Good afternoon, ladies and gentlemen. Welcome, and thank you for joining Merlin's First Quarter Trading Update Conference Call. As we always do on first and third quarter, our CEO, Ismael Clemente, will briefly go through the main highlights of the quarter, and then we'll open the line for Q&A. For those of you who want to ask questions, please press star, followed by number five. With no further delay, I pass the floor to Ismael. Thank you.
Thank you, Inés. Welcome to Merlin First Quarter Results Presentation. It's been a pretty solid quarter overall. In terms of consolidated performance, growth rents went up by 2.7%, which was okay, and particularly we improved significantly our margins, and the FFO went up by almost 17%, which is good. We have significantly shrunk the dilution caused by the capital increase, so we are running at present at - 2.6% FFO per share, which is remarkable given the new share count. In terms of MTA, despite not having revalued in the quarter, we are running at - 4.8%. Let's see what happens in June, but we'll continue probably shrinking the dilution of the capital increase in terms of MTA by 30th of June. It's been a very active quarter in terms of data centers.
We bought two sites in Madrid with 115 MW readily available, which will allow us to develop around 78 MW of IT capacity. In terms of commercialization, a block of 18 MW IT has been let in our Basque Country development in Arasur. Small but super important for us, 6 MW of our repowering capacity that will arrive during the first half of next year in Barcelona has been pre-let, which is important because pre-lets are relatively scarce, at least in the Spanish market. As commented on many occasions, we believe that the possibility of doing pre-lets was confined just for existing clients, and this is exactly what we have done. It is not so easy to do it with a new prospective client which does not know your ability to operate and deliver the exact product they need.
With the two data centers in Barcelona and the Basque Country fully including the repowering, we will become the leader in terms of IT in operation in the Iberian Peninsula that will be further strengthened by the lease up of Madrid when the electricity arrives next year. It is very, very interesting for us because at the time of the capital markets day in 2022, we laid a roadmap in front of all of you, and of course, we try to abide by what we say, and it has been, thanks God, we are delivering what we promised. This is very important for us. On the existing traditional asset base, the quarter has been very strong from an operating standpoint. Beyond the inorganic growth of WIP brought into operation in data centers and logistics, the existing portfolio has enjoyed an organic growth of 2.7% like-for-like.
The occupancy remains super high at 96.7%, which is good, and it's not easy to maintain that kind of levels. Importantly, the FFO has increased at a high double digit, 16.9% compared to the three months of 2024. As commented, almost offsetting the dilution created by the capital increase. Very little asset rotations, EUR 37.4 million of non-core sales, double-digit premium. This is a little bit of a bullshit. I mean, if you allow us to do this because it's been like 11% or 13%, this is low double-digit. I mean, it's not that we have done a 50% premium, but it's okay. We have a further EUR 15.6 million signed that will be executed in 2025, and some other things in the oven that will end up materializing during the year. As commented, no valuation during the quarter.
The MTA per share is a reflection simply of the accumulation of cash flow, so 14.47. We recommended to the board of directors and then to the general shareholders meeting and approved a final dividend of EUR 0.40, given that we have paid EUR 0.18 on account. EUR 0.22 will be paid on May 26 as a complement of the year 2024 dividend. In terms of business performance, the rents have enjoyed a very interesting period, 2.9% in offices, 1.8% in logistics, which is the only segment in which we are lagging behind a little bit, the others, and shopping centers 2.8%.
The release spread, do not be frightened by the -1.3% in offices, it corresponds to one single transaction in which we have renewed and adapted to market a contract to an existing client, and we have lost a little bit of rent in the process, but we have extended the contract that was renewable year by year. It is a contract that we inherited from a past acquisition. We have now renewed till 2032, 43,000 sq m. Plus, we are negotiating now an extension that will be built turnkey of 21,000 sq m for that same client in the same location. It is a complex transaction that will significantly increase our backlog and will further strengthen the average occupancy in the A1 Corridor that you might remember the headaches and the literature it caused in the past.
Now it seems that the problems are a little bit behind us, and the A1 corridor is performing solidly. 4.7% in logistics, which is good because it will translate in like-for-like next year unless we lose occupancy, and 3% in shopping centers which continue performing like a rocket. It's very interesting. Many of you are asking whether we want to modify our guidance for the rest of the year. We are in the first quarter. We better not do it. I mean, let's act with a little bit of prudency. The year is just starting. The world is unsafe, subject to lots of fluctuations. I mean, whatever announcement by the U.S. government can derail the economy tomorrow. We better stay where we are. Yes, I mean, it's clear that we are running on an FFO of 0.15% per quarter.
That should point to around EUR 0.60 for the year in full. I think we said EUR 0.59. We are running a little better, but anyway, it is not a big difference. I believe it's a little bit childish to be so obsessed about the guidance or the not guidance. I mean, around EUR 0.59, if we are lucky, it will be EUR 0.60. It's okay. That's basically all. One comment, which is that offices, normally the first quarter in offices in Spain, given idiosyncratic aspects of the market, is where all the renewals are concentrated. Normally you start the year losing a little bit of occupancy because there is always some churn despite having a renewal rate of around 82%. This quarter has been strong. I mean, a lot of activity, 115,000 sq m contracted. A lot of activity. We haven't lost occupancy, which is remarkable.
If you pro forma the transaction that I just commented that will be with us this quarter in the house and in the year, it will be weighing in our numbers for the next 12 months to come. It will worsen. It will make uglier our numbers for the next 12 months. If you pro forma this single transaction, the release spread has been 4.1%, which is remarkable and goes hand in hand with the idea that I have expressed sometimes to all of you that, believe it or not, rents seem to be accelerating a little bit, at least in Madrid, not so much in Barcelona, but they seem to be accelerating a little bit in Madrid as a consequence of the destruction of stock that I commented. It's just at the very beginning.
We are in the first innings of the game. If the destruction of stock continues over the next years, I believe this is going to have an effect in rents as it had in Portugal a number of years ago. I mean, this is a situation that we have seen in the past. Without further preambles, I think we move into Q&A. My colleague Fran Rivas is here with me because I'm sure there will be a number of questions regarding data centers. Inés Arellano is also here in case there are some specific questions about one number that I couldn't know by heart. That's it. Let's go and move into Q&A.
Thank you, Ismael. The first question comes from the line of Stéphanie Dossman. Stéphanie, whenever you want.
Hello, everyone. Thank you. I will have three or four questions, if I may, mainly in data centers. The first one regarding the new contract sign, how do they compare to the previous contract signed in fall 2024 in terms of level of rent, step-up close, etc.? The second one relates to the phasing of your CapEx in phase II. I noticed in your corporate presentation that you—I'm not sure how to read the change, actually. 2025 CapEx looks to be cut by more than 15% and even by 5% if we add 2026. Is there any unexpected delay in your development plan, or is it a question of lower than expected demand going forward with all the negative narratives since January on development cut from hyperscaler and so on? How should we approach this in the future? Should we expect additional cuts?
Maybe the third one, what do you expect in terms of valuation changes in your data center portfolio in 2025? What is the view of the appraisers? Is it blurred with, again, the bad noise I was talking about? Any change there? Thank you.
Okay. Okay, you're welcome, Stéphanie. Look, regarding valuation, this is an exercise that we haven't yet started for mid-year valuation. I can tell you that the discounts that the appraisers have been employing in our portfolio, discount rates were ranging between 10% and 12%, and they were reflective of the, let's say, level of uncertainty pertaining to construction risks, commercialization risks, a little bit of everything. They were using relatively high discount rates. Now, with data centers finished and fully let, it is normal to believe that those discount rates will moderate, and this will have, of course, a positive effect on the valuation of data centers. I don't know to what extent. I don't have a figure that I can give to you. We will see it in the second quarter results.
Yes, we expect a positive evolution of the value of data centers. In fact, the other asset classes, given the panorama, given the interest rate environment, etc., I believe that eventually most of the bleeding might be behind us. I mean, I would particularly like a little bit more correction in offices to go from our passing 4.9 to something in the region of 5.2, something around that. This is my particular taste. The appraisers may have a completely different view and stay at 4.9 or even shrink, even decrease a little bit the cap rate because we know they are doing that in some other peers in Europe. They are already compressing cap rates. We will not push them in that respect, but it could happen. Let's see. Shopping centers are also correctly valued.
I mean, and there are now significant transactions in the market that give you very interesting price points. Yes, the ones that most people notice is the C and D shopping centers, but I mean, there are also some negotiations on A and B class shopping centers. There is now clearly a reference. In logistics, same thing, although the logistics investment market is a little bit more muted these days because it is, I believe, digesting a little bit of overbuilding by a number of tourist developers in the main regions of Madrid and Barcelona. Regarding the phasing of the CapEx, Inés will take that question, and I will comment on the negativity of the hyperscalers because I know this is terrifying you, the negativity of the hyperscalers, but I will give you a very easy example, and you will understand why there is not such a negativity.
Yeah. Stephanie, basically, on the rhythm of CapEx, bear in mind that we're always talking about committed CapEx. This is not incurred CapEx. With the changes that we've done in phase II, basically reducing our exposure to Portugal and bringing Madrid forward, this is what we have right now as the most updated figures for committed CapEx. Nevertheless, for stabilization figures and timing, we remain the same. Do not be afraid of seeing a different amount of CapEx committed because we think we're going to be on time to meet the business plan that we provided to you.
Of course, the CapEx is alive. I mean, there will be variations during the period. I mean, sometimes we incur a little bit more in the quarter. Sometimes we incur a little bit less. I mean, there will be variations, but I mean, none of the existing construction yards is experiencing any particular delays or are we having problems regarding CapEx? Remember that one thing is the CapEx that we record in our accounts and a very different one, the CapEx that we commit, which the numbers we give to you is the CapEx committed, which includes the accounting plus the commitment. If I buy 30 generator sets from MTU and I commit a further 30, to me, I have committed 60, okay?
Although in my accounting, I will only reflect eventually 30% of the price of the first 30 and just a down payment for the jump in the queue on the remaining 30. There will be variations. I mean, if you want to be very specific, you can come here and talk to our team, but I would not try to reconcile because I believe it is wasted time. On the negativity of hyperscalers, there has been a lot of conversations regarding this. Look, for the ones who were here doing logistics in the past, I remember about two or three years ago, there was a big terror in the market because Amazon was giving back the keys of a number of contracts, and they were not pursuing some deals.
In some cases, for example, the fulfillment center they built in Badajoz and another one they built in, I think it was Huesca or León, those ones were finished, and they have never operated. They have the fulfillment center ready, but nobody has ever operated from there. They have security, etc., but they are not using it because they did it from their own balance sheet. That was specific to Amazon. That was not a problem of the market in general. We continued; we were not super high in Amazon risk, and we continued performing pretty solidly with the rest of the market, with other 3PL operators, without a significant problem. Regarding the hyperscalers, there is the following situation in the market.
The big enterprises are, in some cases, building their own models, in some cases with the help of hyperscalers because the hyperscalers have a very easy way to commercialize with big enterprises because they enter through ofimatics, they enter through the search engines, there is always a way, or they enter through cloud applications. There is always a relatively easy way to establish a relationship for the hyperscaler with the big enterprise. However, once you have trained your model and you start doing inference, the amount of computing capacity that you consume becomes very important because it is going to cost you a lot of money. In those cases, sometimes the enterprises switch the inference to a different supplier, in some cases an artificial intelligence as a service supplier, because they tend to be significantly more cost-efficient than the hyperscalers. The hyperscalers are too big.
They have internal departments that are specific in doing inference, but the cost at which they build, the cost at which they buy, the cost at which they buy the bus bars, the equipment, the racks, and everything tends to be high. They translate that inefficiency in cost sometimes to the clients. There is an effect of migration at some point in which the clients, once they really need to squeeze the model that they have been training and start doing inferencing, they move into a different supplier. Of course, that means that the market share they used to have in the global market share is always little by little being eroded by the myriad of new entrants in the artificial intelligence as a service space that we have commented on many occasions.
We know that Microsoft has been the protagonist of a number of situations, particularly in the U.S., in which they have pulled back from existing deals. They have relinquished capacity that they were entitled to, etc. It's Microsoft that we don't see an abatement in demand. I mean, we see a lot of demand in the market. We see a lot of new entrants trying to build capacity. We believe that at present, at least, there is no reason to be worried. As very well said by David Guerin of Green Street in a recent piece of research, he doesn't know what is the specific amount of new demand that Europe will require over the next five years. He knows it's going to be multiple times the existing one.
That basically means that if you are a relative pioneer and you are at the forefront of the sector evolution, you need to continue building capacity because a lot, half of a lot, twice a lot, that is going to be a very significant increase of capacity, the one that we will see in Europe. He further compared, and it is also a very important piece of information. He further compared the evolution of the U.S. in terms of capacity demand versus Europe. He concluded that Europe is lagging the U.S. by about two years. The capacity contracted in the U.S. in 2018 was very similar to the one in Europe in 2019-2021 and 2020-2022. In 2022, in the U.S., artificial intelligence arrived. The demand for capacity skyrocketed, and this movement has not been mirrored yet in Europe.
If you take the following year, 2023 and 2024, they more than doubled the previous year. Each year, more than doubled the previous year. Those two movements have never happened yet in Europe and will happen over the coming years. We are starting to see some European artificial intelligence now playing around in the market. I mean, at least we know one that is very active and has taken a lot of capacity. Little by little, Europe will be catching up with the U.S. We are not really worried about demand for the moment. If we see it otherwise, we will be the first to tell you openly. Regarding the new contract in DCs versus FAST, you can comment on it.
Yeah. Thank you. Very similar. They are basically contracts of 10-year length, mandatory, with several extensions.
Normally, this is linked to the average life of the different equipment that they are implementing and deploying in the building. This is from the length, from maturity point of view, and from the rent point of view as well, pretty in line. We are beating what we share for phase I and even for phase II at the time of the capital increase. If you remember, basically, for phase I, we were targeting when we did the math at the time, like EUR 112.5 per kilowatt month for these first three assets. For phase II, we were seeing basically an increase up to EUR 118.5 per kilowatt month. Now, if you make the calculation of what we have disclosed, we are above EUR 120. That is basically the levels are being maintained with these specific clients and with the rest we are talking to.
All right. Thank you. Maybe just a follow-up one, if I may. On logistics, should we expect a departure similar to the one of Decathlon? How is the demand behaving?
The departure of Decathlon, as you know, Decathlon is now reducing capacity, including in France. We lost them in Seville, and we're able to replace most of the space by a new contract with Airbus. That could happen with other players. I mean, what is true is that online commerce is no longer growing at double digit. It's growing at single digit. These will have an effect on logistics, no doubt. The flip side of the coin is the excellent evolution of the shopping centers. We always conceived logistics as a natural hedging to our physical commerce, to our shopping center activity. It's played that way on many occasions, including during COVID, where the excellent performance of logistics compensated the decrease in cash flow we experienced in shopping centers. In the future, it might happen the other way around.
I mean, logistics might suffer a little bit because online is clearly no longer what it was. Physically, however, it is doing fantastically well. The two activities combined, you might remember that in MERLIN, both activities are coordinated by the same professional. We have a colleague of ours called Luis Lázaro who is coordinating both activities because we see them as one single activity, particularly now that we got rid of most of our light industrial. I mean, 90% of our logistics today is 3PL related and, well, pure distribution, but it is commerce related. We sold most of our light industrial. As such, I believe that we are talking about the two sides of the same coin. We are going to have a significant exit in the second quarter that we know already, which is in Cabanillas Park B in the phase II corridor in Madrid.
DXO will be leaving. That will provoke a void. It will provoke a vacancy of around 47,000 sq m. We will continue working to replenish that shed. For the moment, the logistics market is good and active. I would not say as active as it was in past years, but it continues to be strong, as evidenced by the pace of prelims that we have been achieving in our existing developments. I would not be too worried about it for the moment. If there are news regarding that, I will disclose in future conference calls, but not for the moment.
Fair enough. Thank you so much.
It's a pleasure.
Thank you, Stéphanie. The next question comes from the line of Marios Pastou for Bernstein. Marios, the floor is yours.
Thanks very much. Good afternoon. Thank you for taking my questions. I've got three questions from my side. I'll ask them all at once. Firstly, it's a bit of a follow-up on the phase II pipeline. I think you mentioned that regardless of the CapEx changes, you are still foreseeing in line with business plan. Can I just check that that refers to the volume of rents you're expecting from 2027 and through to stabilization? So the CapEx isn't changing things or pushing this out. Secondly, on phase I, I think you mentioned previously that a lease would be more likely when the power comes online next year, but maybe an update on Madrid in phase I would be helpful. And then just finally, on guidance, can I just confirm the guidance of $0.54 before making adjustments for capitalized interest still holds true? Thank you.
Okay. Regarding the phase II, in principle, everything is on track. We are forecasting rents of around EUR 320 million for the whole of phase II. Costs are more or less kept at bay. I mean, there are some things which are going up, some things which are moderating a little bit. We are not extremely worried about cost. In fact, in the latest update of our model, we are just like 10 basis points above in terms of gross yield on cost, which is good. EUR 326 million with total IT capacity installed of 210.
Business plan remains pretty much in line with the only significant amendment, which has been the reduction of capacity in Lisbon and the increase of capacity in Madrid, which has been caused by an unexpected event, which is that mid-February, the former Biden administration put in place a U.S. artificial intelligence diffusion rule, and they classified the countries in the world in three categories: tier one, tier two, tier three. The close allies, including most European Union countries, were classified in tier one. They are entitled to import to their territories the latest gear they want from the U.S. For reasons unknown to us, Portugal was placed on tier two together with Poland, for example. That is not the end of the world. It simply limits a little bit the number of GPUs that you can import to the country.
A given operator can import around 50,000 GPUs, which is a lot, but 50,000 GPUs. The base for the calculation is approximately the H100, the Hopper 100 of NVIDIA. Of course, the more sophisticated the GPU becomes, the more reduced the number becomes. If instead of Hopper 100 it is Blackwell 200, the number of GPUs is lower. For future series of NVIDIA like Vera Rubin at the end of 2026, the number will keep reducing because what they are trying to do is limit computing capacity that can fall in undesirable hands. The consensus in the market was that that was probably a mistake that the Trump administration would correct. Probably the Trump administration has had other priorities, and in reality, nobody has really paid attention to this tier one, two, three categorization of the world.
The latest news that we got yesterday, I mean, from an American client in Barcelona, is that the latest they know is that the Trump administration is thinking about scrapping the whole U.S. Diffusion Act. If that is the case, we will rethink Lisbon. I do not know whether we will go from 36 to 108, but maybe we go from 36 to 72 and increase a little bit our phase II capacity just in case in order to make sure that we have more probability of doing our full CapEx deployment and bringing rents to the company. This is basically the only thing that has really changed. We have reduced 72 in Lisbon, but have increased 78 in Madrid. We were looking at two pieces of land in Madrid with immediate availability of power. We have closed on both.
One is subject to demolition and cleaning of the site. It was a former steel mill. The other is subject to urbanization. I mean, basically bringing the utilities. I mean, we're doing the urbanization works. Nothing is really serious. Once we get the hold, once we get the delivery of those two pieces of land, we will start construction. I don't know whether end of this year, but beginning of next should be a good bet. The idea is to add 78 MW of IT capacity in those two plots to replace like for like the 72 MW, let's say, lost in Lisbon. On phase I, I believe I don't remember the exact question. I probably.
Exacerbating in Madrid.
Exacerbating.
Leasing in Madrid.
The leasing in Madrid. Yeah. The leasing in Madrid, the problem that we have is that we only have a commercializable block of around 5 MW because the utility has given us only 8 MW of electricity. As such, that block of 5 MW is a little bit insufficient for IT, so for IA. Very probably, we will commercialize it together with the remaining 14 MW that we will receive next year upon delivery of the electricity by the utility company through two aerial lines that we are bringing, that we are building and bringing to the plot. Once we equip the building, which is something that we should finish by end of the year. I mean, we are receiving the equipment as we speak. We will be fitting out the equipment. It will be ready as of year-end.
We will finish year-end with 42 MW equipped and ready for use, but 58 equipped, not ready for use, or part of it, 16 of those not ready for use. This will be precisely Getafe. Regarding commercialization, once we receive the electricity, I would not be too worried. I mean, Getafe, Madrid has a lot of demand. I mean, we are negotiating with multiple parties, and I would not be too worried about it. I believe during 2026, God willing, we should be able to have it fully lit and leave the phase I completely delivered and full and cash flowing, which at the end is our objective in order to have full rents during 2027 as committed vis-à-vis the market, vis-à-vis all of you. This is what we want to do with Getafe.
Regarding guidance, the guidance that we gave was like EUR 0.54, maybe EUR 0.55, EUR 0.59 pro forma of the capitalization of interest, which is something that, as commented, we do not want to do. I mean, we will give you the raw number, and then the number with capitalized interest will be simply a pro forma that we will give for informative purposes. The pace at which we are running indicates that we are going to exceed the guidance, but it is yet to be seen what will be the excess. We do not have yet visibility. I mean, in the second quarter, after the departure of GXO, etc., we will see what is the cash flow we obtain. If we see fit reguiding sort of mid-year, we will do it, but not for the moment because it is too early to do it.
We believe we are going to beat our guidance, but we shouldn't be too carried away because the year is very long and many things can still happen in the coming months.
Do you have more questions, Marios?
Okay. That's very helpful. Yeah, just maybe just a follow-up maybe on Portugal and just maybe a bit more information as it feels quite significant and we're getting questions on this. I just wanted to check why this maybe wasn't a separate release and, yeah, why we're really just hearing about this now in terms of the changes being made to the phase II pipeline.
Sorry, what is the? I couldn't.
Change? Why the changes in the pipeline?
Why the change in the pipeline? Because of the U.S. Artificial Intelligence Diffusion Act. I mean, the reason why we changed the pipeline is because we shrink a little bit. We reduced the capacity with which we are going to go to market in Portugal. We reduced it to 36 because our clients there in Portugal will need to go through an extra process in the U.S., which is the obtaining of a validated end user certificate. The loophole to the United States Artificial Intelligence Diffusion Act is that if you are a validated end user, and you are operating within a data center which is approved by the BIS of the U.S., you can import the latest technology with special permission from the U.S. government.
Our intuition is that that further requirement is going to, let's say, make slower the process of decision-making of our clients and also might funnel part of the demand to Spain because between asking permission and not asking permission, people is like be water, my friend. I mean, they will go through the easiest route. This is why we have reduced a little bit in Portugal. It has other implications. It is not going to be super economical for us because we are building the generator building, we are building the transformers building, we are building the admin building, and we are going to do just one data hall. Yes, not eighteen as we had initially defined. It is going to be 36, but it is going to be just one data hall.
Of course, if you do three data holes, 108, you will have a significantly bigger capacity of dilution of all the common infrastructure of the park. Life is long, and we will continue leasing in Portugal, quicker or slower, but we will continue leasing there and do the second building and the third building and the fourth building and the fifth building. There is also another thing that is not really helping, which is that Portugal is in the middle of an election process. We will only know the new government of Portugal by the elections are now in May, I believe, and we will only know the new government depending on the agreements that the different forces need to do by June, July, might be September.
If that is the case, clearly this is holding back a number of decisions, including one which is important for us, which is the CapEx that the government of the state of Portugal needs to do in a gas metering station that is right next to our plot from which we are getting the gas that fuels our gas generators. Given the fact that this being a riverside location, we haven't used fuel in this occasion. We are using natural gas for the backup generators. This is why we changed Portugal for Spain. It's not the end of the world. I mean, if finally the U.S. Diffusion Act is scrapped, as commented yesterday by the client of ours, then eventually we will increase a little bit the size of the initial bed in Portugal and recover part of the capacity that we decided to postpone.
Very clear. Thank you very much for the answers. Appreciate it.
Yeah. You're welcome.
Thank you, Marios. The next question comes from the line of Fernando Abril from Alantra. Fernando, the floor is yours.
Hello, Ismael and team. Thank you very much for taking my questions. I have three. First, regarding the CoreWeave partnership, they have now pre-let the Barcelona repowering almost a year in advance, as you said. CoreWeave is a big player and actively seeking to expand capacity in Europe. Do you see a real possibility that CoreWeave could act as an anchor tenant in future larger developments such as the Bilbao Extension or Extremadura? Then a couple of follow-ups. First, on data center rents. You mentioned EUR 66 million in passing rents for the signed capacity. Relative to your EUR 88 million target for phase I, this implies an average of around EUR 100,000 per megawatt month. For the remaining 19 MW in Madrid, which, by the way, I think normally has higher rents, I do not know if this seems quite conservative.
Is this simply a matter of prudence, or do you see real upside to your rental assumptions for phase I and maybe to your phase II assumptions as well? Third, regarding the Madrid assets brought forward to phase II, I know you've mentioned it, but just to be more clear, what is the expected timeline, specifically when do you anticipate construction permits, power sourcing, and the equipment? Just to get an idea of what buffers you have in place to ensure these assets are in operation by year-end 2028. Thank you.
Thanks a lot, Fernando. The one on the Madrid side will be taken by Fran. Regarding the relationship with CoreWeave, yes, it's very important for us that they have committed to a prelet. I must say that it's a relatively natural prelet because they have a significant capacity already in that same data center. They have 20 people, software engineers working in there and our engineers. It was relatively natural that they would take the expansion. It might have not happened, but it does not mean that they are going to continue doing prelets across the board. Yes, I think it's a very positive development. We are looking at other things with them, but whether they can anchor one of the giga developments or not will depend a lot on the evolution of demand in Europe.
If the demand in Europe goes half similarly to what it has gone in the U.S., yes, there will be space for giga developments, and yes, there will be space for doing something together and using the relationship as an anchor to one of our existing big developments in Extremadura or elsewhere. For the moment, do not assume that we are going to be doing everything with CoreWeave because it is a two-way relationship. First, we need to be mindful of a certain dispersion or diversification of rents on our side. On their side, they are also mindful of their own diversification, and they also need to be matching constantly the long-term commitments they adopt as a consequence of leases with the demand they are finding on the market. We will continue, of course. I mean, they are happy with the way things have gone.
I mean, the remote hands agreement has worked pretty nicely for them. They have been really positively surprised about the capacity of our technicians and the way we have equipped on their behalf. Things are in very good terms with them, but we need to see how the relationship develops. Regarding the DC rent, you spotted it right. I mean, if the first 44 is 66 or the first 45.2 is 66, that means we have been letting at an average of slightly above EUR 120. As you might remember, for phase I, our magic number was EUR 112. So we are beating our expectations in terms of rent in phase I. For the reminder, once we get to 64, yeah, it is relatively easy to extrapolate the fact that the 88 is probably short of what the reality should be. Normally, rents should be above EUR 90.
Again, I mean, we will see. I mean, we wait till we fill up Madrid. If for some reason there could be many things playing at the same time, imagine in Madrid, the type of client is a cloud player or a hyperscaler, then things are different. Rents are different because they could be doing things which are not related to artificial intelligence. They could be doing cloud, and eventually, they will not be capable of paying such a high rent. Let's see. Let's see how it goes. Of course, our efforts are concentrated in getting the maximum rent possible. If we are successful, yes, the rent will probably exceed EUR 90 million or even something in the range of EUR 92 million, something like that. Madrid construction and equipment. Equipment, but actually construction.
Real Madrid, on the new two plots, we have two different situations there.
The first one basically is the one which we call the second building in Getafe. On that one, basically, the land we acquire is urban land. There is already construction. It is an industrial facility already active there that will be demolished in the next months. In terms of the project that we need to approve in order to start construction, we are dealing with the same sample and the same area that we already built an asset there. We are foreseeing basically a more smoothly approval process, considering basically that we are almost doubling or more than doubling basic capacity, but in terms of how it does work and the structure of the building, etc., it is pretty, pretty similar to what we have right now.
In terms of timing, we are expecting that this demolition will be ending by the end of this year, beginning of the next. We should be right after starting construction. Best timing we have right now is first quarter 2026 to start construction if asset basically demolition and permitting are going in due time. Interesting thing of this plot asset is that because it was active, the power is supplied, which means that we are paying the power availability, what we call in Spain terminal de potencia. We are paying this on a monthly basis already. Once construction is finished, the power is already there waiting for us. Second plot, which is Tres Cantos, north of Madrid, which is basically area. That one basically is a former industrial facility as well, which was basically active there until several years ago.
One part of the land is already urbanized. The second one is pending urbanization, which can adapt to the type of assets we need to deploy there. The seller is doing the urbanization for us. We will buy, let's say, once those CPs are clean, we are buying final urban land ready to build. Power there is granted. It is not sourced, of course, because there is no building in operation right now. We will delay a little bit the time as compared to the one in Getafe. We expect that they will start urbanization in the second half of this year, probably beginning of the next as well, first half. Again, we will try to do some sort of simultaneously approval for the construction project so we can start construction as soon as the urbanization is completed.
We are not doing this in different time in an overlap, but overlapping the different approval processes to accelerate the deployment in these two plots of land.
Okay. Fernando?
Okay. Yeah. Thank you very much.
Thank you. The next question comes from the line of Alex Kolsteren from Alantra. Alex, the floor is yours.
Yes. Good afternoon. Thank you for taking my questions. Two non-data center related ones. The first one is on the logistics relating the Decathlon to Airbus. Would you comment on the reversion capture there on the releasing activity? And secondly, on the Madrid office, lease renewal with a negative reversion captured. Do you expect all the big leases to mature and capturing similar negative relating spreads? Thank you.
Okay. The logistics one, you will have to repeat it, but I will start with the Madrid lease. This lease, the lease that we have now renewed up to year 2032, we inherited it from an office park that we bought from a venture capital fund. The lease was a little bit, let's say, weird. It was a little bit on purpose. It was a little bit above market. Of course, we took it into account in the pricing of the transaction, but it was clearly above market. It had been injected some steroids and was not reflective of the reality in the area. What we have done now is simply what we have enjoyed. The lease, it lasted. We renewed, but we renewed on a year-by-year basis. Now we have renewed seriously. We have renewed in year 2032, and we have adapted to market.
It is not reflective of a market situation. I mean, if your question, I believe your question means, I mean, you have many other headquarters. Are all of them overrented? No. In fact, as commented on some other occasions, if I have to bet on the direction that the market is taking in Madrid, it's probably a different one. It's upwards. We are now rebuilding the cushion between passing rents and market. We are rebuilding reversionary potential because the market rents are evolving now quicker than inflation. Anyway, in the past years, the office team did a good job in extending most of the important leases which pertain to headquarters of big multinationals. I mean, we extended Endesa till year 2030. We extended Indra till year 2032. We extended Price till year 2033.
I mean, we have the big headquarters over the past years have been significantly extended till beyond 2030 in most cases. Do not be afraid. I mean, do not extrapolate that particular case with the rest of the portfolio because that will be a false read across. What we are trying to do here is simply bring on board more backlog. We want to continue building on the strength of the A1 corridor, and we have buildability which is unused in the Adequa business park. If we can employ that unused buildability and build a turnkey building prelet to an existing tenant, which is enlarging significantly its presence and bringing everything they have in satellite locations to just one single headquarter location, it is a good opportunity for us because it is relatively easy to manage, and it will give us more cash flow and more backlog in the A1 corridor.
We might also take the opportunity to finish all the remaining unused buildability in that park because it is not very significant, and we might perfectly afford the little luxury of building part of it spec and finish it because we are trying to make sure that with the effort of CapEx that we have in front of us for the coming years, given the DC development, we want all the cylinders of our engine to be firing. I mean, if we have a 20V, we want to have a 20V with 20 cylinders firing. We do not want to have a 20V with 18 cylinders firing and two idle because that is not efficient from a cash flow generation standpoint. Lately, I mean, following the divestiture of the BBVA sale and Lisbac, we do not have a problem of LTV.
Let's say our only problem, between quotes, is recovering as quickly as possible from the dilution created by the capital increase through organic growth so that we wait comfortably on a very significant dividend while we wait for the new cash flow stemming out of the DC development, particularly of phase II, which is the one which is meaningful because phase I is relatively humble. It's EUR 90 million. This is the intuition or this is the idea behind the contract you commented. In logistics, I couldn't hear your question very well. Can you please repeat it?
Yeah. Sure. No worries. I was just wondering about the releasing spreads on the letting activity from Decathlon to Airbus.
In the Decathlon to Airbus transaction, we have lost rent, but it's not really spread because it's not exactly the same perimeter. Yes, we have lost rent because Airbus is an industrial client and couldn't afford the same rent we had with Decathlon, which was an online commerce type of rent, a little bit more elevated. This is a deal. We have prioritized, of course, the backlog, again, the obtaining cash flow, and we have decided to relate as quickly as possible rather than wait for another client in the e-commerce space.
Okay. Perfect. Thanks for the explanation.
Thank you, Alex. The next question and final question comes from the line of Ana Escalante from Morgan Stanley. Ana, the floor is yours.
Hello. Good afternoon. I have two questions. The first one is regarding your dividend because correct me if I'm wrong, but I could assume that your dividend policy is still based on your AFFO pre-any adjustments for capitalized interest. However, given you are delaying a bit the CapEx for phase II, could that open the door for paying a bit more dividend or increasing the dividend during this year or next year, even if AFFO, as the reported figure, does not grow much? The second question is regarding some press articles on some potential interest from a sovereign wealth fund in Castellana Norte, in Madrid, Nuevo Norte. In case BBVA would consider selling some of this stake or the totality of the stake, how would you look at that? Would you be interested in getting a bit of that?
Or maybe now that you are focusing on data centers, that could also be an opportunity to cash out from that project and maybe redeploy that capital into data centers rather than using other sources of financing?
Okay. First, regarding the dividend, our policy remains to be 80% of adjusted FFO. And this adjusted FFO, cash calculated, so real money at the bank, is not based on the pro forma in case we were to capitalize our interest. We will remain like that. I mean, the fact that this year or next, we could be missing EUR 0.02 on the dividend payout is not that relevant because at the end, we will not hamper our capacity to pay dividends in the future. The only thing that it does is flattens a little bit the predictability of the dividend, but it does not follow the cash principle. As such, you could be paying more cash than the one you have available. We will continue with the same policy. Hopefully, particularly with the buildup of our data center activity, we will enjoy more cash flow available.
With more cash flow available, we will pay an increased dividend. It is better not to play or mix accounting, I would say, options or tricks with real cash. I mean, it is better to pay the cash you have or 80% of the cash you have. Regarding Castellana Norte, sometimes we read on the headlines of the Spanish press news about it. We have not seen any movements in reality regarding BBVA. I am not sure. I mean, maybe they keep exploring the market. I know they had an investment bank hired some time ago, and they have been sounding a little bit the market, but we do not see them very active in that regard. I mean, now they are focused on agreeing among us what should be the next step, what we should be doing, how we should be developing the different areas, etc.
Of course, we cannot discard that they maintain a parallel negotiation and they sell it or whatever. If they sell, we will stay cool. I mean, basically happy with a new partner, and whether it is a partial partner or a total partner that replaces BBVA in full or replaces BBVA in part. We will continue working with all of them peacefully and happily. Regarding using the opportunity to cash out, no, it is not our intention because we like what we see. I mean, we are real estate. We are not FX Partners. So we like the quality of the offices that will eventually be constructed in that area of Madrid. We do not see as many risks as non-professional real estate people see in there because I know building is always a big tally for many people. "Oh, no, you have building risk.
Oh, no, you have commercialization risk. This is exactly our life. This is what we do for a living. So we are not really afraid of it, and we like the location. We like the, particularly we like the infrastructure, which is second to none, not only in Europe, in the world, because contrary to what many people believe, this is not La Défense. This is right in the middle of Madrid. It's not Canary Wharf. This is right in the middle of Madrid. You have a transportation hub in which you connect aerial train with metro, with subway, with green buses, with blue buses, with the airport, with high-speed train, and with taxi. Very few places in the world you can do that. It's very easy to get there and out. We believe it's going to be a success if properly executed.
We wish we would have a little bit more protagonism in the execution, but our participation is what it is. I mean, we do not have the intention of buying extra participation, at least with the money that we have earmarked for data center development. We might rotate one or two secondary office buildings and employ the money to buy a slightly higher percentage if BBVA is amenable to sell to us. We are not going to do any big movement, and you can rest assured we are not going to be recycling the money obtained in our capital increase to do data centers into this transaction. Okay? In that respect, you can be absolutely reaffirmed that this is not our intention. We like it.
It's going to reshape Madrid over the coming 20, 30 years, and we want to be at the driving seat in this redevelopment because I believe there is nothing of that quality, not only in Spain, but also in Europe at present.
Ana, just to clear up something on the dividend, be mindful of the fact that we pay our dividend based on AFFO, and the CapEx that is deducted from FFO is the maintenance CapEx. Therefore, for 2025 and 2026, because only phase I cash flow is coming, not phase II, remember the first rents from phase II of data centers are coming in 2027, the fact that there is a little bit of a shift in the CapEx deployment does not impact at all 2025 and 2026 figures. The dividend, as we said on a year-end results, the dividend that we got in 2024, which was EUR 0.40, what we said with the guidance is that it is likely to remain very, very similar in 2025 and 2026.
Thereafter, in 2027, 2028, and 2029, 2027 is when we reach stabilization for phase I, and we start receiving some rents for phase II, and then 2029, stabilization and everything. We have never provided with specifics on cash flow, so we will make sure that we get as much cash flow as we can as soon as possible, where we have not provided with any sort of guidance. Let's say for short term, so 2025 and 2026, any deferral in CapEx does not impact our dividend policy.
Very clear. Thank you very much.
Welcome, Anna.
Thank you, Ana. So there are no more questions. It has been slightly more than an hour. As always, we thank you for joining today's call, and we remain at your disposal for any further questions that you may have. Have a nice evening. Thank you very much.