Good afternoon, ladies and gentlemen. Thank you for joining MERLIN 6M26 results presentation. You can find all the materials that will be presented in today's call on our website. I will please ask you to abide by the disclaimer contained in it. Our CEO is Ismael Clemente, along with our two Directors, Inés Arellano and Francisco Rivas, who walk you through the main highlights of the first six months of 2026. We'll then 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. Thank you for attending MERLIN Properties' H1 2026 results. I will be following the presentation that we have prepared for the occasion. Regarding the main highlights in operating performance, financial performance, and value creation for the period, I would like to remark that we continued enjoying strong operation in all of our traditional asset classes with a 3.3% like-for-like rental growth. We continue to enjoy a very high overall occupancy with a 94.7%, pending the incorporation of data centers as of 31st December 2026. In the three traditional asset classes, the one that showed the strongest performance was, once again, shopping centers with a 6.4% like-for-like growth.
In logistics, we had a relatively good release spread, but we lost occupancy, and in offices, we had a good occupancy performance, but we had a relatively low release spread owing to a string of in and outs in a number of peripheral buildings. In terms of financial performance, we enjoyed double-digit revenue growth, +11.7%, which translated into a very solid bottom line performance in FFO, +8%, despite the unavoidable increase in financial expenses. That means basically that on a per-share basis, we are very close to recover the dilution cost by the capital increase, which is, I think, a remarkable achievement.
In terms of value creation, we increased our GAV by 3.7%, with most of that growth coming from data centers, actually, in traditional asset classes, there was a very slight yield expansion, which probably will follow in coming months and years owing to the interest rate environment that we are going through. We strengthened our balance sheet with EUR 768 million capital increase, as you all know. We reduced the loan-to-value to 24.5%, with a liquidity of EUR 2.6 billion, that, well, basically takes care of the most immediate maturities together with the bank syndicate refinancing that we are preparing for the H2 of the year. We maintained our investment grade rating, both with S&P and Moody's.
Most importantly, because that brings a little bit more color to the table in terms of value creation, the most salient feature of the quarter or half of the year was the good execution in MEGA Plan. In the past quarter, we were at 112 MW commercialized. We have now reached 160 because we converted the Arasur 1 situation that in the Q1 was in advanced negotiations. We had signed a head of terms, and we had attached technical and financial documentation. It is now signed, as a consequence, we are 160 MW let. We provided to you an indication on the full- year results presentation of commercializing this year in the region of 200 MW. We believe that we are in a position to far exceed that mark because we have three different avenues that we can explore.
The most immediate, I believe, is the conversion into a full format lease of the head of terms and exchange of documentation taking place in Lisbon, where we had significant demand. We had four different tracks open, two of them very well advanced. We believe that with one of the two, we are going to finish soon documentation and therefore end up having a full format lease that we can report. Beyond that, we have 48 MW in Getafe, which are booked, that eventually we can also work between now and end of the year to convert. We are recently working on a combined pack of 30 and 20 in Tres Cantos and Navalmoral that eventually could result also in a significant lease. Although I believe that will probably extend more into 2027 because we are just starting to entertain those conversations.
As commented, Bilbao Arasur 1 and Bilbao Arasur 2 fully let, which is very important because for many years, all of you have been asking whether we were able to sign true pre-lets, and now we are clearly in that situation. My colleague, Fran, will comment later on that. We are reaching a situation now in which, in reality, we are going after commercialization. We are finishing a product behind our commercialization pace, which is a very nice place to be because we are enlarging and strengthening our dominance in the Iberian Peninsula while we gain a lot of visibility on future cash flow through very advanced pre-leases.
If the Lisbon lease is to be converted, the interesting summary is that 100% of phase I will be let, 70% of phase II, and yet 25% of phase III, which I believe will be a remarkable achievement, particularly for those of you who attended our Capital Markets Day in Arasur in March, because clearly, this is exceeding the projections that we internally had and that we conveyed to you on that occasion. All that while maintaining a disciplined capital recycling. We have sold EUR 75 million as of July, and we still have EUR 90 million of divestments signed that we will be converting between end of this year and beginning of next. Thus reinforcing our internal capital recycling and helping the funding of particularly phase III as we speak. The NTA stands at EUR 15.99 per share, which is very interesting.
That puts our shares at a 4%-5% NAV discount, which eventually, I hope, will be overcome during the week, because, for some reason, today, our trading has been weak. Frankly speaking, I don't know why. This is important because that brings to the table a very significant value creation, which has been always our obsession since we went public in 2014. We have distributed close to EUR 2.5 billion in dividends. Despite that, we have also sent our share beyond 100% above the initial floating price. I think we are complying with our social mission in terms of value creation vis-à-vis our shareholders. In terms of key financial and operating metrics, the gross rental income stood at EUR 292 million, + 10% like-for-like. The total income has been EUR 307 million.
This year, for the first time in our history, we are likely to exceed the EUR 600 million mark in terms of total income, the top line of the company. FFO-wise, we converted EUR 180 million + 8% year-on-year. We cannot simply multiply this by two because there are a number of things that are different in first and H2. That gives us confidence to send or to rephrase our guidance in terms of total cash flow for the year from the previous EUR 327 million to EUR 340 million. That would mean around EUR 0.55 per share above the EUR 0.53 that we gave you in February. Yes, the FFO per share is -1.8% year-on-year.
Again, that is because we calculate the per share metrics based on the total shares outstanding of 620 million, which we believe is the correct way to do it, because if we were to pay a dividend today, it would need to be calculated on the basis of that number of shares. If we were to use the weighted metric, as many of our rivals do, the increase would have been 2.6% and the per share metric EUR 0.30. Our loan-to-value continues to be very low, 24.5%, and the increasing GAAP like-for-like 3.7%, that basically together with the operating FFO, brings our total shareholder return year-on-year to +9.1%, which is, I believe, a very interesting mark. That's it basically for the key financial and operating remarks. I will pass the floor to my colleague, Inés Arellano, that will discuss the traditional asset classes. Fran will talk about data centers.
Thank you, Ismael. Moving to offices, a portfolio worth EUR 6.7 billion that generates 4.8% gross passing yield and 4.1% net initial yield. It represents 53% of our portfolio. The momentum is quite positive, demonstrated by all-time high occupancy level in Madrid. Barcelona is still suffering, reaching our lowest occupancy level at 84.4% due to the exit of a big tenant in Glories that you are all aware of, while Lisbon continues to be a very solid market. Demand is very healthy, although clearly concentrated on the very best buildings, which is exactly where MERLIN is invested. Alfonso perfectly illustrates this opportunity. Finding more than 10,000 sq m of refurbished office space inside Madrid M-30 has become almost impossible. New supply is extremely limited, and there are quite a large number of occupiers looking for flagship headquarters. That explains why leasing is progressing so well.
70% of the space is under advanced negotiations with a leading financial institution well ahead of completion, while the remaining 30% is already leased to Loom. For us, this is value creation in its purest form, transforming an existing asset into one of the most desirable office buildings in Madrid, while substantially reducing leasing risk before delivery. Libertad 201 follows exactly the same logic, in Lisbon. Avenida da Liberdade has become one of the most prestigious business locations in Southern Europe, attracting financial institutions, technology companies, and multinational occupiers. Supply is extremely limited, which explains why the retail component is already fully pre-let to a leading luxury operator, while around half of the office space is already pre-let or under head of terms. By delivery, we expect this to become one of the benchmark office buildings in the Lisbon market.
Adequa represents another very different, but equally very attractive opportunity. As you all know, the northern corridor of Madrid will increasingly become one of the city's main business hubs, as Madrid Nuevo Norte develops. It is a turnkey project for Tecnalia, who's the main occupier of the current business park. That validates both the location and the quality of the product, while allowing us to achieve an expected yield on cost above 12%. We'll continue to apply the same discipline to future phases, including Adequa 7, where we'll only proceed if returns remain attractive. Moving into Plaza Pablo Ruiz Picasso 11, this project is slightly different because it's not only about the building itself, it's about participating in complete transformation of Azca through the Renazca project, creating much greener, more open, and more attractive financial district.
Today's occupiers increasingly value the experience around the office as much as the office itself, projects like this one help to reinforce Azca's position as Madrid's premier CBD. For us, that translates into stronger long-term rental growth and better asset quality. Overall, the office portfolio continues to deliver exactly what we'd expect. Resilient operating performance today, combined with a development pipeline that's already attracting significant tenant interest well ahead of completion. Moving to logistics, which is EUR 1.4 billion, a portfolio generating 5.7% gross passing rents and 4.9% net initial yield. I would describe the market as being normalizing rather than slowing. Occupancy stands at 95%. Release spread, as commented by Ismael, reached almost 4%, and we signed around 39,000 sq m during the semester. Rental growth remains positive, while valuations also continue to edge upwards. What's particularly encouraging is that leasing activity remains broad-based.
Barcelona delivered particularly strong rental growth, while our diversified portfolio across the main Spanish logistic corridors continue to provide resilience. The appendix also shows a healthy mix of tenants and positive release spreads across virtually every geography. Performance in Southport is not an exception. It continues to be very robust, with 97.3% occupancy, with more than 200,000 sq m contracted. Overall, logistics has become a mature income generating business for MERLIN, with consistent cash flow today and an attractive development pipeline for tomorrow. Looking at our commitment pipeline, this is where future growth becomes very visible. We already have 275,000 sq m under development, representing only EUR 96 million remaining investment. That will generate approximately EUR 17 million of stabilized gross rent. Importantly, these are not speculative ambitions overall.
These projects have already been committed because they have sufficiently commercial visibility and at attractive expected returns, around 7% yield on cost. Delivery will take place over the next 18 months, meaning that the pipeline should progressively contribute to rental income from late 2026 onwards. The next stage of growth is within our land bank. This represents another 184,000 sq m of future development capacity, require around EUR 110 million of investment, and capable of generating over EUR 11 million of stabilized rent. The important point here is flexibility. Unlike the committed pipeline, these projects can be phased depending on market demand. We don't need to build this because we own the land. We build because occupiers' demand justifies the investment. This discipline, again, has always been one of MERLIN's competitive advantage, and it becomes even more valuable in today's environment.
Logistics continues to provide exactly what we want from this asset class, stable operating performance today, together with a very attractive embedded growth pipeline. The true growth is going to come from data centers, so I'll leave to my colleague, Fran, to explain a little bit more about data centers. Oh, sorry, just shopping centers. I heard it was going to be after data centers. I think the shopping centers business continues to surprise many investors. It's a EUR 2.2 billion portfolio generating 6.6% passing yield and 5.9% net initial yield. For several years, there's been a perception that physical retail would struggle structurally because of e-commerce. What we've seen actually is something quite different. Best shopping centers have become destinations rather than simply places to shop. Retailers increasingly concentrate their investment in dominant assets with high occupancy, high footfall, and while weaker schemes continue to lose relevance.
Our portfolio is firmly positioned in the first category. During the H1, tenant sales increased by 8.4%, comfortably ahead of inflation, while footfall also grew by almost 2%. That combination translated into 6.4% like-for-like rental growth, the strongest performance of all traditional asset classes, as Ismael said, the strongest asset class of the traditional portfolio. At the same time, occupancy remains exceptionally high, 96.9%, release spreads exceeded 5.5%, and perhaps most importantly, occupancy costs remain extremely affordable at only 10.8%. That last figure is critical. It means retailers continue to enjoy healthy profitability within our centers, giving us confidence that rental growth remains sustainable rather than being driven by excessive rent pressure. It also gives us comfort to carry on our yield management strategy. Overall, shopping centers remain one of the strongest contributors to MERLIN's recurring earnings. Yes, I do pass the floor to Fran. Thank you.
Many thanks, Inés. Good afternoon, everyone. I am pleased to provide an update on our data center business and the key achievements delivered during the H1 of 2026 across all the three phases of our platform. First of all, I would like to congratulate our data center team once again for an outstanding H1 of the year. Thanks to their excellent execution and the very strong level of activity across the platform, we have been able to secure several significant contracts across our portfolio, which I will cover in more detail in a moment. As in previous presentations, let me begin by summarizing the current position of our data center portfolio across the Iberian Peninsula, as shown on the map on page 19 and in the table on page 20.
Precisely on page 20, we provide additional details of our 724 MW portfolio, where we include both operational and development assets, and show the progress achieved across each of the three phases. The perimeter of the three phases remains unchanged, as you can see, with the only adjustment being a reduction of 4 MW in Zaragoza-WIND 1 from the previous 150 MW to 144 MW, in exchange of significantly advancing our ready for service dates that we will see in a moment. Anyway, we expect as well to recover that capacity through the planned repowering of Madrid Getafe 01 in the following months. Moving to page 21, let me review each phase individually. Phase I, 64 MW, are comprising three assets, Madrid Getafe 1, Barcelona PLZF 1, and Bilbao Arasur 03, all of them now fully fitted out and fully let.
Barcelona PLZF 1 and Bilbao Arasur 03 are fully operational and generating cash flow, including in Barcelona repowering project that has been delivered to the client in this month of July. In Madrid Getafe 1, which is also fully let, is expected to generate full cash flow once final power connection works are completed during the Q4 of 2026, where we have right now great visibility. At the same time, we continue to advance discussions regarding a potential repowering opportunity that will add approximately 6 MW of IT capacity in this building. In total, phase I GRI is estimated at EUR 68 million for 2026. From a valuation perspective, I am now on page 22, the successful commercialization of these three assets, particularly in Madrid and following the Barcelona repowering, has reinforced the significant value creation achieved to date.
Rental levels have exceeded, as Ismael was commenting before, our initial projections, resulting those in higher expected valuations based on the latest independent appraisals. From an accounting perspective, we have also recognized the promote accrued to date, reflecting the value created during this period in EUR 101 million. For those who you have previously asked about the promote mechanism, as we have explained before, it is linked to the profitability of each phase over a 10-year investment period. In this case, in phase I, from 2021 to a potential exit in 2031. However, the accrual is calculated in each quarter, assuming what will be the valuation at the moment in time, in this case, June 2026, and consider this at the hypothetical exit date.
As a result, the amount will continue to evolve until the final liquidation in 2031, depending, of course, the impact that based on the IRR calculation, moving towards the final 2031 implies on the number. It also worth noting that part of this promote is paid in advance in year five and seven, as it has been the case in March 2026, where we have resulted in a payment close to EUR 20 million. Moving now into phase II, which comprises 254 MW of IT, the construction across our work in progress portfolio continue to progress according to plan. More importantly, leasing activity significantly ahead of our original expectations, reflecting the transition from the mainly speculative development approach that we have in phase I to a predominantly pre-let near term development model in phase II and subsequently in phase III.
Looking at each project individually, in Bilbao Arasur 02, it was fully let at the beginning of the year with its full 48 MW IT contracted, implying approximately 12 months ahead of the delivery date, scheduled for this December 2026. Out of those, the first 20 MW are expected to begin generating cash flow in January 2027, with the remaining 28 MW following in June 2027. Thus, those different dates refer to the client deploying their fit out. In Bilbao Arasur 01, which has been also fully let during this Q2 , we have secured its entire 48 MW. Again, approximately 18 months before its expected delivery in December 2027. Construction at Lisbon building one and building two, representing 80 MW of IT capacity, is progressing very well.
At the same time, work is advancing on the campus substations, generator buildings, and an administrative building, all of which are being developed simultaneously, saving time for phase III Lisbon assets. As we have consistently stated, commercialization is now driving construction. Accordingly, we are already in advanced negotiations regarding the IT capacity of the entire Lisbon Campus, including not only building one and two that we reported in the last quarter, but also the three assets of phase III, building three, four, and five. Madrid Getafe II is progressing well, with demolition works expected to be completed by year-end 2026, and the construction license anticipated during the Q1 of 2027. Importantly, capacity has been already reserved ahead of the start of the construction.
Finally, at Madrid Tres Cantos 01, following completion of the planning process, urbanization works are now underway, and we expect to obtain construction license by the end of the year, allowing construction to begin during the Q1 of 2027. Pages 24, 25, and 26 illustrate the significant progress achieved in Bilbao Arasur 1 and 2. Many of you will easily recognize the difference compared to the site visit during our Capital Markets Day in March, as well as the progress in Lisbon, Vila Franca de Xira 01 and 02. Moving now into phase III, where we have 406 MW under development. At Bilbao Arasur 4 and 5, where we have power capacity already being secured, execution of the transmission line by the utility company is still pending. Meanwhile, the construction license application has been already submitted and is progressing.
At the same time, in Lisbon, buildings three, four, and five, construction has commenced simultaneously across all three buildings following the granting of the construction license. This decision reflects the advanced stage of our commercialization negotiations for the full campus. As a result, the expected ready-for-service date of these three assets has significantly accelerated. Instead of deliveries originally scheduled for the H1 of 2029, building three, H1 of 2030 for building four, and H1 of building five in 2031, all three buildings are now expected to be ready for service during the H1 of 2029. Just making a quick number under the rents disclosed during the Capital Markets Day, we are talking that we are advancing probably EUR 150 million forward just of accelerating this construction.
Finally, at Zaragoza-WIND 01, where the power capacity again has been already secured, we obtained the Declaration of Regional Interest, called DIGA approval, in July. The next milestone will be the imminent submission of the DIGA, which is the Declaration of Regional Interest for this specific project. It will be an imminent submission for obtain construction license, expected in H1 of 2027, so we can start construction in the Q3 of 2027. The change as well that we have executed in this project is to move from the regional two buildings into one single building of 144 MW IT, that we expect to have that it's ready for service during the H2 of 2029.
Again, the fact that we have advanced on that project is moving our initial second ready for service for the first building in Q4 of 2029 and second building in 2031. Right now, we have saved a significant timing during that project. Finally, as a result of the progress achieved in both phases II and III, we have updated the profile of our CapEx commitments. This now reflects higher CapEx deployment during 2026, 2027, and 2028, while at the same time bringing forward the associated rental income and cash flow generation. Now Ismael will close this presentation with the closing remarks and outlook before entering into Q&A.
Thank you, Fran. Well, once again, just to stress that we saw a relatively strong semester in terms of operations with double-digit revenue growth, good FFO growth, despite higher financial expenses that we have been anticipating to the market for months or years now, and probably will continue in the future. That means basically that the portfolio quality that has been significantly refined over the course of the past two, three years is now clearly supporting the resilient performance of the traditional asset classes. We continue creating value through development pipeline, even in the traditional portfolio. In offices, we have two, three very interesting redevelopments now going up, and we think we are going to obtain very interesting yields on those.
We continue developing some logistics, although it is true that the construction costs are now more difficult to overcome when it comes to justifying going forward with certain projects in certain corridors, because between the increase in land prices, which in our case is not a factor because we have our own land bank. Second, the increase in urban charges by the municipalities, and third, the increase in construction costs. It is now relatively difficult to justify doing a development of logistics from scratch. Likewise, it starts to happen also in offices, except in cases where the building is clearly ours and there is a very big delta between the passing rent and the market rent, which is achievable. However, data centers continue to be our main growth factor.
Thank God, the cycle of data centers seems to be completely dissociated from the consumption GDP cycle that affects offices, logistics, and shopping centers. We seem to be affected more by the worries or hypes that the market feels in every moment about the AI, which, by the way, I believe is a false debate because those worries or hypes should be broadly associated with the evaluations of AI, but not with the adoption cycle. Because in terms of adoption, as we can evidence every day, the adoption continues to skyrocket, and now we are trailing behind our own commercialization efforts. We are commercializing much quicker than we can deliver product to market because there is now a very significant sample of potential clients. All of them are now looking actively for IT capacity across the world, and particularly in the Iberian Peninsula.
At least we know that we are in a sector where the demand at present seems to be endless. At some point, it will probably stall, but I believe this is still relatively far in time. The continued worries about the CapEx expenditure of the hyperscalers, you know my theory, I believe that eventually will end up helping us because, at some point, the market is clearly not rewarding high ROE firms like the technology firms in the U.S. entering into very significant CapEx investments. Eventually, I believe this will favor a specialization of capital, and those who are specialized in building data centers will be the data center builders, and those who are specialized in silicon operation will be the silicon operators. I believe this is a trend which will help us in the future.
In terms of the different phases, the phase I is clearly fully de-risked and cash flowing. The phase II, well, depending on how you measure, it could be between 20%-70% de-risk or 30%-70% de-risk. We have started now entertaining conversations for a number of assets on phase III. Regarding performance for the year, which is the most immediate future, we have slightly increased our FFO guidance by EUR 0.02 per share. I know many of you believe that this is still very conservative, and we can still beat that number. Not so sure, because the two halves of the year are very different in terms of cash flow profile. Anyway, we will do our best to try and beat that new guidance.
As commented, I believe the most salient message for today is that we are very much on track to far exceed the full- year 2026, 200 MW IT leasing guidance. We could perfectly end the year at 340, which would be a very significant achievement. With our current low LTV, high liquidity, and no debt maturities on-site, we remain uniquely positioned to continue funding our growth pipeline and delivering growth to all of you while maintaining a conservative risk profile. Without further preamble, I think we should move into Q&A, and we are at your disposal for any questions you might have.
Okay. Thank you, Ismael. The first question comes from the line of Jonathan from Goldman Sachs. Jonathan, the line is yours.
Good afternoon. Thank you for taking my question. Great progress on the data centers. Not to push you further even, can you highlight what your progress is on phase IV in terms of getting the [inaudible] and Navalmoral? That would be the first question. More generally, I think you've given already quite a lot of color, but the second one is, can you give us a bit more color in terms of the discussions with the potential tenants, and what is competition doing currently? Is it being pushed back? Last question, just construction costs, are you seeing any increase? Thank you.
Okay. Well, regarding construction costs, we continue to keep them more or less under control, although it is clear that particularly in equipment, we are starting to see a number of bottlenecks in terms of delivery times that might eventually end up also pushing up the cost lines. In that respect, what we are doing now is everything that we have announced to you, everything that we have now under construction, we have already done all the procurement of all the equipment for those buildings. We are seriously considering also anticipating a little bit the unspecific equipment, a little bit, particularly transformers, although this is a little bit specific to every design. Gensets, we are considering about the possibility of anticipating also some purchases and storing them in preparation of the most immediate future pipeline that we are handling.
Regarding phase IV and Navalmoral electricity, well, in Navalmoral, as you know, we have been granted 29 MW of electricity, which are good for around 20 MW of IT capacity. Our intention is to start construction of a W96 building as soon as we receive the construction license. The reason why we haven't mentioned it specifically today is simply because we are finishing the environmental impact assessment phase. We have received a number of comments to the dossier by mainly eco-activists. We need to basically reply, or the authority needs to reply to those questions. Only when this phase is duly cleared, we will be in a position to receive a clean environmental impact assessment, and therefore, that will be communicated to the municipality so that they can issue the construction license.
All that, in Navalmoral, should happen before end of September. You never know when you are dealing with administrative procedures. It may take a little bit longer. In principle, we should start constructing at the end of September there. Again, stressing the concept, it's a big box with just a partial equipment because we are anticipating RFS in case during the two-year construction period, we get the significant power that has been requested in that substation, which by the way, has it. This is a little bit repeating what we did in phase I, because we are very conscious that our biggest, not problem, but our biggest challenge at present is to be able to cope with demand. This is why, eventually it will be even better to start two buildings, but I don't dare.
I don't dare to start two buildings, which is two boxes for 200 MW, with only 20 MW of granted capacity. We are a little bit at the mercy of the national grid authority. There is a power contest. We are ranking first. We were the ones that provoked the power contest. We deposited our down payments in February 2025. Situation in Spain regarding grid is always complicated, not so much in Portugal that we can discuss. Regarding tenants, what I can say is that we have discussed always in the past about two sources of tenants, hyperscalers and neoclouds. I think at present, we are starting to see a 3rd type of tenants, which is Chinese, for those who are good with them. A 4th type of tenant, which is big model companies.
Some of them now in the middle of IPO processes, trying to secure IT capacity so that they can make good their own IT operation forecasts to the market. They are trying to get IT capacity straight, without depending on neoclouds or hyperscalers alike. I believe now there is a significant depth in the market compared to the past when we started doing data centers. The depth of the market was limited. Now it's starting to be much deeper. I think this is all.
Super interesting. Can you follow up on the new demand, like the Chinese demand? For model companies, it's fine. Also, are you seeing corporate demand? Like some of these Chinese models that are appearing, they'll need to be run on data centers too, no?
Yes. Chinese demand that we have seen in recent months, we have seen Alibaba. We have seen cloud, mainly cloud at present, not so much Chinese model makers. We have seen TikTok. I think that Tencent Weibo, this is the Chinese demand that we have recently seen in the market.
Okay. That's all super interesting. Thank you so much.
You're welcome.
The next question comes from the line of [Marios]. [Marios], the line is yours.
Great. Thank you for taking my questions. Hopefully, you can hear me okay.
Yep. Perfect.
Fantastic. Great. I have three questions from my side, maybe hit them one by one. The first is on Bilbao Arasur 1. I think at the Q1 you commented there were some hesitations around you signing a full lease. I think there was some uncertainty around the fulfillment of contractual obligations. What has changed there for this to result in a full pre-let versus the advanced negotiations?
Okay. On that one, Marios, thank you for the question. What we commented is that we have always this debate, whether how advanced we can sign a contract, what is the visibility we have on the construction, how it evolves on the procurement of equipment, et cetera, on the permitting of different lines, service stations, et cetera. That's one topic on our side. The other one is as well, how the client is seen to secure capacity to in advance while they try to match reservation of capacity, acquisition of equipment, and final client that will take that computing capacity in the future. That debate was agreed at the very beginning in Q1 with a book reservation agreement, which meant that we agreed that we have a contract, not yet signed, we will look for the right moment to sign it.
Both parties, we have enough visibility, as said on the different topics moving forward. We have agreed that we have reached that visibility by 30th of June this year. If you see in the pictures, we are advancing pretty significantly on the building. All capacity, as Ismael said, all the different equipment from our side is already procured. In terms of connectivity with the utility as well, this is all the different main transformers and lines and equipment has been already ordered. The level of risk is more limited on our side, from the client perspective, it's exactly the same. They have seen more visibility, probably signing already the final client, basically, we agreed to transform that booking agreement into a real pre-let asset still 18 months ahead of ready for service date.
With that, more visibility as what we have at the beginning of the year. This is a trend that, as Ismael was saying, in this one, another example, we are seeing more and more that commercialization is moving forward and moving quicker than even our development capacities. It's always this balance between advancing too much in commercialization and pre-letting almost turnkey projects to people versus how the risk we are facing in case of any delay. As soon as we are seeing more visibility and, of course, we are gaining more confidence on what we are doing, then is when we transform this in the real leases.
Very clear. Thank you. Secondly, on Lisbon, with now all the phases classified as under negotiation, is this part of the Gigafactory project, or are these separate negotiations that you're doing on your own? If you could provide any color on the types of tenants that are looking at this space would be helpful.
Well, we are trying to make it compatible with the Portuguese Gigafactory project. We remain committed to providing the Portuguese Gigafactory a home, we are exploring the possibility of making both situations compatible through a direct dialogue between the potential client and the Portuguese government. I believe, eventually, it might be the same and one single thing. Very interesting.
Clear. Similar to what we've seen in the past, if maybe this project doesn't go ahead, you've then got your own agreements in place to then push forward with a lease without this project going ahead. Is that the case?
Yes. It's exactly as you said. Basically, we have a private deal with a private client, a private counterparty. However, we are trying to make it compatible with including the public side in the equation, assigning to the public side part of the capacity to be recovered in the future through an increase in power in that same campus, that we are, of course, requesting power from the national grid authority in Portugal. There are a number of ways in which that can be achieved, this is what we are trying to get.
Okay. Thank you. Just finally, apologies, going back to slide 22. I know we've discussed the promote fee a few times, but can I just make absolutely certain that the EUR 101 million that is being accrued to date is based on the kind of total exit value, so you're not expecting another similar EUR 100 million or so to be accrued based on the estimated value capture to come, for example? What are your expectations around that total today on phase I?
Yeah. The promote, what covers is the value created over this 10-year period, which means it's not only an exit value itself, it's also basically the rents we are considering all over the years. As I explained before, right now we are in the, let's say, more suited spot of that calculation because we have all the phase I already completed, 100% let. That is basically captured, as you have seen, significantly by the appraisals, because right now what we have is an asset up and running and with a huge market that could be after it. Right now is, let's say, we are achieving the highest level of return, and therefore, the sharing of value is higher.
Going forward, because this calculation will be made on a 10-year basis, meaning that we are in year five right now, we need to move until year 10. What will happen is that we expect to consolidate that value. We also expect that the market will appreciate and convert this as a category more mature market, which means that we will have basically some uplift there in the exit value in 2031. This, of course, our expectation. In the meantime, we will receive rents over the period, which apart from delivering more margin to our projects, at the same time is, of course, reducing IRR from this calculation perspective because we are moving forward from a five-year calculation to a 10-year calculation.
I think the number right now is always accounted from an auditor perspective, on the most conservative way, which is assuming that we have a sale. We have this effective year 10, in every quarter that we are making public our results. Our expectation is if you compare, for example, June 2025, December 2025, there was no significant reprice because there were no commercial decisions at that moment in time. They promote decline slightly because of the more timing on that calculation. Same could happen from now till 2031, because of the elapsing of timing. Okay? That is the way how we are calculating it and what we will expect going forward.
Okay. Thank you very much.
You are welcome.
Thank you. The next question comes from the line of Florent Laroche from ODDO. Just as a reminder, in order to be added to the queue, you need to press star five. Florent, the floor is yours.
Yes. Good afternoon. Thank you for this presentation. I will have two questions. I can ask one by one. The first one is linked to the data centers and capital increase. We can see that you are quite optimistic to continue to sign further leases in data centers. Maybe at the end, in which way the fact that you are funded at this stage only partially, the development of data centers or with capital increase can be maybe a blocking point to sign further leases for phase III. [inaudible]
Okay. We are perfectly conscious, Florent, that we are partially funded for the development of phase III, and we can assure you that we are trying to explore any potential avenues to continue completing our funding. We will be active in the market. We will make sure that we have our phase III completely funded or at least two-thirds funded for the moment, that if the shit hits the fan, at least we can proceed with phase III, with just a little bit more of debt. We need to achieve that point of certainty of execution, which I am sure the market is expecting from us. We will be active during the rest of the year.
Okay. That is very interesting. Maybe a second question on shopping centers. We can see that your KPIs are very good again, this quarter. How is sustainable for you the operational performance of your shopping centers at this level?
I am the first to be surprised sometimes about the robustness of the shopping center performance. We have been in the business for many, many years, and we thought that the numbers we had achieved in 2019 were more or less irrepeatable, that we exceeded 2003 probably already in 2023. We beat 2023 in 2024, we beat 2024 in 2025, and this year again, we are beating 2025. Is this sustainable? Probably more a question for a Bank of Spain macroeconomist than for me. I believe it is difficult that we can sustain this rate of operation for many, many years.
It is true that for good or for bad, we have a number of factors which are helping data center shopping centers in Spain. One is mainly increase in population. Spain is one of the few countries in Europe that, without judging whether this is right or wrong, it is importing population from mainly African and South American countries. That population increase, of course, that goes to shopping centers. Second, the average salaries are going up as a consequence of inflation, although so are going taxes, et cetera. It is just a factor. The average indebtedness of Spanish households is very low, about 41% of GDP.
In reality, we are not in a position like in the U.S. where you have your card debts piling up, student debts, credit card, many different types of debt. In Spain, it's mainly mortgage. Mortgage is going down. The total stock of mortgages is going down as time lapses because the average Spanish mortgage is calculated on a French amortization system. Normally, the monthly payment is equal, and at the beginning, it pays mainly interest that, reaching approximately half of the life of the mortgage, you start paying significant amount of capital.
As a consequence, the mortgage stock in Spanish banks has been going down for a number of years. Now, it's a little bit more stable. There is always the factor of informal economy. Household services have now completely moved into informal. Even residential rents, residential leases have moved into informal, because people doesn't want to be under the radar of the legal system, because if under the legal system, if you have a non-compliant tenant, it will be in your house forever. In the informal system, you take it out with other methods very quickly. That motivates a lot of cash in the system that, of course, is appearing in shopping centers. How sustainable is that? I don't really know.
The good thing, in our case, is that the increase in per square meter sales of our tenants has not resulted in us elevating or increasing our rents on a commensurated basis. This is why the OCR is going down, which means basically we have an ample room for maneuver, where the shopping center industry hit the wall at some point, and it wouldn't find us with OCRs at 16, 18%, in which case we would clearly have a problem.
Okay, thank you. The next question comes from the line of Ana Escalante from Morgan Stanley. Ana, the floor is yours.
My first question is on the type of tenants. Ismael, I think you've mentioned earlier during the call that you are seeing demand arising from other type of tenants, not just neoclouds and hyperscalers. Based on your pre-lets, I know that there is some information that you cannot disclose, of course, but I wonder, how are you looking at that split at the moment in terms of the pre-lets and the bookings that you are signing right now? To what extent you are prioritizing whoever is early or ready to move in, or whether you have already started to try to diversify a little bit away from some of the neoclouds into other companies to minimize counterparty risk?
Thank you, Ana. As Ismael said, we are seeing, for the type of assets we are developing, three different type of potential clients. First ones are the traditional hyperscalers, clearly moving from a more cloud type of request to AI. Still, they are in that process, sometimes securing capacity a little bit ahead of what they will need in the future. Considering the type of client they are, they tend to standardize all the different equipment, all the different assets they have or they will let, and therefore, the time to market is not as quick as probably others. Still, they are in the market, and they continue representing still a small amount of our client base. I'm counting on, as you said, on pre-lets and bookings as well. They're becoming more actively.
We are seeing this because the amounts of capacity requested is a little bit increasing, and the timeline and the ramp-up that they were considering is clearly moving forward. The second type of clients that we mentioned several times are the neoclouds, companies that were created in the last years, seeing a lack of product of computing capacity between what hyperscale is normally contracted and the final users of that computing capacity.
What they have taken, basically, is the opportunity of jumping into the sector. More difficulties, of course, because of the capital barrier. Little by little, we are seeing different categories between neoclouds that they are becoming little by little a bit hyperscalers in terms of size with very good access to capital and depth, and more importantly, normally taking capacity as quick as they secure final contracts. Hyperscalers can, in a way, take more risk of securing capacity ahead of what they will expect to have in the future.
These new clouds, normally because of their financial requirements, equity requirements, they normally take capacity completely, simultaneously to the final client acquiring that computing capacity. This is something good to know that the risk is more limited. I said there are two categories. There are some of the clients, listed companies, CoreWeave or Nebius AI are examples of those listed on the American Stock Exchange. There are others that they're trying to jump into that list. We are pretty sure that in the near term, they will be there as well, increasing the amount of potential clients in this classification.
The third one, as Ismael was approaching, is those that are technically clients of these new clouds/hyperscalers, seen or in light of the scarcity of capacity that they're foreseeing in the future, what they're trying to do is to move a little bit outward in the chain, trying to secure that capacity so they can guarantee their computing services in the future. Whether they operate themselves or they will subsequently contract somebody to operate that stack for them is a different question. They want to secure that capacity. All of this reminds a little bit what we have seen in logistics years ago, where some of the companies, operators, they were seeing no capacity available.
They first secure that capacity, later on, whether they operate themselves or they contract other service providers, they are securing those type of locations. We have several examples in our portfolio where we have final clients taking capacity, despite the fact that when you go to the warehouse, you see the name of an operator instead of the final client. How this is moving, what initially in a market represented almost everything is hyperscalers and some raising effort from new clouds. In our case, because we are more brand new clouds represented a significant amount of our capacity, far above the one requested by hyperscalers.
What we are seeing right now is precisely that these AI models specialists are trying to catch up with these new clouds. What we are foreseeing is that, I would not say one-third, one-third, one-third, probably new clouds and final clients will represent a little bit higher than hyperscalers. It's also true that they are trying to catch up. At some moment in time, not yet with us, at some moment in time, these massive hyperscalers will probably try to jump and, in a way, diversify a little bit more our portfolio.
Another trend we are seeing, and I finalize with that, is that originally we have several clients within the same building. As you have seen, the blocks of capacity that is being contracted by clients is increasing very significantly. We have moved from modules to buildings and from buildings to campuses. That is moving us in order to diversify that clientele to add more buildings into our portfolio. What we are doing with the different phases is to bring more buildings and with that, diversify the type of clientele and the type of tenants that we will have.
Another interesting thing is that the neoclouds, at the beginning, they used to compute mainly for other hyperscalers or large language model companies. Now more and more, we are seeing direct computing for final corporate clients. Big European industrial companies computing, basically inferencing the models they have already trained. They are using the services of the neoclouds for that kind of inference. It's very interesting because that gives also a new layer of reality to the market, which is very much welcome.
Thank you. Very clear. Then my second question is again on the promote. I know that cash flows matter a lot, but as Ismael said once, data centers is a cash drain division until stabilization. However, in the meantime, you are creating a lot of value through development profits or revaluations that come earlier than these future cash flows that the data centers will eventually generate. Therefore, for the market, it's quite important to understand how much of those embedded revaluations you might give away in the form of a fee to a third party. I'm not sure I have understood yet how much that could be, especially for phase II and phase III.
My understanding is that changes, of course, over time because it's dependent on the profitability of the projects. If you could give us a range of, I don't know, whether that could be 20% of the potential revaluation or to be around 15%-30% of the potential revaluation of data centers + 10% of the rents over 10 years, something like that would be very helpful for us to understand how to think about how much value MERLIN is going to take from this super value accretive story.
Thank you, Ana. I said before, the way that the promote is calculated is based on profitability. There are some margin of profitability that we consider that the technology brought and the capacity brought, in a way, is not improving what we could have found in the sector. There is some profitability within the 10 years that deserves to share that profit with our partner. The exit value, of course, in a 10-year discount cash flow has an impact. Having an exit value, whether it is much higher, or higher, or on average, of course, has an impact, but also does the rent over the period. You can more or less determine that if we are getting to net loan cost, roughly on the range of 11% out of this 15%, if you apply more or less a gross to net.
It means that every year, what you are getting in reality, knowing that the uplift in valuation is at the very end in year 10, is more or less an 11% that will be updated. That is more or less what you should consider that will be the range of appropriation, which will be more in line to the rent than to the exit value. Of course, if we are seeing a market that all of a sudden matures very significantly, after 10 years of investment and management provokes an uplift in values, of course, it will have an impact on the total number. I would say that over a 10-year period, the rent has a lot of things to say during that calculation.
In a nutshell, more or less, this is more or less defined to be between the 10% and the 15% of the total profit of the portfolio over time. We will be on the low end if, for whatever reason, the market is not appreciating the assets and the conversation we are doing all over the period, will be on the upper end in the case that the market, as said, matures, it stabilizes, all of a sudden, these are capital markets that are out there that will price the assets higher than they are right now. That is more or less the range. I do not know if with that I have properly answered your question.
Thank you.
Okay, thank you. The next question comes from the line of Paul May from Barclays. Paul, the line's yours.
Hello, everyone. Thanks for taking my questions. Apologies if they're simple, given I'm new to this. Just wondered, firstly, what is the lowest level of ICR on a quarterly basis that you're willing to go to as you accelerate the DC rollout?
Lowest level of ICR?
Yeah, on a quarterly basis, just as you roll out the DC development. Just wonder what you're willing to go to in terms of how low.
Let us check. Right now, we are at 3.7 x.
Times.
If I'm not mistaken, which basically drives as well the 25% roughly of loan-to-value that we have right now. As we commented, we have been very openly on that, we are not willing to exceed above the 32%, maximum 35% loan-to-value. On that sense, even if rents is, or even interests are raising, and cost of debt is going pretty high, with this low leverage, we will never be. We have a covenant of 2.5 x ICR, and I think we haven't been ever below 3x . That is a little bit the spirit of the company, that it's more linked to the low loan-to-value target we have, than to the interest rates. We are less affected by interest volatility as compared to probably other companies who are highly leveraged.
We also take a look at, Paul, at net debt to EBITDA. Both things, LTV and.
The idea is to be below 10.
10 x. Exactly. Way below that figure.
LTV below 35. Our initial model was giving us temporary breaking of that LTV ratio at around 37, but the new versions of the model have gone significantly down because the value creation in data centers is being bigger than we expected and is coming earlier. The model is now giving 34 max, which is good. In terms of net debt to EBITDA, in the original versions of the model, we were going as high as 11.7, close to 12 x, but now it's also coming down, and it's going to be more between 10 and 11-ish during a certain peak.
It will go down to as low as, the model is giving us seven at the end of the period. This is the way we look at it. Of course, I will check into the ICR for a moment. Are you a credit person? Are you trying to price our debt, or are you calculating the amount of capital we need? Are you calculating the amount of capital, right?
It's purely on the equity side, but it's just one of those things that as you obviously increase CapEx ahead of revenue recognition, then there could be an impact that comes through. It'd be great if you could get back to me on where that ICR goes on those models that you mentioned relating to the LTV and the net debt, would be great.
We'll check with the model and go back to you. One thing is important. In current times, debt is not so much accretive. With the current course of issuing debt, particularly if you follow a real benchmark, which is the 10-year unsecured bond of your company. If you take that as a benchmark, there are two things that of course come to my mind. First is that debt is not that accretive, do not play too much with debt because it's delicate.
Second, I believe sooner or later, particularly in traditional asset classes, valuations will start to suffer. Particularly LTV might suffer from a completely unexpected factor, which is the decrease of the V, because the L will be constant or slightly growing, but the V will go down. Be assured that we are a debt-conscious company, and we will try to keep debt at bay, because debt, in reality, is the only thing that can kill a success story in the real world.
Indeed. You're preaching to the converted. Very much an equity fan here. A few questions on the neo clouds, if you wouldn't mind. Apologies. What proportion of the DC revenue at the moment is exposed to neo clouds, including the pre-lets? Is it 100%, or do you have a spread of tenants? How do you feel about the credit quality of the neo clouds? Obviously, debt's been increasing in those businesses, and their credit spreads have been widening. Just wonder what your thoughts are and probably plays into the comments earlier around the spread of tenants. Just wonder if you'd give any details on the contract terms. I think a 10-year duration. I just wonder what extension and termination terms there are within those. It'd be great. Thank you.
Okay. Go.
Considering basically what we have on our books right now, 45% are represented by neoclouds. 5% are represented by hyperscalers and growing, and roughly another 45%-50% are represented by-
Other.
Other. Okay.
That considering 340 commercialization, we are taking the poetical license of considering the 180 let, which is not yet what in purity we should do. It will take time to convert. yeah, considering 340, what Fran told you is the proportion. 45 neocloud, about 5- 10 hyperscalers, and 45 - 50 other.
As said, as soon as we are opening new campuses of big size in certain areas, this can change because if one of the other two campuses is taken by a hyperscaler, then all the mix changes dramatically. As soon as we are having a larger portfolio, then you can extract a little bit more conclusions from our tenants base. In terms of conditions, we normally tend to sign on a 10-year basis, mandatory. Clients normally reserve significant amount of expansions, extensions, or renewal options for them, and mandatory for us. That is more or less across all different type of clients. Technically or initially, hyperscalers tend to be a little bit more longer. We have some of them, but more based on previous times. Little by little, more or less, they are more committing to be more on the 10-year time of term.
This is what we are seeing right now, whether one category or the other. As said, it depends a little bit as well, whether they want to make sure that the capacity is blocked for a certain period of time, if they have any special infrastructure there, or they're bringing some very well advanced, or they're foreseen to come to bring very well advanced NVIDIA type of equipment. Sometimes they go a bit further, but 10, I would say, is a well-established rule in the market right now.
Just on the credit quality, just recently, obviously slight deterioration there in the spread widening. Just wonder if you've got any issues on the neocloud side, or would you still be comfortable signing some new contracts with them?
As said, they are different, I think they are now set in different categories, these new cloud operators. We are, in a way, ranking them, and the market is ranking them, linked to the access they have to equity and debt markets. In the ones that they are listed companies, an equity market is available for them, normally now trading with higher levels as compared to IPO times. If they have very good access to bond market in different formats, that is basically, in a way, rating for us that capacity. If there are debt involved, in 100% of the cases, there is a final client sign. This is what the requirements that both bond holders and financiers are requesting from these type of clients. You know that that capacity is sold at the time that they're signing with you.
The way we can, in a way, check that is on the power consumption that these clients are having as soon as they take possession of the different rooms, and it is what we are experiencing right now. Why they are not rated? Sometimes they are rated on the different issuances they are doing, but not rated on a corporate level, because the rating agencies, of course, are not so happy with investing significant amount of cash flow into new CapEx and not keeping a little bit some money for reducing debt.
They are more focused on getting better terms on debt, they are more in a safe harbor on that basis. They are still in a growth mood and reinvesting significant amount of free cash flow into new CapEx. This is what enables them to get a rating right now. If you see the margins, which is important thing from the top line to the EBITDA, that the margins that they are doing, all of them are extremely healthy. This is what normally we look at as compared to the traditional hyperscalers, which is a much easier way of measuring that risk capacity.
I am sorry to repeat a question or to ask again on the promote side of things, and apologies again, I am relatively new to this. It seems a relatively large amount on phase I. I think it is 27% of current revaluation, 14% of the total expected revaluation. Just wondered what we should expect on, say, phases two and three. Is it a similar structure on phases two and three, or is phase I a larger promote and then it tails off through phases two and three? Also linked to this, the yield on cost that you quote, does that include effectively the cost of the promote, or is that pre any promote payments in terms of the yield on cost? Thank you.
Yeah. The figure that you are now calculating, as said before, instead of calculating over a 10-year period of time, you are calculating over only over a five-year period of time, and just only applying this to the valuation itself. That is the reason why you get to these percentages. As soon as you are expanding this over a 10-year period of time, where the rents will be more significant, and with a significant amount of it, then that percentage will decline. Because I said that promote is calculated on an IRR basis. The longest you calculate, the smaller the IRR is. Of course, over a higher margin.
They will get a lower percentage over a higher amount, that one thing compensate the other is the reason why we believe for phase I, we are more or less in the level that should be at the very end. Regarding other phases, the way that we are structuring it is pretty similar to that. We are always having a look at what is the percentage representing the total profit and the value that is being created, to be commensurate with the size and the value added brought into the table. I would say that as soon as you are seeing more evolving, you will, let's say, have a more concrete number, but between, as I said before, between 10% and 15% of total profit is a good rule of thumb if things are going as they are doing right now. Which we are happy with it.
Cool. Does the yield on cost include that cost of the promote, or should we account for that separately?
Great, we have an original number because we have not generated any promote yet.
Yeah.
We were reporting on a gross basis, now little by little, you will need to make that calculation separately.
Perfect. Cool. Thank you very much.
Thank you. The next question comes from the line from Thomas from Deutsche Bank. Thomas, the line is yours.
Hi, good afternoon. Two or three question. The first is on the new Arasur lease. I'm wondering if you could comment on the lease terms, just roughly, would be very helpful.
Okay.
Do we see any deviation, or do you see any deviations to your initial expectations?
No. I said, the numbers I think we provide for this second phase were more or less in the average of EUR 122.5 per MW per month. The market continue being so intense and the demand has been so big that these levels are continually exceeding from every single contract we are signing. Doesn't mean that there's no sensitivity to pricing on the counterparty. We are competing not only within potential capacity in Spain, we are competing from a European and even sometimes worldwide type of competition. We need to know that we have a sometimes competitive advantage, sometimes we need to be more conservative. Talking about the sizes that we are talking about is full building, one single lease. Of course, it's always negotiating power from the client perspective, even the demand is pretty big. We are exceeding that, so we're happy. We are above our initial projections.
Yeah.
In terms of length, we are signing, as I said before, 10 years is normally the average, and with different renewal options for the client if they will, from year 10 onwards.
Escalator?
Escalator, we are updating normally the rents between 2% and 3%. That is normally the range that is market standard. Our average right now is more towards 3% than to 2%, but it's more or less the ballpark that we are moving on.
Thanks. The second one is on the Zaragoza-WIND. You plan a single large-scale building ready for service in the H2 of 2029. Maybe you could provide some color on current lease negotiations. Seems like given the size, this is something for a hyperscaler. Is this correct, this assumption?
Well, the assumption is correct. We are adapting to a certain set of technical requirements which are g ood for a number of hyperscalers. We are, let's say, hyperscaler ready. We have decided to go for one single building because we believe there is a demand for that specific type of facility. I think more and more people are conscious about internal communication within the DC. Not simply having the silicon, but having the silicon connected through InfiniBand, et cetera, and being able to synchronize the computing of all the different GPUs in one single, let's say, imaginary machine, which gives you a J curve in terms of performance. Under those requirements, we have decided to go forward with that building. We are still pending license.
We hope we can be in a position to start building by around next summer in 2027. Only God knows, because when you deal with public administration, you never know. We will try to be good around summer next year. About two years construction, which gives us end of H1, H2 2029, which is a good delivery date and shortens significantly the two-building structure that we used to have on the basis of the increased amount of demand we are seeing in the market. We are trying to ready more demand quicker, without sacrificing quality, of course, because we are operators. We are trying to ready as much demand as we can earlier than expected, because we see that at present, the big bottleneck is construction, is not so much commercialization.
Okay. The last question is actually coming back on the data center property values again. Just wondering, to keep it simple, if you could provide a rough idea about what you expect regarding revaluations by the end of the year. Should we expect a similar magnitude, roughly as in the H1?
The revaluations are coming through two main impacts. The first one is when we are adding more capacity to be appraised. As we have commented in previous calls, as soon as we get a construction license, that asset moves into our current whip, and the appraisal basically values that property. We have other lands with power, because we are waiting to receive construction license, we have not started yet on that construction, that capacity is kept at cost. That is now appraised. That is what, for example, has happened now in June, where the three assets in Lisbon, building three, four, and five, because we are building as we speak, those have entered into the scope of values.
That also provoked, as compared to the original valuation that this land had at the time, which was very low, now basically is properly appraised by the appraisers. This is one of the impacts. The second one is when once we are within construction, or we are ready with a building starting construction, we reduce the risk of that development by pre-letting the asset. We have several examples as well during this H1 where we have pre-let in advance capacity during the development time. This, of course, has an impact because the risk or discount rate that appraisers are applying to it's been reduced because the risk assessed is much lower once you have 100% of the commercialization risk already offset. H2 depends on when we are converting these bookings and advance negotiations into leases, this will move forward to have a higher or lower amount of revaluation.
Thank you.
Thank you, Thomas. The next question comes from the line of [Véronique] from Kempen. [Véronique], the line is yours.
Hi, good afternoon, all. Thanks for taking questions. I'll keep it very short. Two quick follow-ups. First, on Lisbon, where you mentioned that you're in advanced negotiation, should I interpret it that if this closes, it's a proper pre-let, or is this more sort of like a booked way to look at it?
No, it is a pre-let. At present, it's already advanced negotiations because we have a head of terms, and this is accompanied by an exchange of technical sheets and basic legal documentation. What we need to do now is move into full format lease agreement, and if we can move into full format lease agreement and signed, it will become a pre-let, technically a pre-let.
Okay, that's good. Thank you. Sorry, one last follow-up on the promote fee. How much have you provisioned so far, and what's the strategy regarding that going forward?
Well, we are provisioning every year what the auditor tells us to provision, which is basically a function of modeling the cash flows of the different projects affected by the promote structure on a 10-year basis, but then calculating an equivalent exit on the year in which we are. This is the amount that we provision every year, that amount goes higher or lower depending on the year. As Fran commented, as time lapses, normally the effect, it smoothens a little bit the IRR, and therefore promotes goes slightly down, although multiple goes up and there is more money on the table. Okay? It's a function of both things. For our partner, they get probably less appropriation, but less appropriation on more money, on a bigger pie. Okay? It's the way it works. We are happy with it. We are loyal people.
We have been working with them for a long period. We like to work with them. Of course, we need to be prepared for the future, and we will be. For the moment, we like this way of working because it helps us to get an external research and development department and not simply use off-the-shelf products available in the market. Although it is true that sooner or later, the technology will end up commoditizing a little bit, and the value brought to the table by such research and development department will be slightly lower.
Okay. That's clear. Thank you.
The next question comes from the line of [Stéphanie] from Jefferies. [Stéphanie], the line is yours.
Hello, can you hear me?
Yep.
Yes.
Hi. Hello, everyone. Most of my questions have been answered, so maybe a last one, a follow-up on the funding. I was wondering, of course, you said your share were down today, and I suspect that investors are waiting the next capital increase. Of course, the price will be under pressure as long as you are closer to your NAV. I appreciate it's a bit tricky to answer such questions, but what would be the trigger for the next capital increase? What are you waiting for in terms of, I don't know, getting it closer to the cash flow generation, or how do you approach that?
I think we have commented on a number of occasions that probably the funding method will be a combination of plain vanilla capital increases and convertible bonds. The first one was a capital increase. Most likely, the second batch will come under the form of a convertible bond because a convertible bond delays the dilution, and the dilution happens under much better share price terms. Of course, the model is giving us certain peaks of equity needs, and we need to be mindful of those. At present, most probably we are going to go down the route of a convertible bond issuance.
All right. Thank you.
Okay.
Thank you. The next question comes from the line of Michael Finn from Green Street. Michael, the line is yours.
Yes, thank you very much. I'll be pretty quick. I have two questions, please. The first one is, as you progress through the phase III planning, I'm just curious if you're seeing a major shift in the M&E needs over time. Obviously, that is something that has changed quite a lot, and I suspect it will probably change more. My second question is on the plan for tour, the name of which is escaping me now, but the tower in...
You mean in Barcelona?
Barcelona, yeah. After Meta leaving, yes. Tour.
Okay.
Glories.
Glories. Yeah.
Well, look, Torre Glories, I wouldn't be that worried because Torre Glories is a, I would say, iconic asset in its market. Clearly it is a price maker rather than a price taker. Of course, the departure of Meta is a big hit, particularly because at present, the 22@ area in which that tower is located is very weak. There is being a significant oversupply coming to market in recent years, and it's been really bad luck to have Meta banned from keeping fake news control centers across the world, and as a consequence, losing that client. Sooner or later, we will start recovering occupancy in that asset, and I am not really worried. If it was another asset within the 22@ area, yes, because 22@ is a tough market at present. It's really Comanche area.
Not with Torre Glories because Torre Glories is a very special asset. Sooner or later, we will start recovering that occupancy. Hopefully, within the year, we will already give you some pieces of good news. Over 2027, we will continue reletting and eventually reaching close to full occupancy on that asset. On phase III, you were commenting on what? On MEP, on the types of equipment for data centers?
Yes, exactly. How that has changed over time. Obviously the standard of the asset has obviously changed, and the tenant base has changed a bit as well. I'm just curious, how has that changed over time and what are you seeing going forward?
Actually, this is a very good question and one that motivates some internal discussions. If you pay attention to what particularly American clients tell you will be building lower quality assets, and that includes lower quality MEP fixtures. However, we are long-term operators, and we don't want to do that. First, we are building assets with white rooms which are larger than actually needed with the current densities of rack. That means basically that we are concentrating very visibly. We are concentrating racks in one corner of the room and leaving the rest of the room empty, so you could play paddle in that side of the room. This is good because concrete and steel, although growing in cost, are just a little portion of the total cost of a data center.
We want to have data centers which are sufficiently flexible in case we need to go from higher density to lower density, or more importantly, in case higher density compute in the future ends up consuming less electricity and we can not repower, redensify, or refill part of our white rooms with extra equipment in case one day someone discovers something which makes the existing state-of-the-art racks a little bit less hard in terms of consumption. The second thing which sometimes, particularly large language model trainers tell you to do is not to fit gen sets on an N+1 basis, mimicking the total IT capacity of the data center, and they tell you to only fit 20% of gen sets needed. Equally, we don't want to do that.
That is good for a model trainer that can stop machinery and wait, but it wouldn't be good for an inference user that needs firm power 24/7. We try to do our things well done. That takes a little bit longer to build. The clients tell us that we build Rolls-Royces. Maybe this is true, but we prefer to build Rolls-Royces and keep them adapted to whatever might come in the future than build lower quality types of builds and then, in the future, discover that we are no longer adapted to whatever is happening in the market. So far, we remain faithful to our original designs. We are fitting good quality gensets, although this is becoming now a real bottleneck in terms of purchasing. We are fitting dry transformers, which are really high quality rather than oil ones in terms of batteries.
We are faithful to the zinc-nickel batteries because they have more happenings but less grave. Lithium-ion, you have less happenings or less incidents, but if you have one, you better pray. We try to do things as best as we can in order to make sure that our facilities are adaptable to whatever comes in the future. We are ahead of the future in the way we build. Of course, we are always aware of the fact that new things come to market, and we keep an eye on new suppliers. Particularly, we try always to make more European our build. Bring on board many more Europeans and/or Spanish-Portuguese suppliers because having your suppliers close to you is very important in terms of after-sales support, in case you encounter any future problems in the way your machinery works.
There are also some radical changes coming in the future. For example, our partners of Endeavour are developing a very interesting machinery called Turbozel, which is already available. Aligned Data Centers has a similar thing working in the U.S. already in operation. It's very interesting, but it's very much U.S.-centric because it's good, particularly with gas. Gas in Europe is an expensive thing. We have to be careful with that, but it's very interesting, particularly if you need to fuel data centers which are located in relatively remote areas where it takes time to bring area aligns with electricity and things like that, you might live on Turbozels for a while. Very interesting, intellectually very encouraging debate with the engineers for the future designs and the future data centers that we are going to build. For the moment, we remain relatively orthodox in the way we build.
Sounds good.
Thank you, Michael. The last question comes from the line of Marc Mozzi from Bank of America. Marc, it's yours.
Yeah, thank you. Good afternoon, everyone. The first question is, can you have just a follow-up on the breakdown of your existing type of tenant in data center? Not on the 340 MW you mentioned, but just on the 180 MW you have prelet so far. What is the proportion of neocloud here and hyperscalers?
Okay. 135, 85% approximately is neocloud, 15% hyperscaler.
Okay. Thank you. Then, as data center will become the largest part of your business, when do you think we should expect some guidance on the depreciation impact of the data center equipment that's going to impact your earnings?
I think we are already depreciating our equipment. Talking by heart, I think we are depreciating, the rule is 15. We are depreciating generators 20, we are depreciating batteries 10, but skids and high tension, mid tension, low tension gear, I believe we are depreciating. Fran can give you more updated numbers because batteries depend also on the technology, zinc-nickel versus lithium-ion. Fran can give you more accurate numbers.
Yeah. All items we have, you have first, as you know, 25%-30% is the construction itself. That depreciates over time like a normal building, over 30 years. Then you have all the big equipment, transformers, skids, cabling generator, et cetera, that normally last for between 15 and 20 years. This is basically long-term. Then you have other components which are more exigent, and that one I will include mainly the batteries. Why is that? Because the batteries, as you know, what they do in a data center is to offset a shortage of power until generators are up and running. But at the same time, they are also very active right now with different peaks of the computing that AI is doing. This is the system that the infrastructure uses to offset or harmonize part of these peaks.
Therefore, the usage of a battery right now in a data center of AI is having more work. Originally, their life expectancy was more on the three-year time, technology has evolved very significantly to move that to the 10 years that Ismael was mentioning. That depending on the usage you are doing on those equipment, it could be more shortened on that period. Thank God, the value of that particular item within the big data center is not big.
All of that is properly calculated, right now, of course, everything is brand new, but we are assuming some protection or some escrow in a way of money for potential contingencies on this type of equipment going forward. It's something that, as you said, right now everything is brand new, and we have in several assets, certain ramp-ups of capacity. From 2027 onwards is when you will see, and we can provide you with more detail of how we are treating all these elements as soon as we move in more in operations.
Excellent. Thank you very much. The final one from me is, what sort of pricing did you get, Ismael, on your next convertible bond?
What other pricing?
What sort of pricing or price range should we expect as a coupon? I don't need that, but just as a coupon, just to assume what sort of refinancing cost you're going to face. Are you going to go for a zero coupon convertible bond like we've seen with Vonovia? Hello? Hello?
Hey, Marc.
Yeah.
Okay. Sorry, the line went off probably because there was an alarm of confidential information. Look, we are not in a position yet to decide how we will structure, but you know the principles. I think because we have openly commented with you on some occasions, we prefer a shorter term rather than a longer term because we want to do it more equity-like than debt-like. We are not simply trying to lower our passing cost of debt by issuing very cheap financing. What we want to do is issue something that with the premium, will resemble very much where we believe our NAV will land in three years from now, and in a way, make sure it converts. That will be the idea and the principle under which we are considering the convertible exercise.
Excellent. Thank you very much. Sorry for I just wanted to help everyone to be capable to improve their forecast on the basis of a new CapEx plan and so on. Thank you.
Okay. Thank you, Marc.
Thank you very much. There are no further questions. We appreciate it's been a long call. If you have any other questions, you know where we are. Hopefully, you enjoy good summer break. For sure, we will. Thank you very much, and goodbye for all.