase I and Phase II, but also the former Phase III and Phase IV now remain as upsizing for that capacity that is located in the former four locations that we selected, which are Madrid, Barcelona, Bilbao, Arasur, and Lisbon, and the future pipeline. So in particular, I would like to draw your attention to the fact that this division has significantly changed over 2024. As you know, basically, we have full-speed the equipment of the current three assets already built, but also we have started the construction of Phase II following the capital increase completed at the end of 2024.
On page 32, we are updating the information on a phase basis and on an asset basis that we will comment in more detail in a second. The first point you will notice is related to the fact that now Phase I comprises 64 MW as compared to the 60 MW we had before, and Phase II comprises 203 MW instead of 200, and this is due to the fact that this division has now entered into a, let's call it, industrial dynamic where we are not any more comprised by phases, but just focused on projects and the demand driving those projects. We need also to consider that in this asset class, we benefit from the fact that all the equipment and construction we are doing is modular, so basically, we can move from one side to the other if needed.
It means that probably for next presentations, instead of allocating everything to Phase I and Phase II and then have this type of movement that we've been explaining in a second, probably makes more sense to have in line with other asset categories, operating assets, WIP, work in progress, and pipeline. So if we go to the lower part of the table, which is describing basically on an asset-by-asset basis, the first thing that you are seeing is that that increase in Phase I is due to the fact that now the 6 MW of re-powering Barcelona is, let's say, moving ahead. We got confirmation from the energy company that we will be having connectivity to this increase of power in the first half of 2026. So meaning that we move that capacity, comprising this in the full building for Phase I.
Also, with regards to building three in Arasur, we have now, let's say, achieved maximum design capacity at 20 MW. Originally, we thought we could go to 24 MW but instead of the different level of density that the client is requesting. Basically, we have some limitation on the roof, so we will fix that at 20 MW. In exchange, that power coming from the same service station, we are using it in the other buildings of Bilbao- Arasur, of building two and building one that we have in Phase II, reason why basically that capacity is increased now to 96 MW in total.
Also, we are getting following a little bit, and we will describe this in a second regarding Madrid- Getafe, the possibility of increasing a little bit the power there, which also we are setting now in Phase II, but as soon as we get confirmation, probably we'll move as well to Phase I . That's the reason that probably going forward makes more sense to talk about operating assets with a pipeline. In terms of just upsizing, the news there, as we will talk in a second, is the inclusion of a new project called Madrid North with 30 MW immediate and 130 capacity of upsizing to 130 that we have allocated to pipeline as well as the two projects in Extremadura. So going to more detail on a phase basis, I'm page 33 right now.
We have seen basically that we have changed a little bit this slide as compared to the previous one. The reason why is that during this year, we have advanced negotiations for several assets, and with that, we are allowed basically to reduce the expected stabilized year from the original 2028 into 2027. In terms of equipment, we ended the year as reported at 26 MW fully equipped and that will jump into 32 by the end of this first half, and we will achieve 58 MW by the end of 2025 in fourth quarter. We are taking, we'll comment in a second as well, that we are taking a bit of equipment in advance of the power that we'll receive for Getafe in 2026.
The only part basically that for Phase I that is pending to be equipped in 2025 is the 6 MW of the repowering of Barcelona. Basically, the reason why is that we ordered now, but as this implies basically some equipment linked to generators, the timing of receiving that equipment matches with the time we are getting basically the supply as in first half of 2026. As compared to previous slide, 2027 has disappeared, and now 2025 and 2026 will, let's say, capture all the equipment that we need to deploy in that phase. That basically has moved us to EUR 88 million in terms of stabilized rent and just ran a little bit up to 14.5% stabilized gross yield on cost for this Phase I.
Moving to page 34, and in terms of commercialization, as Ismael described before, for Barcelona, we were able to fully let the assets to an AI hyperscaler at the end of last year, so the only capacity we will have there is the re-powering, the 6 MW as soon as they come. And by the way, we give a high probability that the client, which is the largest client we have in that asset, will take as well this capacity when ready. Regarding Bilbao- Arasur, we commented in last presentations that we have two options here because we were in conversations with a large hyperscaler for building one, and so for building three, the one which is already built, you see there, and the second building.
As our CEO commented at the time, this was basically something that we fought a lot because in one side, you can let two assets in a row, of course, in exchange of sacrificing some level of rent. But in the other possibility, let's say, was to continue with other clients and letting little by little with higher rent, but not basically with such amount of commercialization in one spot. So we decided to go with the large client. Over 2024, we passed all their onboarding technical homologation successfully, which basically comprises not only the equipment, but also operations and even financial capabilities to provide this type of services. But we knew of both parties that basically this building will be fully equipped by mid-2025. On top of it, they require some level of investment, additional investment on special fittings of the rooms.
And in order to make or to comply with this date, we set a deadline in which basically the client needs to decide whether they go binding with us so we can guarantee this ready-for-service date or not. The deadline basically or the time basically elapsed, and the client was not in a position to be binding for that. So what we did basically is just to resume the previous conversation we have with the other two clients that are right now basically competing for that space, and we are open to them that they are not in exclusivity. Okay? Curiously enough, one of these clients is requesting, but by the way, both clients are going for the full capacity for 18 MW that are available right now in this building which is constructed.
One of those is requesting us similar requirements to the one large hyperscaler that we were commenting a minute ago, and also asking from our side basically an option for building two, which basically gave us a little bit of hint that probably the former client is still around. Here is an important thing which basically we are seeing, and we have received several questions on that sense, which is basically the amount of projects that are being announced globally and in Europe and, of course, in Spain, which basically do not show. They show a lot of capacity and a lot of access sometimes to power, but the reality is that there is not clear ready-for-service date, which is what is important for this type of clients.
They need to make sure that when they receive the equipment, which are a large investment, we have commented several times that it's sometimes between two and four times the amount of money that we are investing equipping all these assets. So these guys, when they get all of these commitments, they need to make sure that once they receive that equipment, it's a place to put them in service. So all of this noise, what it's doing is that at the end, they are super reluctant to commit to certain dates if they're not seeing that the landlords can deliver that space. This has provoked that a lot of, and of course, these situations provoke that new clients come into the scene.
In this case, basically, there is a new concept which is AI as a service of companies which are getting a huge amount of capital invested because what they provide is not only the space, but also the servers already connected, which makes that these large hyperscalers at the end, they use this type of providers to secure or to enter into areas which are already fitted out and ready for service. And probably this is what we are having right now, which is in front of us. Regarding Madrid- Getafe, we have here was a mistake. We have 70% booking. In reality, we have 100% booked for the capacity which is available right now.
It's the same client that we were reporting time ago, and the fact that it basically is not converted into lettable spaces because this client wants the additional capacity of power that will come in the first half of 2026. So we were holding that conversation now that we are approaching that available capacity is what basically is making more probable that we can convert this booking into lettable space as soon as we have a certain date. We are prudent people, so what we don't want is to commit to a date if we don't have certainty and consider that we have suffered several delays for different reasons to get supply of that power. As soon as we have a certain date, then is when we will transmit this to the client and sign it. Moving now to phase what we call Phase II in page 35.
Following the capital increase, basically, we have started works in the sites, mainly Bilbao Arasur and Lisbon. We have first paid, of course, the municipal tax to initiate the construction. We start compacting the land. We have tendered the construction works, issuing the purchase orders for equipment. We are incorporating new staff for this asset. So all of this has been started right after the summer once we were in a position to make this type of commitments. That's the reason why we are reporting here CapEx commitments. Here in this asset class, there is also something which is specific regarding the equipment where you normally do not pay a lot of amount of money upfront. You normally limit that to 10%-25% of the total investment.
And then, once you commission, once you receive and put in service the equipment, is when you pay another significant part of the payments. And of course, you leave some percentage basically for several months as a guarantee of that good performance. So what we are reporting here is the CapEx commitment. Of course, in order to incur in this CapEx commitment, we need to have the counterbalance of associated funding. And now that with the capital increase, we have it is when we have launched all of these orders that are basically drafted here in those amounts. On page 36, starting with Madrid- Getafe and taking advantage of the experience we saw in Barcelona where we are re-powering that building, we have also asked for the same type of increase of power. There is a particularity here in case of Madrid as compared to Barcelona.
In Barcelona, basically, the new power is coming from a second substation, not from the original one, which is from a technical point of view, it's a little bit more complicated in order to isolate one part of the building to the other. In the case of Madrid, it's a little bit easier because it will come through the same line, which in other words means that the delay we are suffering in getting that power supply could be partially compensated for the fact that we could get a little bit more power and basically improve a little bit the numbers once we are in full operation. Regarding Bilbao- Arasur, we have completed the preparation works for the land for the second building. We have started piling on site at the beginning of February.
Once we have finished already the tender of the construction companies, we are starting as well construction there. Regarding the equipment, in this case, we are ordering for building two full equipment, the 48 meg in one shot and not in different phases as we did in the past. For third building, we are progressing with the construction license, and we expect basically to be ready to start construction by beginning of 2026. Finally, Lisbon, we have started compacting the land right after the summer, which is progressing as expected. We are incorporating the piling as soon as the preparation of the land is being completed in the different phases.
In this case, as you know, it's a little bit different to Bilbao- Arasur, and this is just because of the proximity we have to the Tejo River, which basically means that we have a longer period of time of compacting and piling as compared to what we have experienced in Arasur. We expect to have by last quarter of this year all the works completed, not only for the first building, but for the full campus. And what we can do basically is start a little bit before on the construction of substations and generator building and first building as soon as the areas basically are completed. In parallel, what also is a change that is interesting to comment is that seeing that the clients are requiring a higher density than the original projected.
Remember that here what we have for this 100 MW is five buildings of roughly between 18 MW and 20 MW each. What we are doing right now, what we have already requested and is looking very good, is to increase the density per building, increasing it to the 36 MW, which basically will mean that we will basically absorb that capacity of the 100 MW in a couple of buildings, and then we can move what we have commented for Phase III upsizing into the same plot. In commercialization, we are advancing with one particular client, but we need to be conscious basically that we need to confirm at some point in time the ready-for-service date. So as soon as we are advancing, we will have more clarity on these conversations.
Finally, as well as we did for the campus in Arasur, that's where we are basically, as you know, setting on-site photovoltaic plants very close to the site. We are doing the same in Lisbon, which is basically another of the requirements that normally the clients requested, which is basically to have energy production very close to the data center. Moving to page 37, the former Phase III that we call now upsizing. On Lisbon, as commented, basically this 80 MW on re-powering will be used in the first five buildings. The 120 MW extension links to the adjacent plot that Ismael commented at the beginning of the presentation. And the news here basically are the inclusion of Madrid North, which is a project that we have closed, I mean, just this week, with an initial amount of capacity of 30 MW.
But very interesting, this land allows us basically to go with further capacity that we are planning for 130 MW more. And basically, due to something which is pretty interesting, which is that the substation that provides this capacity is completely adjacent to the plot we are talking about. It's a big plot of 200,000 sq m of buildability. So there is capacity there to grow. And so what we are seeing right now in Phase III as an upsizing of another project in Madrid North of 30 MW, in a way basically to cover as well the city with Getafe could be converted in the future into a data center campus as we have in other parts of Spain.
Finally, moving into page 38, this is Phase IV with two projects that two of them basically we announced already, which is the one in Navalmoral de la Mata and Valdecaballeros in the region of Extremadura. As commented in our announcement, the reason for selecting this area is because this is the region in Spain where there is more sustainable energy generation area as compared to the consumption levels of the region. In terms of capacity, this area comprises roughly like 30 gigawatts of capacity installed, and only close to 5, not even 5 GW of capacity is the one that the current region is using, which means that it exports 25 GW of capacity. Taking that in consideration, we have been looking for a while for projects that could make sense in that region. We found two of them.
One is Navalmoral, which especially of this project is that this is a very well-connected area. It's very close to the motorway, which means that it's basically next adjacent to the fiber cabling that connects Lisbon and Madrid. And also because it's urban land, so it's ready to be built as soon as basically we have to find clients for the project. We have announced that a little bit ahead of a big conference that was held in the U.S., at the beginning of the year, PTC. And after sharing this with the big hyperscalers and clients that were taking part of that congress, we received a considerable interest of several parties. And in particular, one which is very interesting on the two sides. Navalmoral is more immediate, but Valdecaballeros still needs to be done some urban approvals maybe through the connections to the site.
So I mean, this time passes very well the needs of that client. So we will basically comment in future calls as soon as we progress with that. So in summary, 2024 has been a game-changer for our Merlin Edged division. Phase I at full speed of capacity. Phase II already started and with a lot of work in terms of construction, delivery of capacity and commercialization during 2025. But we think that we are very well-positioned as one big operator in the sector in Europe.
Okay. Thank you, Fran. Well, basically, just to summarize and close today's exposition, the company has clearly delivered very strong performance in all key financial and operating metrics during the year, like-for-like growth, occupancy, re-leasing spread, FFO generation.
So it's been a great year, a year that started a little bit dubious in the first quarter, but at the end ended up being a very, very good year for the company. We have reached historical maximums in occupancy. Our traditional asset classes are very close to full optimization, which goes hand in hand with what we have been telling you for many years now, which is that the reason why we decided to move into the data center business is because we felt that following the digestion and owning of the mature asset class portfolio, there will be a point in which the company will have no significant further growth to offer to the market.
And we wanted to find a new area of activity that was sufficiently enticing for the market in order to continue growing the company and convert it into a significant player in the European arena. In terms of value creation, we will continue during 2025 to develop logistic land. We have a wonderful portfolio distributed in key locations in Madrid, Lisbon, Valencia, but now the refreshment in Vitoria and Seville. And the yields on cost that we can achieve on the development of our own product are very significantly above what those similar assets trade for in the secondary market. So we will continue insisting in that route of growth. And the main route of growth will continue to be our mega plan, as Fran was commenting.
This is an important year for us because even logistically, we are receiving about 26 MW-32 MW of equipment before year-end that we need to store and install in order to have it ready for receiving the power during the first half of 2026. In the cases of Madrid and the re-powering of Barcelona, we need to start operations with whoever the final client is in the Basque Country during the second half of the year. We have only included in our cash flow projections a couple of months of cash flow in the Basque Country data center. If we can clinch a contract a little bit earlier and we can obtain some extra cash flow, that would clearly add into our guidance for the year.
We continue beefing up our options and properties with a view to honing our mega plan portfolio in anticipation of what we believe is going to be a very significant wave of speculation once the available energy positions of the different substations across the national territory are made public during the year. I mean, that will clearly spark a wave of brokers and traders trying to assemble land with electricity in order to basically flip positions, and we are one of the few real off-takers that exist in the market, so instead of simply waiting for traders and intermediaries and brokers to come here and offer us positions with electricity at gold prices, we are taking a more proactive stance and accumulating some land banks similarly to what we did in logistics in 2016, 2017, and 2018 in order to enhance a little bit the profitability of our projects.
Because yes, land is not super relevant as a cost for the total cost of a given data center project, but better get cheaper land than more expensive if you want to extract that little extra profitability that we want to extract in everything we do. Of note, the agreement with the regional government of Extremadura in that respect, the first plot in Navalmoral de la Mata is fully permitted and infrastructured, and electricity is within close proximity, so it is a location in which we could eventually start had we closed an agreement with an existing client, but this one is relatively ready and available.
The one in Valdecaballeros will take more time because we need some road infrastructure to be done by the authorities, and we also need to bring the dark fiber from we are deciding which route are we going to take in order to bring the fiber to the location. But it's also a very interesting location with a sea of renewable power immediately around our site. I mean, 1.3 GW of photovoltaic, thermosolar, and hydroelectric plants in very close vicinity together with direct line to the nuclear plant of Navalmoral precisely, which is the one that is also giving us the backup for the same project. Regarding outlook for the year, as commented, 2025 will be a hard hats and helmets year. I mean, we will be building. Data centers will continue draining resources from the company during its buildup.
It will be a negative contributor of cash flow during the year, despite the fact that this year, at least the income figure is going to be much more meaningful. The one we have in the budget is around EUR 25 million. So it will be much more meaningful. But still, cost will outweigh income production in the data center division. And this will continue basically through stabilization. I mean, we will clearly be positive in 2027, but during 2026, ± , maybe we'll get into positive territory, but it will not be meaningful for the company. The year will be uneventful in terms of debt maturities. The next one we have is in November 2026. So towards the end of this year, beginning of the next, we will start moving in the market and with the bank community trying to prepare in advance for the November 2026 maturity.
Clearly, if we move too early, we get extra cash, which is not good for the company because at present, the cash remuneration is not good, and the extra cost of taking additional debt on the balance sheet is not good. But we are prudent people. So if we need to sacrifice a little bit of cash flow, we will do it because we want full safety and make sure that in November 26th, we repay our existing bond. We are suggesting to our board to pay another EUR 0.22 per share dividend on account of the 2024 financial year. That should be recommended by the board and approved by the general shareholders meeting that we will be holding in April and paid during the month of May, more or less. The estimated FFO for 2025 will be around EUR 0.54 per share, flatish as compared to last year.
I mean, beg your pardon, but as commented, we have significant increased financial costs that we are digesting into our P&L, and also, the data center division is growing, but growing means also staffing it up and taking a number of expense decisions, which are for the moment detrimental to our P&L, but I'm sure we'll bring very significant growth in the future. The data centers, many people asked us this morning how diluted we expect them to be during 2025 in the region of EUR 0.04 per share, so in the absence of data centers, we will be at EUR 0.58, around EUR 0.59 in terms of guidance of FFO. But of course, we need to continue paying for that division till it is mature and a big contributor to the P&L of the company, and this is basically it.
I mean, I see that the market didn't like what we did in 2024. So I took the opportunity to buy some additional shares this morning. Thank God. I mean, market volatility is always good for people who are in the know, understanding by in the know that we, I mean, technically, we are out of blackout period since 8:00 A.M. this morning. So I bought after that. But I fully believe in what we are doing. We are not here in the data center business looking for extra hype in our share price. We are not here simply to brag about the fact that we are now part of the AI noise. We are here in order to look for cash flow. I mean, we are trying to build a significant revenue contributor to the company.
We are here in order to double up our cash flow generation and, if at all possible, also double up our dividend payment capacity. So that will take time. It requires a lot of patience. Patience is something that the markets normally do not have. So I know it's going to be a bumpy road between now and 2027 when people start to see that there is something behind the data center name. It will be a bumpy road, but I am fully confident on what we are doing, and I'm sure that you will all enjoy the fruits of what we are seeding today a couple of years down the road. Some of you have requested some remark on what is happening in the world. Of course, I am not a geostrategic expert. I have no fucking clue of what lies in the mind of Mr. Trump.
I only know that we operate in Spain, Portugal. Both countries bear a very significant deficit with the U.S. in terms of trade. So if tariffs are imposed on Europe, it will not be the guilt of Spain and Portugal. I mean, Spain particularly carries a 10 billion deficit in trade with the U.S. I know that Europe, in general, bears like a 50 billion super added, mainly in the export of goods because we are deficitary in the import of services, but clearly not Spain, Portugal. Spain has been put on tier one in terms of import of technological products from the U.S., which is very good news because that was really reason for concern. So we can enjoy here, or our clients can enjoy here, the state-of-the-art technology in terms of GPUs, CPUs, TPUs, and everything, which is very, very good news for us.
And we will continue trying to make sense of our relationship with our current U.S., and European clients in the future in our data center division. So without further delay, I think we can move into Q&A. I pass the floor to Teresa, who will organize the Q&A session. And we will be here to respond to your queries and questions as best as we can.
Sure. Thank you very much, Ismael. Just to remember, if you want to ask a question, you have to press star five. The first question comes from the line of Marc Mozzi from Bank of America. Marc, the floor is yours. No, sorry. So the first question comes from the line of Florent Laroche from ODDO BHF. Florent, the line is yours.
Hi. Good morning, Ismael. Good morning, everyone. So thank you for this presentation. I would have three questions, if I may. So my first question would be on the advanced discussions that you have for 18 MW in Bilbao. So I understand that you don't know when it will start, but maybe could you please tell us what you have taken in your guidance, if you have taken something for 2025? My second question would be for offices and shopping centers. So we understand that you have optimized notably occupancy. So what kind of dynamic can we expect for 2025? And my third question would be on data centers. So we can see that you increase a lot the potential IT capacity in terms of megawatts. So all these developments cost a lot of money. How do you manage your development risk on this data center before starting development of a new building, of extension, of new works? So that would be my three questions.
Okay. All right, Florent. Look, regarding the first one, I will let Francisco respond to you. Basically, I guess you were referring to the fact that we had conversations during 2024 with a certain client for building three and two in Bilbao, which now has been replaced by conversations for building three and then an option on building two. You can explain.
So the demand basically for building three will be ready, so what we call ready for service by the end of second quarter 2025, so end of June, beginning of July. Once you basically, if you have a contract signed, you need to count basically you have a couple of months where the client normally do their fit-out, or we do the fit-out for them, which is normally what we are doing so far because of the nationality of the client.
Then once you have that ready, it's basically when they start paying. So that's the reason why Ismael was counting. We were saying before that we were counting for a couple of months of cash flow during this year for this asset. Things can change depending on the level of fit-out that finally they convert, but this is basically the visibility we have right now. So that's basically the impact. So the guidance basically is based on that assumption.
Okay. Well, regarding offices and shopping centers and the dynamics that we can expect in 2025, I think that in shopping centers, increasing occupancy will be a hard thing. So occupancy will remain more or less flattish. We are going to continue capturing some reversionary potential, mainly through variable rents and more income because the centers are better attended. I mean, there is more footfall. And as a consequence, our tenants are selling more. So the marginal propensity to pay rent increases. And as such, we could capture part of it in the rent renewal plus inflation. So the like-for-like will not come by positive variation in occupancy. It will come mainly through re-leasing spread and a little bit of inflation passed on with maybe a little income from more spaces and the put into operation of Marineda if and when it is finished. I mean, we are finalizing the construction works. It will mainly hit 2026, but there will be also some contribution, very little, in 2025. In offices, in principle, we are projecting a relatively flat year too in terms of occupancy. It is not easy to grow occupancy significantly with a big portfolio like ours.
But I know that during the year, we will also be negotiating a number of big listings that eventually, if you hit one of those, you can add a little bit of net occupancy to your portfolio. So very interesting dynamics in that market. Rents are clearly going up slightly, but steadily. So I like that. We are passing on inflation without a problem. So the like-for-like in this case, occupancy, we cannot bet on it, but we could add a little bit of occupancy. Clearly, rents, re-leasing spreads will little by little recover. I mean, we have had a period of relatively low re-leasing spreads because we ran too fast by applying inflation following COVID in the high inflation period and took a significant dent into reversionary potential. But now, little by little, we are again creating a cushion between passing and reversionary.
We should be getting some re-leasing spread during the year with positive inflation. And regarding data centers and money funding, two things. One, Phases III and IV, don't even bother about them because physically, logistically, we will be occupied building Phase II till 2029. However, I would also like to introduce another concept, which is fungibility. So don't be surprised if in Phase II, we decide to run faster with the first building in Navalmoral and slower on the third building in Lisbon. If that happens, don't be surprised. I mean, as Fran commented, most of our equipment now is on purpose is modular. So the gensets are exchangeable from our different skids. The UPSs, the transformers are now exchangeable among our different sites. And depending on demand, we could adapt by moving that demand to one or another particular place.
But in summary, we will be building 200 to 200 and change megawatts of capacity between now and 2029. Don't be led to believe that we are going to be building 1 GW of capacity between now and 2029 for one very simple reason: because it is impossible. You need to send project managers, engineers. You need to have purchasing managers. You need a lot of logistics. We, on purpose, chose to do second phase mainly in two existing locations. Had we decided to do second phase through 10 different locations across Europe, I would have perceived much more significant risk of execution in that plan. So we wanted to be strong in the locations that we know and where we have already a lot of personnel deployed. So playing with that, think about this concept, which is fungibility.
Eventually, we can exchange or move parts from here to there, but we are going to be building exactly the number of megawatts for which we have been funded through the capital increase and with the eventual bond tappings that we will be doing during 2025 and 2026.
Okay. Thank you, Ismael. The next question comes from the line of Marc Mozzi from Bank of America. Marc, the floor is yours.
Yes. Thank you very much. Sorry for messing up with the star five. I have three questions. The first one is regarding your 2025 FFO guidance. Can you give us a bit more color around what are the moving parts, specifically on the financial cost, net financial cost, including potential income from your cash in hand, which is EUR 1.5 billion if I'm correct? And if you were to give us a high range, what could it be? Because if I understand you correctly, your EUR 0.54 of FFO in 2025 is a kind of a very, very conservative guidance, and you're going to easily beat it. So just give us a little bit of color on that one. Thank you. That's my first question.
Okay. All right. So if I may repeat.
Okay. So basically, on the moving parts, Marc, as Ismael commented before, our data center ramp-up is costing us, let's put it this way, even though we are increasing our top line in data centers by EUR 23 million, roughly, in 2025, from EUR 2.2 million- EUR 25 million, that's being eaten by the properties and the overhead in the division plus the financing costs that we need to bear. As you know, the homework that we did in 2023 was great because it enabled us to pay the bond in May, the EUR 600 million bond in May with available cash. That's great, but it's also costing us more financial expenses. So these are the two moving parts: the data center drag and the huge financial interest that we're having, which on a net basis is even higher because, as Ismael commented before, in 2024, we've been benefiting from nice financial income coming from the deposits. But unfortunately, we cannot get that remuneration anymore for the cash flow that we still have. So having a lot of cash is great, especially for rating agencies, but it's not very efficient if you wish as we move forward with the data center rollout.
If you want a, let's say, range or a bracket, I can tell you that it's complicated to provide at present. I mean, particularly without knowing the date of operation of the Arasur lease and the hedging and the evolution of the interest expense during the year. On a good day, we can move to 55, but don't expect, I mean, 58 or something like that. Be mindful of one thing, which is that we, again, as commented, we don't capitalize expenses. So if we were to capitalize expenses, which is probably the figure that many people have in mind, that 54 will be 59, and that 55 will be 60. So cosmetically, the figure will look much higher. But the reality is that we are spending a lot of money. Just last year, we grew the gross financial debt by EUR 400 million.
This year, the gross financial debt will continue growing as we continue spending in the different CapEx lines. As a consequence, there is a significant interest cost that we need to swallow and we need to digest until we reach stabilization. The bond that we are issuing now in May has a coupon of 1.75%. Our current cost of financing for long-term, something between seven years and beyond, it's well above 3.5%, I mean, between 3.5% and 3.75%, depending on the tenor. The bond that we are going to be issuing next year in November has a coupon of 1.875%. So all that needs to be accommodated into our P&L. However, at some point, we will reach a steady state that will not be exceeded. Once we are there, our P&L in terms of financial expenses will be absolutely stable.
But for the moment, we are still adjusting to the new environment in interest rates. Okay. Thank you very much. My second question is on your dividend. I'm not sure to understand why you felt the need to cut it by 9% because we're talking about EUR 20 million of saving out of CapEx of EUR 2.5 billion of additional income of roughly EUR 400 million . I'm not sure I understand what sort of signal you would like to send us from that cutting dividend. Look, Marc, in reality, we are not cutting. I mean, the figure we are distributing is exactly the same. And we also are continually adhering to the same policy, which is basically 80% payout based on our existing adjusted FFO. We have taken into consideration your reflection and your commentary.
We have been thinking about what to do because it will be very easy for us to increase the dividend and pay 0.44 EUR. As you rightly commented, it's only EUR 20 million more. But two things. First, there are people in our shareholding base who are, of course, dividend seekers, and we need to please those. But we also have people in our shareholding base who reason in a different way and say, "You have raised money from the market in July and are giving back money to the market in the following two years." Why? I mean, I prefer that you spend that money in more generator sets for your data centers.
I mean, don't give it back to me because it is not very sensible from a purely financial standpoint to be asking money with one hand and be giving back the same money to shareholders with the other hand. I mean, so we prefer to be a little bit orthodox. I know we run the risk of looking idiot. We run that risk in many decisions we take, but we prefer to be a little bit stringent and make sure that we do things by the book and hope for the market to understand that when you do a capital increase, your EPS gets a little bit diluted, and hence your DPS gets a little bit diluted, and take it as a normal thing and then hope for a better cash flow and DPS generation in the future. That has been our reasoning.
But we don't have the revealed truth in our hands. So it might well be that you are right, and giving EUR 20 million extra this year and EUR 20 million extra next year and keeping the dividend at 44 is better understood by the market. I mean, I can tell you we still struggle with Spanish media and the regulator to explain that we pay dividend twice per year because in most newspapers, we are purported as paying 1.9% dividend because it's twice per year, and they take the last figure reported, and they only take half. So sometimes, with your best intention, you take decisions, and then the market doesn't even fucking understand what you're doing. So I know it's hard, and you could be perfectly right, Marc.
But we have taken this decision, and I beg your pardon, and I hope you understand that there is no big difference in giving EUR 20 million more payout to our shareholders while we are still spending heavily in data centers. Yeah. Fair. It's just sending the wrong message. That's the only point.
And my last question is regarding your NTA. How much of your data center capital gains, potential capital gains for Phase I has been yet recognized in the NTA? Or let's put it differently, how much do you still have to recognize on your Phase I in terms of capital gains in 2025 and 2026, of course?
I'm sorry. You're talking about how much has been the level?
No. No. No. No. No.
How much equity gain?
Yeah. What is the expected equity gain in the future? I mean, to me,
Or how much on the other side? I mean, I don't want you to give us how much you're going to expect, but give us how much you have recognized already as capital gains in your Phase I?
Our data center division presently is valued at around EUR 600 million. And to me, if you take NAV x multiples and you take into account that cash flow production is going to be in the region of EUR 85 million-EUR 90 million, I think they should be valued at no less than EUR 1.2 billion, no less. So you still have the same amount to be recognized. Another EUR 600 million.
Just for Phase I?
Just for Phase I, yes. And then you have, of course, all Phase II and all the land that we are accumulating, which, as you know, we carry that cost.
Okay. Thank you very much. Very clear from my side. Thank you. All the best. Okay. Cheers.
Thank you, Marc. The next question comes from the line of Thomas from Deutsche Bank. Thomas, the floor is yours.
Hi. Good afternoon, everybody. A couple of questions. The first one is on data center demand overall. I mean, there have been some concerns recently, as you know, triggered by DeepSeek and also maybe the America First approach. Just wondering, what's your experience so far? That's my first question.
Okay. Well, of course, we were concerned like everybody else when we saw the DeepSeek news. However, what we did is basically consulted with our own engineers and with our clients. I mean, we are in the middle of a certification project with NVIDIA as a preferred partner. And we asked them, and we asked the hyperscalers, and we asked the artificial intelligence service operators. And we saw little preoccupation. And the truth is that we saw little preoccupation.
What they tell us is that neural networks exponentially consume more energy as you add one layer of complexity, so once you have a neural network which is already working, you add one extra layer, and the consumption doesn't increase by 10%. It increases by significantly more every time you make more complex your model, so together with that, we haven't seen one client even blinking in the current conversations. I know you are all concerned about a certain TD Cowen report stating that Microsoft let go of leases in the U.S., amounting to 200 MW, etc., well, all I can say is that I don't know what were the specifics and the circumstances of those negotiations and what were the specific reasons why Microsoft let those contracts go, but I don't see any abatement in the demand at present in the market.
What I do see is that, particularly with hyperscalers, is that as much as they are wonderful, magic organizations, technologically speaking, and they do wonders in what they do, organizationally, they are complicated. They have grown very, very big. And in some cases, the administrative, the admin that accompanies that growth hasn't grown at the same speed in terms of efficiency, etc. So they tend to be very, very, very complicated animals. The infrastructure people can be very eager to take a certain position. They might drive you mad with different iterations of equipment layouts. Now I want liquid cooling. Now I want air. Now I want 130 refrigeration. Now I want only 60. Now I want larger racks. Now I prefer shorter racks. So they can drive you mad with all that. And you have a frank and open dialogue with the infrastructure people.
But then when the technical conversations are finished and they go for approval to California, sometimes it is quite complicated. So to fill that void, what we have seen is that they are using the AI as a Service guys because the AI as a Service guys are younger companies, much lighter, much leaner in terms of structure. They are full of cash, and they are taking blocks of capacity as if there was no tomorrow. And in many cases, at the end, they end up putting that capacity at the service of the same hyperscaler who couldn't get the approvals on time in California. This is what we see. We see also some Chinese demand in the market, which is new, very interesting.
We are starting to see some hints of European demand, which is a very, very promising circumstance, particularly because by operation of the EU Data Sovereignty Act, starting 2026, about 3 GW of IT capacity need to be repatriated to Europe, which at present is completely spread throughout the world, so needs to be repatriated into Europe. So for the time being, we are not super worried. We will grow with the market. Of course, our long-term ambition is simply to have a fair market share of a tiny market like the Iberian Peninsula. So at present, in Spain, the total installed capacity is like 140 MW. Come 2030, the estimates are that it should grow to something between 600 MW and 800 MW.
All what we want to have is a fair share of that, which is between 200 MW and 300 MW, which goes hand in hand with our Phase I and Phase II deployment. So not overly concerned about the worries that are now isolating the market regarding DeepSeek and etc. We believe that there is enough space for us to continue growing in our home market.
The other question is on pricing power in the data center business. I mean, how would you assess the situation? I mean, it seems like pretty strong on your side. Maybe also what was the reason momentum of market rental growth and what do you expect ahead? I think currently it's the EUR 120 level per kilowatt per month. If you could provide some color on this.
Pricing power, Thomas, at present, even we are talking only in abstract because very few leases are being signed. But even in abstract, we see that the rents continue to go up. Average rents now for a seven-year lease oscillate between 120 and 130, more or less. And this is significantly above what we initially underwrote when we built our DC model down in 2020 and updated in 2022. I mean, just to give you an example, the updated Phase I average rent for us was 112. And regarding Phase II, we underwrote 118, and we are comfortably exceeding those rents at present in all of our leasing conversations. So the pricing power remains. It is true that if you want to lease 120 MW in one block to a hyperscaler, then you better be prepared for some price tension.
I mean, they will stab you to death, I mean, on the back. But precisely what is new in the market and provides you with some defense in that regard is that three years ago, the only counterparty you could have in order to lease more than 20 MW was a hyperscaler. Today, you can lease more than 20 MW to 15 MW or 20 MW different counterparties in the market. And what is really scarce is IT capacity. So you are the holder of a good which in the market is scarce, which is IT capacity, immediately deployable. And the number of procurers, the number of offerers for that capacity has grown significantly. So generally speaking, the fundamentals for offer or the demand supply in the market are good for the owner of the capacity. Anyway, I mean, the only thing which is certain in life is tax and death.
We will see how the thing evolves in the coming years.
And last one is actually on the competition situation in the data center space. I mean, maybe you could describe a bit the supply side. It seems like there's a lot of allocation of capital to the segment?
Okay. I guess, I mean, we can talk about our home market, which is Spain, Portugal in general. As my colleague Francisco was a little bit hinting some time before when he was presenting his data center rollout, the biggest problem we have at present is the amount of noise that has been created. So the market believes that everyone is doing data centers in Spain.
Then the truth is that if you take a car and go travel Spain in secondary roads and big conurbations, the periphery of big conurbations, the number of machines that you will see breaking holes on the ground in order to build a data center is exactly zero at present. In fact, there was one little data center being built in Cerdanyola del Vallès in Barcelona that was originally started by Panattoni. Panattoni flipped into AQ. AQ flipped into somebody else. So the truth is that nothing has started yet. It has already flipped like twice. And similar to that, everything. I mean, many people in this business, in order to heat up its share price or brag about the fact that they are super modern and they are working together with the AI moguls. I don't know.
But the truth is that there is a lot of noise, very, very little reality at present. So competitive landscape as we speak in Spain, if you, Thomas, come to Spain and you require 20 MW immediate, understanding by immediate before end of 2025, you will have one counterparty to talk to. And that counterparty is talking to you now. And then for 2026, I think virtually the same, maybe two, if one of our colleagues finishes what we believe they are into. Maybe three in a good day. But very, very, very little competition at present in our market. I don't know in the rest of Europe, and I am not super familiar with what is happening in the FLAPS. I really don't know. I cannot be very specific about it. But in our home market, very little competition.
And importantly, since you decide that you fancy to do some data center, why not? Till the moment you erect, open, equip, source electricity, and start operating with blinking red and green lights in your data center, if you are Superman, three years. But more normally, it will be by five years. So time to market is of the essence. And I don't really see many people even starting developments at present. I guess they will start in the coming two years. A number of new developments will start. But for the moment, there is nothing.
Thank you.
You're welcome.
Okay. Thank you. The last question comes from the line of Fernando Abril from Alantra. Fernando, the floor is yours.
Hello. Thank you very much. I have two or three questions. Please, first, with regards to from a data center tenant, besides the beauty of Extremadura, why do you think a tenant may opt to go to Extremadura or to Navalmoral de la Mata instead of Bilbao? I would like to understand the drivers behind it. Is it lower rental cost, lower or cheaper energy cost, or I don't know, the proximity to networks? Your views on this would be very helpful. Then second, you provided very good color on the logistics pipeline with rental potential and pending CapEx and so on. Do you have same figures for the office pipeline? You have several buildings. You are working on several buildings like in Lisbon or in Madrid. Yeah. Any color on that would be also helpful. Thank you.
Okay. Well, starting by the second, we don't have it handy, but we will prepare it for you. We will prepare it for you and respond. The only two to the best of my knowledge, I mean, just talking by heart, the only two buildings which we are developing without a pre-let are Liberdade 195 in Lisbon and Alfonso XI in Madrid. In the case of Liberdade, I can tell you that at present, we have already signed heads of terms for all the ground floor space. That means around close to 30% of the building space at rents that beat our existing underwriting. In the case of Alfonso XI, we have just started conversations exactly for the same, for the retail. I mean, we normally start by the retail, and then we start filling up the different floors as we progress in the refurbishment.
We don't want to be engaged in serious conversations with tenants while we progress in construction because then you incur significant risk of delays. But be reassured that we trade lots of square meters of offices on a daily basis, and I don't see any big leasing risk in those two projects. So anyway, we will make sure that you get, let's say, refresh on how the commercialization of these two projects is evolving as we progress also in construction. So in the coming two years, we will tell you how everything is going in case you are afraid. Then regarding Navalmoral versus Bilbao, the key is very simple, Fernando, and probably is something which is not easy to understand. You need to be where the power is.
It is the Basque Country cannot take for granted that producing 7% of electricity of the one they consume, so having an electricity deficit of -93%, then grid will continue providing them happily all the electricity they request. However, if you position yourself in a region where what is a fraction is the consumption compared to generation, like Extremadura, where Extremadura consumes around 15%-16% of the electricity it produces, then eventually, if you are right there where the electricity is produced, you can plug into that electricity, and that electricity will not get transported, losing 40% of the generation capacity to the Basque Country. If you add into this equation fiscal solidarity, you cannot take for granted in certain regions that the grid will be solidarity with you if you are not fiscally solidarity with the rest of the country.
You have to be where the generation is. In the Basque Country, thank God, we have our supply granted. So we are in good shape. We have a fantastic location because this is where the cable lands. So we have direct access to Marea, Grace Hopper, and now Anjana in our site. So great location from a cable standpoint, better than Extremadura because in Extremadura, if you take into account that the cable enters through Lisbon, more or less, Sines, you still have like 250 km of land cable, which is lower efficiency than the sea cable into our site. So slightly less efficient than our data center in Lisbon. But again, our data center in Lisbon, it has been very hard for us to get 250 MW of electricity that will be good for around 180 MW of IT capacity.
It will be very hard to get to 450, which is what we need to get of electricity in order to top up our capacity at 300 IT. Imagine trying to get to 1.5 GW in Lisbon. That is overly impossible. Trying to get 1.5 GW of electricity in Basque Country is impossible. Trying to get 1.5 GW of electricity in Aragon with 180 coverage generation versus consumption is impossible. You have to get where the electricity is. So the engineer might be happy or unhappy. But if you create enough critical mass, there will not be just one engineer sad and lonely. There will be 2,000, 3,000 engineers that eventually will be living in El Gordo, in Isla de Valdecañas. They will be playing golf when they are not on duty. And eventually, they will have a very happy life, very cheap life.
They will be the kings of that region, and they will be where the electricity is.
Okay. And Ismael, the strategy with Navalmoral is to be more prudent and not to start any development until there is certain pre-leasing or certain demand, at least on your books, or I don't know, more speculative?
Oh, yes. I mean, we don't have funding for a EUR 10 million development. So don't worry. I mean, the idea there is to, if you want to do the whole thing, to do it hand in hand with a partner, with a hyperscaler. If you do just one building, it's because you have a specific demand. So if one of our clients says, "Rather than being in Lisbon, I prefer to be in Navalmoral because I have better visibility for expansion," then why not?
I mean, instead of building one of our modular buildings, the A60 or the A100, instead of building them in Lisbon, we bring the plans to Cáceres and build it in Cáceres. No problem. But it will always go hand in hand with our clients.
Thank you very much.
You're welcome.
Thank you. The last question comes from the line of Celine from Barclays. Celine, the floor is yours.
Hi, Ismael. Just one question. So obviously, the 2025 FFO guidance is not what we expected. So how can you make us think about 2026 in a more positive way, especially on your earnings trajectory? Thank you.
Hi, Celine. Thank you. Thanks for your question. Look, it's very hard to predict 2026 at present. There are many moving parts that can provoke an effect in our P&L and our FFO generation capacity. The most important of it all is basically interest rates. So if for some reason, interest rates, if you really believe that interest rates are going to go down in the European Union, then eventually, we will be able to replace our existing debt or add existing debt at a less meaningful cost, in which case, clearly, this will have a beneficial effect. Second, if we are faster in leasing out data centers, either in the Basque Country or in Madrid, eventually, that will also provoke an increase in our cash flow. So there are a number of things that can weigh in the final result of 2026.
However, in order, I know you don't like realistic views, but in order to be realistic here, I would assume that 2026 will be positive compared to 2025, but relatively flattish because take into account that in our assumption, the debt cost in 2026 goes from the 2.6% at which we expect to close 2025 to around 2.9%. So those 30 extra basis points of cost in debt will significantly erode our cash flow generation capacity, let alone with the fact that we are going to be onboarding significantly more debt as a consequence of the data center development. So we have to be realistic, and this is what I can tell you at present. However, of course, we will do our best to make 2026 as visually attractive as possible for you all. But we have to be very, very prudent and realistic.
Thank you.
You're welcome.
Okay. Thank you very much, Fernando, for the questions. Thank you for bearing with us for almost two hours. If you have any further questions, don't hesitate to contact us. We'll be happy to call. Have a good weekend. Thank you very much.