Good morning again, everyone. Thank you so much for joining to our ACS Investor Day. Today is a very important day. This comes as a follow-up of our Capital Markets Day last year, April 2024. We would like you to get out of this presentation with a few messages. The first one is the recap from the Capital Markets Day: what we said versus what we've done. Everyone will realize that we've been working hard, not just to meet all our commitments in the last Capital Markets Day, but also to increase our targets and outperform our own objectives. The second message, equally important, is about the entire strategy. We are going to focus on data centers. Today is about data centers.
As I said in our Q3 presentation, I would like to explain all different verticals and where we are in the development of each one of them. Why? Because all of them will be high-growth areas in the same way that data center is today a high-growth area for us. You look at our numbers in 2025, and you realize the relevance of our digital strategy in today's performance. Equally important will be the impact of the other vertical areas in our future. It's very important for us to focus on two things. First one, our firepower, our capital allocation. We'll spend significant time talking about it. As I said in my Q3 presentation, we're very comfortable with our net operational cash flow performance and how much firepower it's going to give us during the next years.
Yes, there's potential investments that we're following, but we're not relying any more on them as we used to rely when we explained our Capital Markets Day. Also, it's the valuation. The valuation of the additional revenues we will generate in the business through data centers, but also from the assets that we are going to be creating at the platform that we announced today, and we'll explain and describe the platform. Also, the platform that we announced in Germany through the edge data centers a few months ago, and then in the rest of the business. We'll try to give an overview of where we believe that ACS will be over the next years, from now to 2030 and beyond. Let me start with the strategy recap. The first one, the top-line growth, right?
We were talking about different verticals, how we were going to focus on them back in April 2025. We explained that we wanted to be in a lot of different areas, infrastructure. That is why, because we believe that the entire infrastructure in the world has to be reset. We also believe that each one of the different verticals, when it comes to infrastructure, digital, energy, artificial intelligence, defense, etc., they are connected. Nowadays, it is very difficult to build a data center without understanding the energy component. Also understanding that energy in five to ten years from now will mean nuclear. Every time we qualify in a defense project, they ask us about our experience in digital, AI, nuclear, civil, general building, our supply chain, and systems. The same thing when it comes to critical metals. All the different verticals are related.
If you really want to be an engineering and construction firm in the future, you need to make sure that you are a leader in each one of these areas. You will see as well the efforts that we're doing in modular construction, because it's part of the future. We are taking a lot of our modular workshops from the past, where we used to do our ring segments for tunnels or girders, beams, etc., and we're really converting them into modular workshops for a lot of infrastructure in the future. Thanks to all of this, now we can say that we have improved in 20% the backlog that we had back at the beginning of 2024, from the $74 billion to the $89 billion that we're at right now.
This is a consequence that the strategy that we put together at that point was the right one, embracing the future and embracing the high-growth areas. The other thing that we said is that not only we wanted to increase revenues, but we wanted to increase our margins. We went from the 5.3% EBITDA margin that we had back at the beginning of 2024 to 6% that we have right now. We expect to continue increasing that margin as we enhance our engineering capabilities, as we continue delivering more value, as we embrace the projects in the future, but also as we continue becoming more and more lean operational, which takes me to the next point. We have been simplifying our structure in two ways. The first one, you have seen the merger of Latera and Druckhaus in North America, some acquisitions, how we are removing layers.
Also, functionally, we've been transitioning into a more high-tech organization, which means that certain components have to be centralized. That means supply chain, engineering, systems, strategy. For us, centralization doesn't mean that we're putting additional layers on top. It means that some of our companies become champions on certain projects. You will see the role that Turner is having in our global digital strategy and artificial intelligence as well, including semiconductors, but also the role that Hochtief is having in nuclear, etc. We're having champions in each one of the areas. We are centralizing the approach to make sure that we work as one group, one team. One group, one team is much more than one slogan. It's a way of living within ACS where we all feel part of the same objective and the future.
As we do so, and we embrace artificial intelligence, we've been able to get, through the integration of the operations, $70 million of additional value per year in the reduction of our structure. We will be continuing increasing that amount significantly as we move forward. You are not seeing that amount yet in our figures, obviously, because of the cost of redundancies, and that offsets part of that. Our intention is that that will start being reflected in our numbers in the future. Net operating cash flow, if you recall our CMD back in April 2024, we were announcing an average of $1.1 billion-$1.3 billion net operating cash flow for the future. Clearly, we have outperformed that figure, and we are about the $1.5 billion.
Finally, the long-term value that we have created and we expect to continue creating for our shareholders of more than 60% return to date if you take into consideration our evaluation plus the EUR 3.50 per share since April 2024 in the one-year-and-a-half period. We expect that if we continue this trajectory, we will be able to increase dividends in the group. Continuing with our CMD and more specifically focusing on some of the areas, right? Starting with integrated solutions, where Turner has been a real champion in their performance and in the way they are approaching the future. We announced for Turner a margin of 3.5% EBITDA by 2026. You will realize that in 2025, we will be achieving that 3.5% average with a strong Q4 that they will go close, I mean, above the 3.7%. Turner EBITDA is expected to grow in 2026.
If you consider the increasing guidance that we have announced for Turner with a 58% in euros, and you look at the performance of 80% growth, FX adjusted in US dollars at 100%, just looking at the 30% additional increase in EBITDA for 2026 realizes the strong performance and the future of Turner. Turner has right now like a 40% advanced tech portion of their projects in their backlog, and data centers represent around $12 billion. The data centers and the advanced technology projects, we expect them to continue growing in their balance sheet as we move into the future. An important part of this presentation will be infrastructure. As I always say, and as I said before, our strategy had always three pillars. The first one, reducing risk. We want sustainable cash flows. We want certainty of cash flows.
We don't want to start fighting for lump sum design build where the risk is uncertain and there's always one project that goes wrong and jeopardizes the entire strategy. Decreasing risk is a very important part of our strategy. The second one was to make sure that we were embracing all the verticals of the high-growth areas in the future. As traditional civil and traditional general building is expected to continue growing depending on the country between 4% and 8%, but it's still low to what we want to achieve. That's why we embrace the high-growth areas, and that's why we are achieving the levels of growth that we're achieving. The third part was to make sure that we were generating stable EBITDAs coming from our assets. That's where it was very important, our greenfield strategy of managed lanes in North America.
We announced that we're going to invest in those projects. We won Georgia 400 last year. We are pre-qualified in four projects right now, about to be pre-qualified on a fifth one. We will be spending some time on those jobs. On the brownfield side of things, the figures of Abertis that I gave during the Q3 presentation. $4.4 billion EBITDA by 2025, but more importantly, post-SANEF in 2033, $4.4 billion-$4.7 billion as we are projecting today without further M&A and without further additional projects, which, of course, we are working on them. We are looking forward to continue to deliver.
From engineering and construction, as I said before, the important thing was to increase the percentage of low-risk projects above the 85%, which we have accomplished, and then making sure that they were also jumping into the growth areas where right now they have an order intake above the 55%. The more they get into those jobs, the more engineering, the more collaboration, and the better risk profile of projects they can achieve. A few numbers also looking at the recap from the Capital Markets Day in 2024, starting with revenues. We said we would achieve $48 billion by 2026. We are achieving by 2025. We said that we're going to achieve ordinary net profit by 2026 of $850 million-$1 billion. We are achieving in 2025. The net operating cash flow, we said that we would have from $1.1 billion-$1.3 billion average from 2024-2 026.
We are achieving in 2024 $1.5 billion, and we're comfortable and confident on the level we will achieve in 2024 and the level we will achieve in 2025. Our ACS backlog mix represents and shows our shifting to high-growth areas, as I just explained. No need to spend a lot of time in this figure. This represents the value that we have given our shareholders of more than 60% since our Capital Markets Day. The share in April 2024 was at $38.66. We are, as per yesterday, at $77.80, plus the EUR 3.50 per share that we gave. That brings a total of 60%. More importantly, we're looking into the future to continue delivering high returns. This is just a recap of what I just said.
Making sure that we have an end-to-end solutions provider role and solutions for the world in all the verticals, making sure that we invest in infrastructure, in assets that provide the sustainable EBITDAs, and then achieving that through three different levers. The first one, operational integration, very important. The one group, one team philosophy that I touched before. Second, talent. This is potentially one of the most challenging things in the world right now, how to attract talent, how to train talent, how to retain talent. I think this is one of our advantages at the same time. Why? Because the value of our brands, because of our granular presence everywhere in the world, in all those jurisdictions where we work, North America, we have presence in 47 states in the U.S.. We have presence in all provinces in Canada. We have presence in Europe.
We have assets in South America. We work in Asia-Pacific from India to Hong Kong, and we have significant presence in Australia and New Zealand, all the places that represent a lot of high-growth areas in the future. A very disciplined capital allocation, which, as you can imagine, will be an important focus today. Let's start with the first one, our role as a leading end-to-end provider. These are the different areas. Our core infra, our digital and tech strategy, today's focus, defense, energy, including nuclear and critical minerals. Let's touch a little bit in each one of them. We are already leaders in data centers, but we do have the capabilities, and we are working a lot to make sure that we become leaders in each one of those areas. This is the chart.
This is the timing of when we expect some of those areas to thrive. Data centers is clear sector that has been very strong in 2025 and 2024 results that the company has been able to perform in that area and will continue performing and will focus. We believe that defense is another area where we will be showing good performance in the following years, and we will be experiencing a big growth. The semiconductor space, I'll show you some of the examples that we're doing. There's a few big fabs subject to funding, subject to geopolitical decisions, but we are very well positioned on those. We'll have the critical minerals strategy outlined afterwards. I answered a question in the Q3 about it, and I will show a slide showing our experience.
A nuclear, which is a very long-term strategy, but we believe that the best way to approach it is starting today due to the complexity. Let's start with the core infra. The most important thing is the growth, sustainable growth in a market that we're seeing. We see that transportation and sustainable mobility will continue growing at 5% CAGR until 2030, 3% in the case of biopharma, healthcare, education, social, and sustainable infra, 3%, and 4% general building. Below, you will see the strong local capabilities and the backlog. This is a very important slide. It is very important because every time we talk about our main core business, people think about boring, traditional, and slow growth. However, 85% of any infrastructure globally is traditional core infrastructure. Every time you do data center, 85% is traditional. When you do a nuclear plant, 85% is traditional.
It's the walls, it's concrete, it's armored, it's civil works, right? The same thing applies to semiconductor fabs. The same thing applies to defense, to each one of the military bases, even for mining. This is key because if we are able to learn the 15% on engineering, on systems, etc., as we are, then we have the capability of the 85%, not just subcontracting, but performing and having the people to deliver. This is being key in our success in the high-growth areas. This slide represents the artificial intelligence ecosystem. The future is about AI and energy. This is AI, starting with the lower layer, which are the semiconductor fabs, the chips, the semiconductors.
I will show a slide now about our experience and what we're doing to position because it's very important if you really want to be someone in the AI chain, you need to try to be in as many layers as possible because they are all interrelated. Data centers, second layer, starting from the bottom. It's clear our experience, what we're showing. We are global leaders right now in the construction, and we want to be global leaders in the development. The third one, the cloud infra services. We're jumping into cloud services, cybersecurity, etc., through our platform in Germany, the Edge Data Centers, where we are providing the end-to-end solution for the colocation and serving clients as public authorities, defense, etc. The models, we are not positioning ourselves so far. That's where OpenAI is, that's where Gemini is, that's where Claude is, DeepSeek, etc.
That's not a layer that obviously is natural for us, and we're not looking to have a role in that. However, we see ourselves with a role in the applications when it comes to infrastructure of AI. The next part is robotics. We have agreements with different robotics and smart hardware companies to develop and influence the construction of the future. Starting with the semiconductors, we've been working in more than 20 top-tier projects so far. You will see some examples on the screen. I won't go through all of them, but we do have a pipeline of more than $15 billion that we are currently pursuing. This is what we have already done. Some of them are semiconductor fabs. Some of them are biopharma projects, in some cases with higher standards than the semiconductor fabs when it comes to clean rooms.
You will see on the left the market. We expect the market to reach $900 billion by 2029 at 6% annual growth. The market, when it comes to semiconductor fabs, is very unstable or uncertain because sometimes slowdowns and then accelerates very, very fast. So the investment is very predictable. What is not predictable is the timing. That is why so far we cannot project what is going to be the revenues moving forward in a clear way as we are doing with data centers. We will eventually, once we understand and once all those projects get the appropriate financing. The important thing is that we have experience. The important thing is that we are working on them. Look at the footprint that we are generating in the last years and the kind of projects that we are currently working globally. We move into data centers.
Obviously, a lot to say today, and we will spend a lot of time going through all of these. As a summary, first, engineering and contractor globally with more than 9 GW commissioned as per today in data centers. $14.3 billion currently in our backlog. We expect to develop 3 GW by 2030, 2033, but we do have, and that's as a developer, and we do have a pipeline of 11 GW. When it comes to Asia-Pacific, we're top five contractor, but again, we expect to be increasing that as we go. This is a very interesting slide because this shows what we're doing when it comes to artificial intelligence. I'm of the opinion that 10 years from now, only technology companies will exist. Technology companies doing healthcare, doing education, technology companies doing banking, technology companies doing construction. We are transitioning in becoming a technology company.
These are all the systems, all the applications that we are developing in-house in the end-to-end engineering construction world using AI capabilities, right? I won't explain. It's one of those names, but all of them are internal products. We're expecting to commercialize eventually most of them. We are embracing AI to the point that we're developing our own systems to perform our projects. How are we being able to achieve this? The most important thing when it comes to AI is to have the ecosystem. We have the largest ecosystem globally, not just in terms of geographical diversification, but also in terms of different verticals. We're using that ecosystem to develop the products. Of course, we're basing a lot of our products in third-party algorithms when it comes to generative and third-party algorithms when it comes to neuroanalytics.
What we are doing and focusing is the application itself, not the model for obvious reasons. Defense. We've been working in more than 60 projects in the last years, representing more than $8 billion executed. Some of the examples are on screen. You're aware of them. The military bases we're doing in Australia, the field logistics in Australia, the military bases in Eastern countries in Europe, some of the projects in Germany, Pearl Harbor, submarine, dry dock that we're currently working on, the missile transport bridge in Vandenberg, California. It is a market that we expect to grow 5% and to reach $1.7 trillion by 2035. Obviously, most of it is not addressable by us because we are currently focusing on what we call Defense 1.0, which is infrastructure, and we will see our opportunities when it comes to Defense 2.0.
There is an EUR 80 billion addressable market in infrastructure by us by 2030. Critical minerals. More than 140 critical mineral projects since 2022. Some of the examples on the screen. Most of them are very small in nature because they are pure engineering projects. How have we been able to achieve it for all the bolt-on acquisitions that we have been announcing for the last few years? Novapro, Prudentia, Pipar, Minsol, Mintrix, etc. As a result of those acquisitions, we have been able to grow in the engineering space. Some of these projects will be converted in EPCMs. Some of them will not. Some $1 million engineering project could potentially be turning into a $2.7 billion EPCM very soon. Our intention is to try to convert as many as possible.
By the same token, some of them will offer equity opportunities that we will be analyzing one by one without being the main part of our strategy. We will be analyzing at Oak individually, and we'll give some figures very, very soon. They are not gigantic because we want to be careful with this market. It's not our natural market. Yes, the infrastructure associated to it, the equity we will need to obviously analyze with the right level of care. Then nuclear. Let me go back to our experience in nuclear. Hochtief Nuclear built 13 nuclear plants from the 1950s to the 1970s. Since then, Hochtief has been maintaining and dismantling most of them. Since 2022, we started revitalizing Hochtief Nuclear, and we've been working more than 80 projects since then. Some of them decommissioning or dismantling.
As a consequence, the Sellafield project that we announced a few weeks ago came to us, $685 million revenues for Hochtief. There are other projects that have nothing to do with dismantling, that have to do with design of some of the components of the balance of plant or the reactors. We are growing significantly in those areas. Why? Because we want to generate capabilities in engineering when it comes to nuclear. We are qualifying again the entire group globally, nuclear, and we are including engineering capabilities, supply chain, modular capabilities. In the short term, this is very useful because it allows us to continue being successful in our data centers, especially the ones above 1 gigawatt that every client expects at some point to be connected or at least to have the possibility to be connected to nuclear.
But also when it comes to some of the big fabs, we're currently working, whether it's semiconductors, batteries, etc. It's not just a long-term strategy. It is part and allows to be successful in some of the things we're currently doing. Then comes our role as developers. Although I'm not going to start getting into a lot of details on the evaluations of data centers because that's a big part of the presentation today, and I will allow the team, and then I will come back for some closing remarks, this is a very important slide. Let's go through it. The first one is our transport strategy. That's basically managed lanes in North America. We have been pre-qualified in the 285 project in Georgia, the I-24 project in Tennessee, I-77 in North Carolina, and we're looking forward to the 495 project in Virginia. Five managed lanes.
We already won Georgia. We are assuming we can get from one to two projects. That means that we will inject, and these figures are comparing the 2024 capital markets day versus today. From now to 2030, we are thinking that we will inject $1.5 billion compared to the $1.2 billion-$1.8 billion we said in the CMD capital markets day. The $1.5 billion is a project and a half, more or less. That is because we are putting the threshold in 2030. If we were going to 2033, that obviously amount would be increased, right? As we will see later. The equity value that we are thinking, and we need to go to market consensus, and I am going to open a parenthesis. Every figure we are giving when it comes to valuation of managed lanes or data centers are market-tested figures. It is market consensus.
We are taking the lower end of the latest market transactions. Every time we speak about a managed lane post-construction, every time we speak about a data center when it is ready to build, a data center when the lease has been signed, a data center when it is full in operations, we are taking the ratios from the market. What we know is what we are going to be giving the market at any given time. We are applying the market figures to those valuations, right? We are not coming up with any number. When it comes to the valuation in the case of managed lanes, we know that the lower end of the market nowadays for a project that is mature post-construction is about six times equity. In some cases in the past, we have achieved up to 10.
Such is the case of 288 or some other projects from the competition. Anything below or anything before the market or the asset gets mature, then we need to start playing with discount rates, and it's a little more subjective. That's why we're going to be showing the 2030 threshold and the 2033 because that represents when the market, our assets are mature versus our approach to them. We're expecting an equity value by 2033.6. Obviously, if we're injecting $3 billion to $4 billion, $3 billion multiplied by 6 would be 18, 4 would be 24. That would be once the assets are mature. That won't happen in 2030. It will happen more probably by 2032, 2033. We move into digital tech. We're expecting in the capital markets day, we said that we're going to inject $1 billion to $2 billion.
We're expecting to inject $2.2 billion from now to 2030. We said that the valuation of equity would be $3 billion to $5 billion. We're expecting $11.5 billion equity value by 2030 to be increased afterwards, as we will show. The edge data center was not included in our capital markets day, but we know that we're expecting to inject $200 million with an equity value by 2030 of $1.4 billion. Then we have our energy projects. The energy projects, because we have been focusing on data centers and transport, we have decreased the amounts that we were thinking to invest back in the capital markets day. In the case of the energy demand, we've gone from $1 billion-$1.5 billion all the way down to $300 million. Therefore, the equity value that we're expecting on that has decreased as well.
When it comes to next generation mobility, we keep the same numbers. They are going very well. SkyPorts has been a very good investment. As you have been hearing the news, we got not just the Emirates VertiPorts, but also the ones in France, the U.K., New York. It is growing significantly. We continue being optimistic on the strategy. Critical minerals is very early stage. That is what we said back in the capital markets day. We keep saying that right now. We need more visibility. We need to understand our capabilities and where we are going in the future. The total, keep this number because we are going to discuss a lot during this capital markets day or this investor day on how we value the $18 billion of our assets by 2030.
At the same time, we will be talking on the valuation of our underlying revenue and EBITDA business associated to our strategy. I spoke before about the three levers, but I would like to right now turn over to our Chief Financial Officer, Emilio Grande, so he will take you through our financial position and the capital allocation strategy.
Thank you very much, Juan. Thank you very much, everyone, for joining us today. I'm going to, before I hand over to our DC data center business colleagues to explain the most interesting part of the or the most focused part of the investor day, let me just give you a quick financial update of where we're sitting and a bit of a roadmap with particular focus on the capital allocation and our investment strategy, which Juan has outlined already at a high level.
Let me get in a little bit more detail. I think the first big message from my side and the more important one is we're sitting in an excellent position to move forward with all the plans and opportunities we've got ahead of us. I look at it from two perspectives. First, from a cash flow generation, Juan has touched on this, but if you look at the growth on our net operating cash flow over the period since last time we met in the capital markets day, early 2024, you can see that 23% CAGR growth adjusted for working capital. I'll touch on that in a minute, right?
This is sustainable long term, and it will get higher, as Juan has indicated, but this is sustainable growth of cash flow because it's based on top-line growth and margin expansion, which, by the way, we expect to continue in the future, right? I'll touch on how we on the firepower based on this and the critical messages, but the focus is to continue to grow this in terms of top-line growth and manage working capital, which goes to the risk and business mix profile that Juan has touched on. Our current balance sheet position, that's another asset at this point that puts us in an excellent position to move forward from a financial perspective and deliver on all our ambitions in terms of investment and growth.
This is the numbers we just released for Q3, $2.2 billion of debt in the balance sheet with a $3 billion EBITDA last 12 months. That's 0.7 leverage. Obviously, we're going to improve these numbers by year-end. As you know, we've got seasonality in Q4. We've got some financial transactions going on, so we will improve these numbers. That's going to put us in an excellent position from December, which is our starting point for all the capital allocation numbers I'm going to provide in a minute. In terms of roadmap ahead from a financial perspective, obviously, focus continues to be on the net operating cash flow generation going forward, focus on the fundamentals in terms of working capital management as well, but promote growth and margin expansion and do the right management of our cash at a project level and throughout the organization.
That's a key part of our financial roadmap, as it has always been, but obviously, we need to continue to deliver and focus on that strongly. Operational integration, as Juan has touched on, is very relevant. I'm going to focus more on the efficiencies plans. We have already delivered to date $70 million annualized savings realized to date. That's a combination of Dragados, Ratero Dragados, Scenic, and other parts of the group. This is just the tip of the iceberg. We are working on a broader plan across the group, and we will provide further update on that. I mean, it will be significant. This is just a report of what we've done to date, which we expect to show in the P&L and in the cash flow generation as well in the coming years.
I'll jump now to, obviously, the more important point in terms of capital allocation and how we plan to address that. Let me touch first very quickly on some more detailed numbers on what Juan has outlined. In the capital markets day in 2024, we talked about investing between $3.5 billion and $5.5 billion into greenfield infrastructure, right? What have we done to date? We've already invested $700 million between 2024 and 2025 to date, and we are saying we're going to continue to invest $4.5 billion. That puts us in an overall number of $5.2 billion in the period from 2024 to 2030 and $4.5 billion from, say, January 2026 to the end of 2030, which Juan has provided already a breakdown and we'll hear more about in the data center space.
In terms of value accretive M&A, you have been following the announcements, but we have spent around just over $700 million as well in terms of different acquisitions, mainly in the mech and elec space, engineering, delivery capacity to expand our capabilities in Europe, for example, through Dornan or through other companies, Maverick, Schliessman, etc., and is promoting growth, obviously, in the business. The acquisition in the several bolt-on acquisitions in the critical mineral space and the acquisition of these, of which we still have a 40% remaining coming back to us soon, right? This is above $700 million. All the companies follow the same strategic direction we provided in the CMD. Highly synergetic. They are all performing. They are all growing and providing growth to the overall group once inserted in the group ecosystem.
In terms of brownfield acquisitions, $850 million invested and committed into Abertis in the period, which we will be disbursing for the A63 in the last quarter of the year, the $200 million. This is the picture of what we said, what we've done, and what we're looking to do going forward. Let's look now on how we are going to deliver the $4.5 billion investment in greenfield. You can see on the left-hand side of the screen. M&A and other brownfield investments. The key message here, I think Juan has advanced it, but here you can see the numbers.
Given the extraordinary growth we are experiencing, and I showed that graph before where we had at nine months $1.6 already, if we assume a generation of just over $1.5 billion of net operating cash flow per year and we deduct shareholder remuneration, that means $900 million per annum, which gives us $4.5 to invest, right? Okay, let me keep going then before we hopefully not lose it again, but just in case. I was talking about the $0.9 we're expecting to generate after shareholder remuneration per annum, which gives us ample firepower to deal with the investment plan we've got ahead of us. When we did the capital markets day a year and a half ago, we were expecting around $600. We've got additional $300 per annum to invest based on current performance of the business. That $1.5 is obviously an average for the period.
We expect growth. It is a reasonable conservative number for the period. Juan has touched on the Q3 presentation on the divestments in the Q&A session we have already done to date. Obviously, we still have potential for further divestments that would add to this firepower position for our plans. The bottom line of this is we have got $5.5 billion-$6 billion of firepower without increasing our leverage position and without monetizing the new investments. This is going to give us headroom in the rating position going forward because, obviously, maintaining the debt levels with the expected increase in cash flow generation, FFOs, and EBITDAs, that will give us headroom. We are not planning to use it. It is just an additional firepower lever that has always been there, and we will accumulate more.
Also, we've got the possibility to monetize assets, as we said as well in the capital markets day, which we will see through the data center session. That's, in particular, an asset with a high turnover and high value and liquidity position in the short term. We will see how that accumulates very quickly. We are sitting in a very comfortable position, highly visible operational cash flow by the business to invest in all these initiatives through a capital allocation process, as we've been doing to date, where, as I said before, we've done around $2.3 billion since last capital markets day between the $700 million in infra, the $700 million M&A, and the $850 million in Abertis. We are sitting in a comfortable balance sheet position. We've been balancing out to date, and we plan to continue to do that going forward through our capital allocation process.
With this, I'll hand over to the data center colleagues. Thank you very much, everyone.
We are in the midst of a generational transformation, one driven by the extraordinary demand for data center capacity. Across the ACS Group, we are uniquely positioned to meet that demand. Together, we deliver integrated end-to-end solutions at scale, with speed, and with certainty. In North America, Turner is a builder of choice for the world's leading hyperscalers. In Europe, the pace is accelerating, and in Asia-Pacific, investment is growing as companies rapidly scale their digital infrastructure in the region. Across every region, the story is the same. Clients turn to us to build faster, smarter, and at greater scale. They choose us because we deliver. We bring deep experience and technical strength to every phase of the data center lifecycle.
Driving this growth is a race to build the infrastructure that will power the next generation of artificial intelligence. The projects beginning today will bring critical capacity online by 2026. That capacity will be absorbed and will fuel the next wave of digital innovation. To meet this demand, we are scaling up. We're growing our teams, expanding our global reach, and strengthening our engineering and technical capabilities. This will enable us to continue to deliver the most complex facilities with precision and speed. This is a time of extraordinary opportunity, and together, across the ACS Group, we are ready to deliver globally.
Good afternoon. Peter mentioned a couple of things that really should resonate with all of us today. He mentioned extraordinary opportunity, extraordinary demand, and he mentioned a generational transformation that's happening.
There hasn't been a revolution like this in the world since the Industrial Revolution in the early 1900s. I would argue that even that can't compare to what's gone on, what is going on, and what will come. This is still in its infancy. The digital revolution is just starting. He talked about extraordinary demand. Does anyone imagine that the demand for data capacity will go down in the next 10 years? You can't imagine it. You can't imagine living without your phone, without your computer. It won't go down. The opportunity, and as Peter said, it's an extraordinary opportunity, is how we meet that challenge. I think the ACS Group, I know the ACS Group is uniquely positioned to meet that extraordinary challenge as we move through. What is driving the industry? What are the trends?
What do we see as a dominant builder of data centers moving to development? What do we see out there? We see the AI growth. That's the easy one, right? You hear about it every day. Think of this: $3 trillion-$4 trillion of AI infrastructure investment by 2030. $3.4 trillion. Data sovereignty, as the data volume grows, so does our need and our desire to protect that data. It affects the market. It's a trend we see every day. Accelerated construction techniques. Ten years ago, five years ago, we were expected to build a 50 MW data center in 18 months-22 months. Today, that expectation, the same data center will be complete in 9-11, and it's not fast enough yet. It has to be 6-7-8. You see the advent of rapid deployment techniques.
In Des Moines, Iowa, last year, we built a 38-MW data center in three months in a temporary structure. Rapid deployment will really drive the market from an accelerated construction standpoint over the next millennium, probably. The other thing that's accelerating construction is the use of modularization in the construction and engineering of a data center. We now design data centers to do the modular. We used to design a data center and build a data center and then figure out what we could modularize. DFMA techniques tell us, "Design it for modularization upfront," and then the building becomes faster and it becomes easier. We're making great strides in developing that modularization. Strong public and private investment, capital from all sources. Public is getting involved. The recent announcement of the five EU Gigafactories to support AI across the European Union.
The public funding and capital influx in the United States and in Asia-Pacific. Finally, the private equity, as we'll talk about today, involved in grabbing a share of the market and bringing that capital influx to it. Lastly, hyperscalers self-fund. These are trillion-dollar companies. They really do not need an influx of capital. They bring the capital to the table. Last trend we see is the rapid increase of edge computing, and Byrne will talk about that later. Through our product, the XCO that we've developed and are constructing and executing through Hochtief in Germany, this brings computing power to the edge, to the end user. If you're going to have smart mobility in cities, you need computing power at the source. This market alone is projected to increase by 20% annually over the next five years. Those are the trends.
Let's dig into it a little bit more. What are the numbers? I'm not going to read everything, right? But 16% annual growth in our data centers, demand for data capacity over the next five years annually is a pretty big number. 16% growth in any market is pretty big. Data center demand will continue to grow. And three factors drive demand. First and foremost, AI, of course. 15 x increase in data center demand due to AI and generational AI. Cloud migration. Everyone thinks the cloud has been, we're all talking about AI. That's the past. It's really probably the fastest growing sector. 20% of the data value now is in the cloud. By 2030, 70% of the data value will be in the cloud. A greater proportion of our data center work today is still building cloud computing facilities.
Probably 65% of it is still building data or cloud facilities to handle that volume. Lastly, just the sheer amount of volume, sheer amount of data that is demanded and consumed and implemented is driving the market. We have three strategic pillars in the ACS Group. First and foremost, it started with a great foundation, and that is on the engineering and construction side. Dominant position globally as a world leader in data center construction. We are parlaying that expertise and that knowledge into the investment and development side for data centers. The third pillar is being a cloud provider through providers like our Urizon product linked to the XCO edge computing in Germany. Three pillars, three ways to get there, and we will talk about all of them today. Our addressable market, the strategy enables us to do this.
Right now, we can address, as builders and engineers and constructors, 100% of the market. We can build edge. We can build hyperscale. We can build colo. We can address 100% of the market. Our move into investment and development of data centers allows us to address 60% of that market. Our target is to build for hyperscalers, hyperscalers, and even second-tier hyperscaler clients through giant colos. We can address another 60% of the market. It is not additive. We are not going to do 190% of a market, but it is 60% more addressable and available for us through the investment and development. Finally, we can address 30% of the market by being a cloud GPU provider. We are going to move through the different pillars. Jim Brownrigg is going to take us through the EPM side.
Thank you, Mike.
Yes, and Mike mentioned, if we stop and look at the pillars that are part of our strategy, looking first at engineering and construction, right? Engineering and construction, if we look at the available market, right, looking across the regions that we work in, first in the United States or in Americas, the market is going to grow from 42 GW- 137 GW, over 100%. Again, in Europe, again, over 100%, right? And again, in APAC, over 100%, right? Growing by that much IT capacity, which is easily transferable into investment in infrastructure and into the data center, right?
More importantly, if you look at some of the hot markets or the markets that are growing and expanding quickly, like Ohio, Virginia, and Texas, those are markets that we work in today that we have expertise and capabilities, that we have resources, that we have relationships in the supply chain, that we have relationships with the local governments and authorities having jurisdiction, right, that we can deploy readily and easily in those hot markets. Similarly, in Europe, we're similarly active and working in Spain, in the Nordics, and in Eastern Europe. Again, the same items. In the markets, with the expertise, with the resources, with the experience in the supply chain, and with the experience with the local authorities having jurisdiction.
Then again, in APAC, very similar again, growing 100%, and our late in Asia group being active in Malaysia, Thailand, and Indonesia, having again, resources, experience, expertise, relationships in the supply chain, and relationships with authorities having jurisdiction. Now, if we take a look at each of those individual regions, if you will, or those individual markets, let's talk about the Americas first, right? If you take a look at what we're doing today in 2025, we'll do about $8 billion in revenue that will grow to well over $20 billion by 2030, right? A significant growth rate capturing all of the addressable market that Mike spoke about, right? All the addressable market, those 100%, if you will, of those markets that we pursue, right? If you look at our backlog today in the Americas, it's about $13 billion.
By the end of this year, that backlog will be $17 billion. Backlog is growing significantly year over year, right? We're the number one builder of data centers in the Americas. We take a look at Europe, right? Expanding in Europe. Again, a market that we're active in, a market that we're growing in all the addressable markets. If we look at our expansion in 2025, obviously with the acquisition of Dornan and their work in the data center market, our expansion of Turner into Europe, and of course, working with Dragados and Hochtief, really growing that market from $1 billion in revenue this year to well over, or nearly $3 billion by 2030, right? We have a pipeline of over $13 billion worth of work in Europe, and again, working at all of the addressable markets. Now looking at APAC, right?
$3 billion of work completed or under development in data centers, a pipeline of $16.5 billion, and a top five contractor, again, working at all of the addressable markets, again, with a growth rate that will get us to well over $2 billion by 2030. If we look at our end-to-end, or excuse me, our engineering and construction services, it really goes chronologically through the work, right? It's design and engineering. It's supply chain and operations. It's site and civil and execution, core and assembly and execution, modular construction, which Mike spoke about, and I'll go a little bit deeper momentarily, and MEP assembly and commissioning, right? Where we have experience today and we're working. More importantly, where are we going, right? Where are we going? Where are we expanding?
We're expanding because we've reinforced our ability in design and engineering with the acquisitions of Dornan and Fleischmann and Maverick, right? We're reinforcing our ability in supply chain by expanding SourceBlue globally to serve all of the areas in which we work, right? What are we doing in execution? We're reinforcing and expanding our expertise, partnering across the company firms to work on projects in Europe, partnering across the company firms to work in APAC. In modular construction, really deepening not only our expertise, but our ability to deliver, our ability to deliver faster, our ability to deliver more effectively, and ability to deliver where resources are more limited.
In MEP assembly and commissioning, again, with our acquisition of Dornan and Fleischmann, our ability to have that deep expertise and really a captive set of expertise in mechanical electrical, which is critical in the data center market. Let's talk a little bit about why is modular? Why are we talking about modular? Why are we talking about DFMA? Right? What are the challenges we face today in the data center market? Number one, speed to market. Getting to market faster is revenue faster, right? As a technology company, serving your clients faster. Speed to market, number one. Number two, resources. Skilled trade, labor, continuing to be a challenge in most markets. How do we deal with that challenge? How do we deal with the lack of resources? We need to do it through modular. We need to do it through prefabrication.
Modular design, that means DFMA, design for manufacturing and assembly. That means designing around modular, not taking a design and trying to modularize it, but designing around modular. Designing around modular, productizing a kit of parts or assemblies that come together. Vertically integrating that supply chain becomes incredibly important. You have to vertically integrate your planning of components and subcomponents and assemblies to arrive on the job site at the right time or just in time in order to assemble that building. Vertical integration of the supply chain. Again, last-mile approach to both logistics and to client needs. Our client needs are evolving very quickly, right? They're evolving into expanded water-cooled or expanded hybrid water and air-cooled. They're expanding into different types of floor loading, right, to accommodate bigger racks or more dense racks, requiring more fiber entries, right? How are we going?
How do we adapt to that in modular? We leave that last mile out, if you will, of configuring the data hall around actually how they're going to use it, what servers they're going to put in, what racks they're going to put in, and how they're going to cool it. A traditional data center. If you think about building a traditional data center, if you start with a pad, right? Earthwork is done. You have a pad or a pad built, if you will. You got to do the foundations, right? We got to pour the slab on grade. We got to go vertical with structure, right? We had to go vertical with enclosing the building, getting the roof on. We had to go inside and finish out the interiors, put the partitions in, move the equipment in, right?
Make all the MEP connections, start up the equipment, commission the equipment, right? A very end-to-end sequential process because we're stick building the data center. What does the future look like? The future looks like a modular data center. Let me show you an example, right? We start with this kit of parts. A kit of parts or a kit of modules that will come together at the job site with the generator to begin with. Next, with the fan wall and HVAC equipment, right? Most data centers still today use a fan wall, but now we modularize that entire fan wall and that gallery. The CDU equipment that's coming into the data center to feed the water-cooled servers. The internal plenum module, right? The plenum module, the hot oil containment that's going up into the plenum.
The integrated mechanical and electrical that's coming in to feed everything above the plenum, right? That's getting everything to the racks or to the servers, if you will. Of course, then the electrical equipment modules, right? These are modules that we're putting the switchgear in, the UPS is in, the STS is in, if you will. As we finish getting in the electrical modules, we then come in with the hot oil containment that's going to sit underneath the plenum. That hot oil containment today comes already with all the busway or bus bar, if you will, on it. Already comes with the lighting on it. It comes with a cable tray or the cable rack on it. Comes with the structure to support it. Comes with the hot oil containment fully enclosed. Modular air-cooled chiller platform. No longer are we just lifting a chiller to the roof.
We're lifting the chiller plus the access platform plus the connections so that we can quickly connect it and commission it. If you look at a traditional data center that Mike spoke about or that I spoke about earlier with being stick built, if you will, and Mike talking about typically from pad ready, you know, if we were really good, we'd get it done in 10 months. If we're a little slower, maybe 13 months or 14 months in the U.S. If you look at it today, we're able to deliver this in the six to nine-month timeframe, right? Now we have to build out our ability to design, our ability to vertically integrate the supply chain, integrate design with construction, integrate our ability to commission and install this modular. That's where we're going. To some degree, we're very much already there.
What does this all mean, right? By 2030, right, we expect to be doing $20 billion-$25 billion in engineering and construction data center revenue, right? You saw the projections, low-end $20 billion should be easily $25 billion. An EBITDA margin of 5%-6% means we increase the EBITDA by 2030 in data centers to $1.2 billion-$1.3 billion. What does that equate to? It equates to an ACS equity valuation of $13 billion-$15 billion, right? We grow the revenue at good margins, right, to create EBITDA, and then, of course, translates to equity value of ACS in 2030.
Thank you, Jim. Good morning, everyone. One strategy, three pillars. We're going to focus on the second one, which is our strategy for large-scale colocation data centers for hyperscalers.
We started with our strategy more than two years ago when we realized that in the face of the amount of data center capacity that was required by the market, there was an enormous opportunity for the ACS Group to become a data center developer. If only we could align and activate all the experience within the ACS Group in developing, investing, and building infrastructure. If we could capitalize on pre-existing relationships with hyperscalers through our construction business, and if we could take advantage of our local presence in key strategic markets for data center development, we should become the developer that builds and develops faster and better than anyone else. In proving that to hyperscalers, our vision is we will become the partner of choice to deliver the critical physical infrastructure that they require to meet their own business targets.
Ours is a very unique value proposition in the market. As of today, there is not any other development platform of this scale with such global presence, and that is industrial in nature. I think it has been clear through the presentation from Jim that engineering, construction, and supply chain management are within our DNA, and that makes us unique. That sets us apart from any other competitor in the market. We have announced this morning an incredible partnership with GIP BlackRock, who is going to be joining us in delivering this vision. What is our vision? Our platform is a standalone organization where we have assembled all the knowledge and all the skill set to deliver end-to-end services to our future customers. These skill sets and this knowledge would allow us to move and transform our investments in land and energy and transform them into fully stabilized and operational data centers.
The experience and the skill set includes land development, energy development, engineering, construction, project finance, operational capabilities, etc., etc. When required, a data center platform will reach out to different companies of the ACS Group around the world to complement these services. Growth is very important to us, and we have set a target to have a portfolio for ACS of 3 GW of data center capacity, utility power, 3 GW by 2030. 1.7 GW of that data center capacity is already part of our portfolio, and it's under active development and in some instances under construction. To complement the different to three, ACS is monitoring the market and is currently assessing opportunities of up to 11 gigawatts in the different geographies where we're focusing: Europe, North America, APAC, and we're also contemplating opportunities in Chile.
For the 1.7 GW, we will be counting on the partnership for GIP, and this initial portfolio is constituted by seven sites. We focused primarily on tier one markets adjacent to tier one locations and in some of the fastest growing markets around the world. We have two assets in the US in the Fort Worth-Dallas area, Ruterglem, Virginia, Melbourne, Australia, and four assets in Spain, two in the region of Madrid, two in Zaragoza. Texas and Virginia, forgot to mention, are some of the most important data center clusters in the US. I'm going to let Jim give us an update of where we are in development in all of our assets of the initial portfolio.
Great. Thanks, Vicente. Looking at each of these individual assets and kind of where we are and where we are with capacity. Starting with Madrid, Madrid One, in Alcalá, right?
Our first data center, we're well underway. I'll show you a few pictures here shortly. It'll finish in three phases with 140 megawatts of utility. It's in construction, right? We'll be in COD by the end of 2026, right? COD basically means commercially available to the customer and getting ready for RFS, right? The next one that'll come online is in Texas, in DFW's area. DFW, which Vicente mentioned, is a very hot area, right? Market that's expanding, a market that clients are coming to us for. Hyperscalers are coming to us asking us for capacity, if you will, in Dallas-Fort Worth, right? 335 MW. We've already started early works. Virginia, Ruterglem, if you will, just in that kind of just south of Northern Virginia, the biggest data center market in the world, 300 MW. We're just starting early works there.
Basically, we're pushing dirt this week, if you will. In Melbourne, right? 240 MW. We're just starting early works as well with the earthwork, if you will, and building the pad for the utility substation. Next, excuse me, Madrid Two, which will come online in 2028, again with 137 MW, which is an expansion of just about 11 km-12 km up the road from Alcalá, which is our first data center campus, if you will. Next is Zaragoza One, right? Any of you that have been reading about the data center market in Zaragoza, huge expansion. Expansion by a couple of the hyperscalers. They're making significant investments in Zaragoza. We have 300 MW there. We're in design and zoning ongoing. Our team's done a fantastic job to get through zoning very, very quickly. A fantastic achievement. Again, coming up with COD in 2028.
Lastly, Zaragoza Two, really a land bank with access to power, 400 MW, where urban planning is ongoing. What's really important to this is that power access is secured in most of these sites. We still have a little bit of work to do, but power is what's important, right? When we sit with a tenant or a client, it's about power, it's about land, and then it's about our ability to integrate and deliver in a very fast fashion, a low-risk and integrated form. Madrid One, I mentioned Alcalá, right? Where are we at today? Some of our team members that are here today walked to the project this morning. The building's in, has all the undergrounds in, the building's enclosed. We've started, interior partitions are in, mechanical electrical systems or conveyance, if you will, of those systems are in, and now we're moving equipment in.
We're moving electrical equipment in, we're moving generators in, we'll start moving fan walls in, right? We're getting very close, and we're going to hit a COD, if you will, by fourth quarter of 2026. This is what our data center will look like with roughly 15 MW of IT ready and available to our tenants to move in. Also looking at leveraging our integrated approach, which I mentioned earlier. What differentiates us in the market? What differentiates us is that we're able to provide an integrated offering. Everything from site selection, access to power, access to design and construction, supply chain, commissioning, and turnover. We leverage all of the ACS companies that have expertise to do that. We can integrate, we can start sooner, we can go faster, we can make better decisions, right? Supply chain and procurement I mentioned.
Our ability to leverage the SourceBlue supply chain for a global OFCI program, leveraging that spend across the 1.7 GW Vicente mentioned, leveraging that spend to not only get the best pricing in the market, but to really control the supply chain. Critical path on these jobs is largely the M&E equipment. Leveraging that volume, if you will, or that aggregation gives us preferred manufacturing slots. Leveraging our construction expertise. I mentioned earlier that we're active in all these markets, right? We have resources, we have data center expertise, we have supply chain relationships, relationships with the local government and understanding how to get through the permitting process, which is really critical on this. We also are able to leverage GMP collaborative open book contracts. What does that mean? That means it allows us to mitigate and manage the risk from end to end.
Being able to integrate at design, at supply chain, at construction, at commissioning, we're able to deliver it in a mitigated risk fashion because we're collaborating, and we're collaborating day one as we start the project. Lastly, operations and maintenance, right? Many of our companies have worked across operations and maintenance, leveraging that expertise, whether that be Clece, whether that be Dornan, whether that be Turner, right? Whether that be UGL, how do we leverage that expertise, if you will, to enhance our operational and maintenance ability?
Thank you so much, Jim. All of our sites that are able to deliver data center capacity for customers by 2026 and 2027 are already under active commercialization. In fact, we're targeting to sign our first lease no later than the first half of 2026.
Having the construction capabilities sitting at the table when we negotiate with hyperscalers has been instrumental to get traction within our commercialization efforts. Why? Because that provides our customers certainty, credibility, and we are able to react faster than anyone else to the first signals of commercialization and ensure that the construction resources are available to us and are mobilized on time where other competitors may have to start reaching to third parties to get those resources ready. In terms of commercial terms, we're focusing on leases of more than 10 years. We're contemplating two types of leases, a full-collar lease that includes all the operational services.
Important to note that what we do is operations and maintenance of the physical infrastructure, facilities management, but it does not include any maintenance services of the hardware or the IT component of the data center because that belongs to our clients and those are inside of the data halls that we rent to them. We are also contemplating triple-net leases where all those operational services of the physical infrastructure are performed by the clients themselves. When operational services are included, these contracts are based on service-level agreements that define very high levels of performance standards that include high levels of redundancy as well as response times whenever there are incidents in the facility to minimize the impact on the operations of our clients.
Data centers have very strong barriers to exit for our future customers because for every single dollar that we invest in physical infrastructure, our client invests two, three, or more in investments of the hardware and IT component of the data centers. That in itself creates a natural barrier to exit for our clients. Together with the comfort that we take from operational experience, we're getting very comfortable negotiating the clauses around early termination in these contracts. One of the things that we are more proud about is the enormous amount of talent that we recruited from the industry. We currently—sorry, can I just continue, please? Yeah, you bet.
We have been able to effectively go to the market to the organizations and companies that work in each of those elements, in design and construction, excuse me, in design or in engineering, in construction, in operations, in equipment manufacturers, if you will, other colo developers, if you will, right? This is a smattering of the companies that we have recruited our team from, a fantastic group of people with deep expertise in each of their sectors, if you will, each in their disciplines, but equally important expertise that brings a broad group of relationships, relationships with manufacturers, relationships with operators, relationships with hyperscalers, our tenants. I am very proud of the group of people we have assembled from the industry to help deliver on this mission of being a developer and data center operator. Looking at an overview of how do we shape our platform, right?
As Vicente had mentioned—
I can take these. Okay. Thank you so much. Apologies for a lot of things are a sugar-like issue, but I'll try to continue. A few words about our collaboration with GIP BlackRock. First of all, it will be a joint venture, 50/50, co-controlled by both partners, where the initial portfolio of 1.7 GW will be transferred. In terms of the platform structure, there will be a distinct entity that will be dealing with investment and another one that will be dealing with the services. The investments will be done through project-specific entities that will allow us to raise project finance, standalone project finance, separated one project from the other. Separating investment from services will allow us to continue providing services to our clients, even in those instances where we would contemplate opportunities for equity recycling, the investing from stabilized assets.
Finally, the original portfolio will be transferred for our original initial consideration of $1 billion. ACS will receive also a number of payments in total, EUR 1.2 billion, as we hit different commercial and business targets that include EUR 200 million in the case that new projects are added into our platform. This implies a total valuation, sorry, Mike, of our platform, initial portfolio of EUR 2.2 billion, contemplating both the initial consideration and the earnouts. How do we create value through the lifecycle of our assets? There are fundamentally three important milestones for value creation as we derisk the project from land and power into operational assets. The first opportunity is at ready to build. What happens at ready to build?
We have managed to unlock everything that needs to be sorted for us to be able to develop a data center in one of our sites, which basically means having secured fully the power and having the zoning for data center activity in that site. We feel very comfortable with this stage, and in fact, we've managed to prove in Australia, in Spain, and the US that we are able to move through this stage faster than anyone else. Commercialization is the next big milestone in value creation. That's where we sign a lease for our customers. We are able to raise finance against the future revenues, and we're able to start vertical construction of our data centers.
Up to this point, we would have had to invest in everything that is required for us to come to a window of opportunity of 18months-24 months to ensure that the data center is delivered to our clients in this timeframe. This timeframe, as Jim has explained before, is becoming shorter and shorter, and therefore, having the construction capabilities, as well as the modular solution for the rapid deployment of infrastructure, is becoming more and more critical. Finally, as the data centers become operational, there is the greatest opportunity for value creation. At current market multiples based on recent transactions, our portfolio could reach valuation of $20 million-$25 million per megawatt IT of enterprise value and $12 million-$15 million per megawatt IT for equity value.
Based on these multiples, we have applied these multiples both to our initial portfolio of 1.7 GW and the additional projects that ACS will be incorporating to reach the 3 GW by 2030. Applying the multiples that we've seen in the previous slides, we are proving that we set to reach a total equity valuation of $11 billion by 2030 and $14 billion by 2033. 2033 is when we expect that the 3 gigawatts will be fully operational. We have run a similar analysis based on DCF to prove that both analysis converge and that the valuation both for 2033 and 2030 is similar in both cases.
In order to achieve this, we will have to inject net equity injections of $2.2 billion, which we have calculated netting off the gross equity needs for the entire portfolio of almost $5 billion, with netted off the money from the initial contribution of our initial portfolio, the earnouts, the equity already injected by the ACS Group, as well as distributions and opportunities for early monetization of some of the additional assets that will be included into the portfolio. We do not need to wait until 2030. What we are trying to prove in this slide is that there is going to be a ramp-up between 2025 and 2030, and that by 2028, our equity value would have already reached at least $7 billion. This happens as our projects, both in the initial portfolio and the additional projects, continue to mature, more leases are signed, and more gigawatts become operational.
Important to note that the numbers in these slides reflect ACS' 50% stake in all the projects in the portfolio, and that is based on ACS' own estimates. Just to recap, there are a number of factors that we wanted to highlight for you as we finish this part of the presentation. First of all, signals of growing demand in the market that data center capacity will be required by hyperscalers and other AI players in the next years and decades. Secondly, our unique value proposition to the market, our industrial nature, and the backup from all the companies of the ACS Group. Third, our de-risk approach to investments. We continue to focus on well-established markets where the data center capacity is fungible and with projects that have largely secured the power in order to enable this data center capacity.
That, in addition to the new partnership announced with GIP BlackRock, set up strong foundations to be successful and to deliver the equity value projections that we presented to you today.
Thank you so much. Coming back to our vision to be a reference player in the data center market and delivering end-to-end services, solutions to our clients, I want to now, after we first heard about our great capabilities in engineering and construction, being a leading EPCM contractor, and in addition, we heard about the large data center market, I want to take you now with me to the edge, where another growth area is. The growth is driven by low-latency need. The low-latency need develops further applications. We allow our clients to bring new tech into the industry, like, for example, IoT, for example, edge AI, or in the future, robotics.
At this stage, low-latency matters, and it's not just low-latency, it's as well the client wants us to deliver a sovereign, a very resilient solution, which is totally different to other applications. Having said this, we want to deliver these services as well in the course of defense, in the course of banking to government that needs sovereign solutions. In addition, when we have here the cloud services we want to provide, we can reach the end users, positioning ACS as a player in the cloud services market and as well in the AI inference market, and later on, someone who enables robotics and other services. How did we do this? We thought it's a need to develop an own product. That's where we started to look at the layers of a data center value chain you see here.
We started really from scratch like four, five, six years ago, looking at what is the need of the end user. The end user is interested in compute, the end user is interested in storage, and nowadays, he is interested in having an end-to-end solution in AI. We started to look at the value chain, and we saw we are already placed in the lower layer, one to three, where normal PPPs take place and where we do lifecycle optimization. We start in the infrastructure layer, looking for sites, permits. We enable fiber and grid. We do the construction and operation and maintenance, and we secure power, cooling, and sensors in order to operate such a data center. Having this in mind, we always optimize the lifecycle in order to find efficiency gains here. It was a natural step to look into the other layers.
In the platform layer, where server network distribution is relevant, and at the infrastructure layer, where virtual services and hosting is relevant. If you want to really do a lifecycle optimization and look at the whole thing and vertically integrate, you find a lot of efficiencies. That is where we started to develop a very sustainable reference, sustainable data center product for the edge. The data center itself is a direct liquid cooling now with a closed water system. It finally ended up as a data center, not just having a facade and a green facade from the outside. It is fully built in laminated timber, and it is capable of capturing CO2 already in the construction phase. It is not just about that. With the direct liquid cooling and the all-over optimization, we are now able to have a top-notch energy efficiency we can bring here.
You have to imagine we come along with the power usage effectiveness of 1.1, which you have to think about like that. If you want to have one power for a computer, you need 1.1. You need the 10% more energy to bring this compute power into action. The reference on the market currently, or the average, is at 1.4-1.5, which means others have to pay three, four times the energy in order to bring the same compute into action. As a natural step, we built, as well as a commercialization layer, Urizon Cloud Elevators that deliver the service to the end user. Which end user does want to bother about a site, a permit, about who does the construction, the operation, and maintenance, and who does the all-over facility management here? What is the right network, and what are the right service?
That is what we brought together and optimized to this cutting-edge solution. With that came a faster time to deployment as we built this modular approach with two, five, and 10-MW blocks that are set up in a manner like a factory. We can easily, quickly fulfill demand on the market with our own supply. With that, we enable, and the cloud layer is something where we position ourselves. We enable clients quickly to move on and develop themselves in the market. This is supported by a pipeline we established with first financing. When we started, we had secured a framework with five data centers in Germany, but we already, during the first data center execution, extended up to 15 and now up to 25 data centers all over Europe with a partner, Palladio, that is a 50% partner into our data centers.
We start in Germany and extend our operations into Austria, Switzerland, and the Benelux. That is our go-to-market idea for this part. The platform we built here will be the foundation of the ACS edge platform, where we extend until 2032 up to 60 data centers. The beauty here is we are capable in delivering these data centers, these edge data centers, close to the cities, close to the client, as we are so distributed over Europe and can really deliver close to the clients. You have to have the capabilities to integrate infrastructure into the local ecosystem. The data center has to look differently. Somebody said, "Look, this looks like a spa or a nice hotel." It has to be quiet. You have to really apply to what is necessary in order to build near to the city infrastructure.
Having said this, our go-to-market strategy starts in Germany in a very dense area. We are close to potential clients, and we have already visibility on the first contracts in our Urizon Cloud. Then we will roll it out. You see here our first data center is already in operation. Mid of 2026, the second one goes into operation. We ramp up with our platform as this is made to scale quickly, exponentially. We ramp up to above 30 data centers in 2030. From there on, the exponential curve will go on. We see the Urizon, so the cloud layer, scaling faster than the physical layer, which allows us, every time we have some demand, to put some data centers into action here. The modular approach and the way we fit out the data centers is optimized over the lifecycle.
We optimized financing, and we optimized the whole structure in order to have here the best value out of it. We put equity value of approximately $400 million up to $800 million into the platform, and sorry, $200 million into the platform, come out with an equity value of $400 million-$800 million. The very thrilling thing is being now more diversified and having access as well to other layers where we offer cloud. The cloud is not that asset-intense as this comes with different margins, and it comes with different multiples we see in the future for this platform. I would say conservatively, we valued this with $1 billion in 2030, but we see as well room for improvement on that side. So many things. Just two words: resilience and sovereignty. We heard a lot.
When you talk about data centers and you have here a mesh of these centers, edge data centers, you can imagine that government and public entities are very interested in that today. If you take out one of these data centers, the others take over. You can imagine it like a mesh of a wireless LAN that you have at home. This is our contribution as well to resilience here. Thank you, Burns. Just to finalize, a final slide where we can see the contribution to the value of the data center business by 2030 per the three fundamental pillars of a strategy: engineering and construction, large data center colocation, and edge data center or cloud services reaching more than $25 billion by 2030. Thank you so much for your attention.
We look forward to providing further updates in the future as we reach successfully our business targets. Thanks a lot.
Okay. Hopefully, you found all very interesting. I'm going to recap and give some closing remarks. The first one is what's our objective, which is to become an end-to-end provider and to accelerate our plans in the new verticals. As I said before, they are all related. Our game, our play is in the infrastructure space, right? The future infrastructure needs to reset, and it's all interrelated. This is very important to understand what's coming in the future. All vertical growth comes together. You cannot separate one from the other. That's why it's so important to grow all of them at the same time, obviously subject to timing in the market.
The second thing is we want to keep our leadership in the traditional core business because 85% of all what we're looking at in data centers, but also in the future vertical growth, it's related with traditional. It is very important to have the ability to have our granular exposure to each one of the regional markets so we can mobilize people, and we can be very close to obtain the permits, obtain the energy, obtain the knowledge and the network with the subcontractors, the supply chain, and the sale performance capabilities. Very, very important. You cannot do all of these without the core traditional capabilities. Obviously, scale up on our AI because AI, artificial intelligence, is relevant not just to understand what we're facing, but also to embed internally for operational efficiencies when delivering these jobs. Our role as a developer is very important.
I will move it to the slide right now of how we're seeing the equity that Emilio was explaining in terms of capital allocation, in terms of higher power, and the potential growth of that. Hopefully, through a presentation, you realize that there's two variables when it comes to managed lanes. The first one is out of the five managed lanes to come, how many of them we will secure, how many of them we will be able to win. We are assuming two out of five. That would come on top of the Georgia 400. There's plenty of other managed lanes coming in the future as well. It would be a matter of timing. The second one is when it comes to data centers. We know the 1.7 gigawatts where we are with that.
Both Jim and Bethany explained very well the status from a construction perspective of all those jobs, all of them with the energy. We're just starting on the construction, or we are far advanced in the construction, but also the different milestones when it comes to that data center. Securing power, which in our case, all of them have the power secured, to having it ready to build. At that stage, we're talking about $2 million per megawatt valuation. This is market standard. We get to a lease. Price goes up to $4 million per megawatt. The next milestone is once everything is under operation. At that stage, we're talking about $22 million, more or less, per megawatt enterprise value, between $12 million-$14 million, around $13 million per megawatt equity value. This is a market number. Of course, it's subject to change.
It's subject to evolve through the years. What is important on our side is not so much to focus on that; it's to focus on the delivery, to focus on making sure that we have the right level of megawatts per year in each one of the stages. We will be communicating transparently every time we achieve another milestone so you can have a much proper analysis of the valuation of our portfolio. We get specifically about the data center. We believe we can get by 2033 to 60. Those are the small ones. You will realize the value of the cloud services associated to it. That's quite relevant. Obviously, specifically, which was the main purpose of today's Investor Day, our leading position in data center, which in the short, medium term, it's going to be very, very, very important for us.
With all of this, this is where we see our equity valuation from a development perspective, from an equity perspective, through our investments. Let's focus on the one on the right because this is where we believe we will be in 2033 because by 2033, we will have at least three managed lanes under operations, which means applying the ratio I mentioned before, six times we're talking about $18 billion with $3 billion invested, accumulated investments by 2033. If we go to data centers, you saw the evolution of the value. Again, this is looking at each one of the sites, each one of the megawatts available under operations, under lease, etc., and we will be incorporating. Based on what I just explained, we're assuming that our large data center platform will be valued at $8.3 billion by 2033.
Our small one, or the additional 1.3 gigawatt that it's advanced, would add $6 billion, and then the small ones, $2.5 billion. And then the energy industry and natural resources. So that's on 2033, right? That's a $35 billion-$40 billion equity value. Now, let's move into the 2030 because then not everything will be under operations. That's when we need to start getting into this kind of cash flows, where it is. We have all that information, and we have all that information available for any one of you that wants to go to the detail of each one of the analyses we are putting here on the screen, right? So we are fully available not just to answer any question, but to provide all information on the analysis so you can get comfortable with the conclusions. In 2030, managed lanes are more complex because they won't be in operation.
That's when we need to do an estimate. It's more subjective. We are coming to the conclusion of $3.6 billion by 2030 of the $1.5 billion invested. Those prices will not be in operations at that point. It's more challenging to put a number to it. When it comes to data center, in theory, it's much more objective, subject to the ratios I said, if the market didn't change. The question is, is the market going to change when it comes to assumptions? What we know is how much we're going to have at each stage in our data center platform. That's when we come to the $7 billion conclusion of equity value for large-scale data centers and to the $4.5 billion, the additional 1.3 GW, and the small one, the $1.4 billion, which Burns explained before. Energy, we're assuming $1.5 billion versus $700 million, not material at these stages.
I think that we, as I said in my presentation at the beginning, this will be more ad hoc, and we'll be analyzing. We need more information before we start injecting a lot of money in our projects, in our verticals. This is how we come to the $18 billion valuation by 2030 of our assets, $35 billion-$40 billion. This is ACS share, right? This is attributable to ACS. We have taken our percentage of the one-third in the managed lanes or the 50% in the platform. This is our share, okay? Again, all this information is available for scrutiny for everyone that would like to have a follow-up. This does not include Abertis, okay? Abertis, we're working separately, and we will continue providing information as we go.
This is the bridge if we add the EBITDA valuation coming out of the engineering and construction business of data centers, but just data centers. The first part that you see on the left is how we are more or less dividing the current market cap into different stages. How much of that is the infra investment equity value? How much of that is the end-to-end offering equity value, and how much is the data center? It is a little bit subjective, right, as you can imagine, because it is not so easy for us to understand the consensus of the market for each one of the areas. Let us just focus on the additional value, right? Not so much how we divide the initial market cap, but the additional value. The first one is the $13 billion-$15 billion data center engineering and construction equity value.
We believe that $9 billion increase versus the valuation today is coming from what Jim was explaining before. The additional revenues, the additional EBITDA that we are going to incorporate into the business at a ratio, I mean, more or less of 5%-6% EBITDA versus revenues. We take the additional revenues that most of it we have already frameworks in place, plus we have the visibility of our own projects. There is a lot of certainty on this amount. The question is, at what multiple you evaluate that EBITDA? We are assuming 12-14. That is where we come to this valuation. The equity value increase for the infrastructure investments. This is the $15 billion-$20 billion generated before. This is the $18 billion that I was showing in the previous slide.
This comes from development, from equity, from managed lanes, from data centers, the 3 GW, plus the edge, and a little bit on the industrial side. All of that comes to a valuation of $45 billion-$50 billion versus the $20 billion market cap today. This does not quantify the EBITDA and the rest of our business. This is not quantifying engineering construction or the other verticals. This is just pure additional value coming from data centers and the infrastructure equity investments we have been describing, okay? We leave everything else that we are not quantifying at these stages from growth in our more mature core business or defense or Germany, nuclear, etc., right? 2026 is going to be a year of multiple news because, obviously, three out of five managed lanes will be awarded in 2026. It will be important for all the reasons that I explained.
We believe that it will be an important year for Abertis because of the extensions, because some of the negotiations are ongoing right now that will be announced, and because of potential M&A. It will be an important year because we will be communicating transparently the states of each one of our assets when it comes to data centers. Of course, I hope to continue giving very good news when it comes to all the verticals. I am going to stop here because I think that it is very important, and it is the time to get into the Q&A. I would love to have all my colleagues join me in the states to answer all your questions. Any question?
Hello? Thank you. I am Arpitell from UBS. Four questions primarily focused on the data center development.
Number one, we spoke about sort of the pricing mechanisms for construction to ensure there's a maximum price which is paid. Two parts to that. What about the risks of delays in terms of timing, so not the cost? Also, following construction, if you have, for whatever reason, delays to the lease implementation, is this just lost revenues forgone for ACS? Or do the clients which you lease the space to, do they then receive compensation for this lost time? That's the first question. Secondly, you spoke about the sites being power-ready for ACS. I'm aware of a few sites, I believe, in California which are power-ready, not from ACS, but more broadly data center sites. The infrastructure is essentially sitting idle because the grid infrastructure is not actually sufficient to carry the electricity supply to those data centers.
In the locations where you are building data centers, what gives you confidence that power can be supplied to those areas? Are there any measures you are taking to ensure that that will be in place for once the data centers are operationally ready? The third one, can you just help us get a better understanding of the costs associated with the maintenance of the data centers, the different moving parts? Which of these are fixed, variable? Just trying to get a better feel for the margins that this business can generate. A fourth one, if I may, just on the difference in sort of the data center demand environments in Europe and the U.S.. A few of your competitors have flagged a bit more softness in Europe relative to the U.S., so just wanted to understand the trends there. Thank you. Okay.
Jim, do you want to start with the construction ones and then?
Yes. The first question regarding the growth of the market? GMP. Oh, sorry, GMP. So GMP, so a guaranteed maximum price contract is effectively a collaborative open book contract, right, that allows us to work collectively from inception, if you will, all the way through design, through construction, and through implementation. It allows us to integrate all the offerings or all the individual functions as well as each of the companies such that we can mitigate and manage the rest together, right? Our outcomes of GMP contracts are very, very reliable as opposed to a traditional lump sum or design-build lump sum. The outcomes are incredibly reliable. It's the most common delivery method of data centers in the U.S., for sure, because the risks are so dynamic and you need to move very quickly, right?
Speed to market is so very, very important. A GMP contract is good for everyone. It's good for an owner from a developer standpoint because the risks are mitigated and shared and collaborated versus being at risk, excuse me, an adversarial relationship where the risk is one or the other's, right? There is risk transfer throughout the design and then the construction, but it's transferred at the time that's able to be mitigated and managed. It ultimately becomes a guaranteed maximum price or a fixed number, if you will, but it happens over time where you can collaborate and therefore have a more predictable outcome or really a cost-certain project.
The energy. I can take that one for sure. That's absolutely right. That's what we live as developers.
Sometimes the energy has been sort of promised, but the utility company is unable to deliver it because they have not undertaken the infrastructure upgrades required to deliver that power. How we manage that risk is we would never go ahead purchasing a site where the power is not halfway through that moment of being available at the site. We would not make any commitment to future customers or start investing capital in vertical infrastructure until there is certainty that the power can physically be delivered into the site. I would add a couple of things to what Vicente was explaining. We go through a very, very thorough due diligence when it comes to the projects. The due diligence is not so much about the ability to get the energy or the permit because that is basically a big risk.
We are trying, especially as we begin, maybe in the future we change and we take a little more risk. Right now, we are very much risk-averse. If we have been jumping into some of these sites, it's because the energy is ready. Having said that, you're right. Energy is not just having the permit. It's making sure that you're able to do all the extensions. By the same token, throughout the due diligence, we made sure that there was not going to be any problem with any additional extension in the distribution line, any additional extension in substation. We did not want, especially at the beginning, because we need to make sure that we show success. We cannot just take the risk.
We have been very, very clear in the first 1.7 GW, and we will continue being very, very clear in not taking uncertain risk when it comes to this potential additional substation or distribution, right? At least with the first packets. The other thing that you asked for is the risk, what happens with the list? Typically, hyperscalers and some of the big clients, they want, by the time they really start getting into negotiations, they want the 12 months of visibility. At that point, any potential risk that could cause a delay, it is gone. You rely on yourself at that stage about finishing the main building and making sure that you install the EPDUs and the connections, right? The risk is quite limited at that point. Yes, there are penalties under the lease agreement if you do not finish on time.
By the time you sign the lease, you are very clear about what you have to do, and you are very well advanced with everything in place. If we were signing the lease at the very beginning, obviously, the risk could be higher, and we avoid that. That was it. What was our market enemy? Hiya.
Thanks very much for the detailed presentation. It's Graham Hunt from Jefferies. I'll just ask two questions. Firstly, I wondered if you could give a little bit more detail around the lease agreements that you'll be signing for these data centers, just in terms of duration, energy hedging, risk around renewals, and changes to the terms of those lease agreements. Just additional color there would be helpful. Second question, you talked a lot about equity value creation out to 2030.
How should we think about this in terms of cash being returned to shareholders? When do these assets start becoming cash generative? What's your thinking around shareholder distributions as these assets ramp up? Thanks.
Vicente, you take the first one on the list.
First one, for sure. The lease is obviously the leases that we're negotiating at the moment, and we've already exchanged legal documentation with a number of potential customers around their confidentiality provisions. As we explained in the presentation, we're trying to target leases that are more than 10 years of duration with potential extensions of that lease. We think that that is pretty standard within the industry, whether that is a full colo lease or a triple lease, as we explained. The rest of the provisions, I think you made reference to early termination. Was that the question?
Early termination, as I tried to explain earlier, we find those types of provisions in the leases. The way we take comfort is from a number of things. We take comfort from the payment that you will receive at that point if there is an early termination. Obviously, there is early termination for convenience or early termination because of a cost. We take comfort from the natural barrier that I was trying to explain earlier. We do not know any precedent of an early termination of a lease in the markets where we operate, first of all. Secondly, there is a natural barrier to exit for those hyperscalers, given the amount of investment that they do when they come to our data centers as hosts.
Third, we take comfort from our operational experience, having decades of experience of operating infrastructure assets that are more complicated than a data center. We have a lot of confidence in our ability to deliver to those standards. Thirdly, as I said, we negotiate under which conditions a tenant might be able to exercise the right of an early termination.
In terms of the cash flow generation from these assets, obviously, you have got two very different types. You have got the data center asset, high or fast completion to operations, plus the operations income coming in. The cash yield is pretty stable after that once it is in operation. It can take, since inception of development to completion, it can be around three years. Construction could be anything below two years. It is quite fast evolving in terms of delivering cash.
The managed lease, obviously, is a longer wait, and it's more backed on the value generation, but as soon as it's coming in operations, it's the same thing. The point I would add is, in terms of our capital allocation strategy, what we're doing is doing this with equity. So we're not relying on debt to fund all this investment because we are conscious that these are assets that will start generating cash in some time. Adding to what Emilio is explaining, at the end of the day, and we started explaining three years ago to the market that ACS, it's a very good company for those that are looking for yield, but it's also a very good company for those that are looking for growth. It's a very good balance. Putting aside, once all those assets become mature, we will continue increasing revenues, EBITDA, and dividends.
Actually, I announced before that most likely we're going to start increasing our dividends. We are building, we are creating revenues and EBITDA from the start because we are not just the developers. We are the companies building all our projects. You will see, in addition to the work we do for third parties, you will start seeing an increase in our revenues and EBITDA, right? All of that will hopefully increase the profits and therefore will increase dividends. You have the long-term assets once those recurrent EBITDA start coming in. We are not considering in our base case any recycling of equity. The data centers nowadays are very liquid, and managed lanes are very liquid once they are in operations, right?
At this stage, they are not our base case, but eventually, we could consider recycling part of our managed lanes on Abertis or outside, but also recycling part. Our agreement in the platform already considers potentially recycling part of the equity into a yield curve. And that's embedded in the current agreement with BlackRock GIP. Whether we do it or not, we discuss, but there's other instances where we could start anticipating part of that value.
Hi. Thanks for taking my questions. I'm Alberto Valencia from Malantra. The first question is on Turner. You provided quite staggering numbers of the total investment in the data center industry. It seems that your estimate is that Turner will be, of course, the leader, but it's still a small portion of it. I was just trying to understand how do you see things playing out?
I don't know if this is that you think that there is no sufficient capacity for all those big numbers of investments to actually get done, or if you think there will be a large number of players providing the supply to build all that capacity and there will be maybe some consolidation. Also, for Turner to achieve those numbers, just to understand how are you planning to upscale capacity or if that will cannibalize some of the capacity you have for other sectors. Just that on Turner. My second question would be on the Greenfield ventures on data centers. In some parts of the presentation, I think you were talking about power, megawatts, and others on IT megawatts. If you could make the distinction clear so that we know how much of the megawatts actually are for IT.
The last question would be on management and the plans. If you plan to change the current long-term and the plans of the management team to align with the targets you provided for 2030 or 2035?
Thank you. I'll take the capacity issue with Turner. There are a lot of ways to address capacity. We talked about one of them within the industry of marginalization, which takes the trade labor constraint and almost doubles or triples it by doing half of that labor, if not more, in a fabrication facility and off-site. The trade labor I worry about more than Turner's ability to be able to provide capacity to increase our revenue and run these projects. That said, we've done a very good job and will continue to manage redeployment of our resources. We're in 45 different geographic locations in the United States.
Not all of those locations are at full capacity right now from a market standpoint. A market goes down, a market goes up. We redeploy assets, our folks, our best resource to areas geographically and also markets where they can have the greatest impact. It takes a lot of work, but if we did not do that, we would stagnate in a geographic location or we would stagnate in a market. Right now, the data center market offers us such a unique opportunity. They are great customers. They are at great margins. We have done a lot of work to redeploy our assets to be able to meet that capacity. The IT, I can take that one.
You are right that we refer to both instances in the presentation, total power utility and total power IT.
Those are related through the PUE, which is a factor of energy efficiency within the data center. Ultimately, the IT power is the amount of power that is left for processing capacity in the data center. The higher the PUE, the more inefficient the building is; the lower, the more efficient it is. We're running our calculations at the moment with 1.45, which we think is very conservative. The final PUE, we will know once the design is fully finalized and the facility is completed. We think that is conservative, but therefore, in a business plan, there is room for an upside in that sense. More megawatts IT would end up going to our customers that we represent at the moment in the presentation. Would you mind to repeat the third one on managed lanes? Sorry.
My third question was on executive incentives and whether you will change or does the board, I do not know, plan to change the incentive plans to align the incentives with the targets provided today for 2030 or 2035? We have been working over the last three years exactly on changing and aligned incentive plans with the work, right? I do think that so far we have been able, we need to adjust every year. I am not expecting anything different from 2025 and 2026 when we approve the new one. We have been already in the last three years changing scorecards, changing STIs, changing LTIs to make sure that we align all of that in each one of the scorecards to the objective of every company or development department, etc., and then everyone with the same ACS objective.
Even more, the stock options plan that we did publish three years ago for the first time was including everyone within ACS, Hochtief, Turner, Flatiron, and to make sure that everyone was aligned into the one group, one team culture. Absolutely. The other thing that I would like to mention, when we speak about revenues, right?
If the question, which is a very good question, is, do you think that the world is going to be able to build trillions of infrastructure in defense, trillions in data centers, trillions? I mean, the question is absolutely perfect. I do not know. What I can say is that our revenues that we are showing here, we have the capability because most, I mean, the cap, the threshold, the limit of what we are offering here and showing, it is more based on our capabilities to build, not so much on the potential market.
The potential market is infinite, trillions. I wish we were able to do all of that. But we're basing all our estimates in what we believe we can do with our resources and the additional resources we believe we can bring into the company. I wish then the entire market was addressable from a resources perspective because then obviously we could increase. And maybe if our modular strategy continues being successful, as Mike was explaining, and if we're able to continue increasing our supply, we will be able to get more of the market because most of data centers who are not talking anymore about 7 MW, 10 MW, 100 MW. We're talking about gigawatts. When you get into the gigawatt threshold, not so many companies, there's not so many players. So the restriction is in the players more than in the market.
So three questions, please.
First, it's on the deal with GIP. One is on accounting. Will the vehicle be equity accounted, fully consolidated by ACS? Second, I'm not very clear how much cash is ACS receiving upfront. Is it EUR 1 billion? Is it EUR 500 million? And lastly, on the value uplift from the managed lanes, you mentioned a multiplier of six times cash on cash multiple by 2033, if I recall correctly. And by then, SR400 will have only been in operation two years, which is not a lot to prove its worth. I was wondering if you are baking in any expectation for the sale of this asset by then already, or if it's just your expectation that you will retain the asset in full by then. When to start? On the transaction structure? Okay.
In terms of the transaction structure, the transaction is valued at EUR 2.2 billion, 100%, all of the assets in the platform, right? We are essentially selling 50%. I mean, we do it in a way where the assets are contributed, we contribute our share, and the asset receives 100% of the amount, right? At the end of the day, the full valuation is EUR 2.2. We receive EUR 1 billion net of what we have to put. However, there is half of that, which is subject to earnouts, right? We receive upfront roughly EUR 500 million. We receive another EUR 600 million via earnouts as a net of all the combination of the transaction. Is that clear? Okay. From an accounting perspective, it is equity accounted, as Vicente explained. It is a standalone platform with the full team, 50% partnership, no full consolidation by ACS. It is a joint partnership.
On the managed lanes, our experience in managed lanes is much better than six times, right? At the end of the day, you take a managing project with IRRs on day one at financial close, above 15%, 16%, 17%. You go through all that, the construction, which is taking seven years. You get at the end of the construction, you have two years of visibility of cash flows. I mean, 288 was 10 x by the time we got to that point. If you look at some of the competition, it was more than six times. We are taking the lower of the cases we have seen in the market. The six times is based just on going from a 16% discount rate to whatever is appropriate when you have stable cash flows and the visibility of the cash flows instead of discounting seven years in advance.
We're not assuming changes in traffic or changes versus the financial model. We're not assuming changes in CapEx. We're not assuming changes in operational maintenance. In 2015, we invested EUR 360 million. When we sold in 2022, we're talking about EUR 2.7 billion. When it was removed from us, market value was EUR 4 billion, and that's why TechStock decided to get it at the close of the contract. No, we don't believe that it is crazy. We're not trying to beg anything. We're just basing it, it's pure mathematics, right? Of course, there could be mistakes for good or for bad when it comes to traffic projections, CapEx or OpEx. That can happen, right? At this stage, we want to make sure that we keep assets. A very good way to keep assets is to transfer to Abertis. We retain 50% of the asset, right?
Or if Abertis at the point does not want to acquire the asset, then we can go to the market, but always with the objective to remain a big part in the asset because we want to generate long-term cash flows. We do not want just to be in the business of recycling equity because then we generate a lot of value in the short term. We increase and we have the peak when we sell, but then that is gone, right? We want to make sure that we grow in the future with that. Hello. Flavio from BNP Paribas. Thanks for the presentation. I think ahead there are lots of exciting opportunities. That is my view. Some people in the market think that ACS is investing at the peak of the market and this is an AI bubble. I am just curious from your perspective, how do you see things?
I mean, clearly you're investing in it, so you think it's not a bubble. But just what are the KPIs you're monitoring? What is the feedback you're receiving from clients, from the hyperscalers about the trends in AI and data centers? So I'll take that one. When you look at the AI market and you look at the figures, we see figures from $5 trillion-$6 trillion, $10 trillion, $15 trillion. And that's only in infrastructure. If you start adding semiconductors, if you start adding models, platforms, etc., is it a bubble or not? And there's a lot about, there's a lot of literature and analysis with different views on that because there's, of course, a lot of different trends in the market that could influence that demand. The first one is timing.
To what extent we are going to be by 2040, where we believe we're going to be when it comes to smart hardware, robotics, data, 5G, 6G, satellite connections, etc., per state. How deep in the society all of that is going to be because right now everyone is having a view on that. The second one is there's a race to control all of that. Potentially some companies will get out of the race, and that could influence demand. Then you have the quantum computing, which of course has an influence in data centers. There are a lot of different elements that will define the extent of the bubble. There are two important questions on this. The first one is not just the demand, but the supply.
I think that the AI race or data centers specifically, it's going to be more driven by the supply, what can be built and by when more than the needs, right? Coming back to my previous example. We're talking about trillions. That's infinite. Infinite divided by two, it's infinite. By 10, it's infinite. By 20, it's infinite. By 100, it continues being infinite. The demand is infinite no matter how much of a bubble it is. If we were a major semiconductor company like Nvidia, yeah, of course, we would be playing on an infinite world. More or less, of course, would have an impact in my projections. We are tiny. We're talking about $20 billion production out of a $70 trillion universe. We are not affected by that.
We know that what we have in front of us will be needed by 2030 and by 2033, without a doubt. The question is not so much this. It is how can we multiply this by 10, which is what our clients are asking us? So far, we do not know. We are working on it, right? There is no concern on our side about what we are presenting today. If the question was, what do you think about the $70 trillion universe of AI, then we would be two hours talking about it and the potential of a bubble. That is a different discussion from what we are facing here today.
There are some online questions. Luis Prieto from Kepler is asking about our EUR 1.2-EUR 1.3 billion. If the expectation for data center engineering construction, is it a base case scenario?
Could you elaborate something about what could be the worst and the best scenario about that? What are the E2E contractors that you have used for the valuation of data centers EMC operations? Why this multiple we have applied? The second question was related to the recent answer that you have done, which is what are the key risks of the announced development platform? Where could things go wrong? Jim, you want to start with the first one?
What was the first one again? Sorry. Yeah, it's about the expectations, how we have reached this $1.2 billion, $1.3 billion, who could be the worst scenario or the best scenario as well. Sure. On the EBITDA side, right? Yeah.
If you look at that valuation, from a standpoint of us growing at a minimum to a $20 billion revenue in 2030, right, I think that easily the upward side of that is $25 billion, right? That range, when you look at the U.S. market, continues to stay strong. That is at least $20 billion and climbing, right? I think that is the correct range, right? Could it be a little bit above that? Yes, right? We have seen substantial growth in the U.S. market, which is substantial, right? Just the growth from $4 billion last year to $8 billion this year and go on. I think we are in that range, but I think there is some room to grow it. Yeah.
The other question was about which are the comparables, the peers that we have used for the valuation of this business. One second. Go ahead.
I mean, when it comes to, there's a couple of things when it comes to evaluation, specifically about EBITDAs and revenues. The first one is we have big visibility on the revenues, and we have big visibility of what’s EBITDA we’re getting from those revenues when it comes to data centers. At this stage, two types of clients. The platform itself, which is very defined when it comes to revenues. Third parties, very much defined because we have framework agreements. As we get into larger data centers, larger framework agreements, larger the visibility. The EBITDA percent that we’re applying is the EBITDA we are getting, right? We have full visibility on our cost and what’s EBITDA. The evaluation of the multiplier of the EBITDA is when we need to go to the market, right? There’s a lot of discussions about Turner.
The multiplier of Turner, is it 13, 14, 15 times? Our competition, we're seeing 16, we're seeing 15, we're seeing 17, depending on which one you get in this market. We're just applying the lower part, which is the 14, multiplied by the EBITDA, right? Lower case when it comes to the multiplier with a very certain revenue and EBITDA streamline. He was asking as well about the risk involved in the development of the platform. How you see the future risk and what things could go wrong? In the platform or in the partnership with Abertis? In the platform, I mean, the biggest risk is not being able to place the capacity. That would be the fatal scenario for us.
As Juan was saying, we have no doubt today that the amount of gigawatts that we're developing will be placed in the market, given the reaction that the market is having. As I said, we have some of our assets that deliver capacity in 2026, 2027. We are already negotiating the leases. In some instances, there is interest from a number of potential tenants. We are very confident that given where we've chosen to invest in the sites, the capacity will be placed.
Good afternoon. Nicholas from Morgan Stanley. Just coming back on the build-out of data centers. You were saying you had $14 billion backlog right now. You will hit $17 billion by the end of the year. $8 billion of revenue. So you'll be running at right now, you're starting at five, six at the start of the year.
You're running at $9 billion, $10 billion run rate on the quarterly basis. On a yearly basis, annualizing the quarterly basis. Why don't you see even more growth in the short term? I mean, you've grown tremendously in 2025. You've got so much in the backlog in 2026. I mean, why stop at 30% growth at Turner? I mean, I know these numbers are big, but what you already have secured is even bigger than that. I'm just wondering, you've got your usual caution. Could it just be from 30%- 50%? We've seen a little bit this year. That's the first point. Second point on the JV with GIP, the portfolio is not really skewed to the U.S., quite skewed to, I mean, lovely Spain. I mean, we like it, but it's not super geographically diversified. Why is that?
I mean, it's not a fair reflection of the state of the market right now. That's what I'm trying to say. Again, you've got your roots in Spain. You've secured some great sites and so on. Does the full portfolio show greater skew to the U.S. from here, the 3 giga, or is it still kind of the same as what you have right now?
Jim, you want to start with the first one?
Absolutely. Great question. If you look at, we had significant growth from 2024 to 2025, right? You look at the backlog numbers and do a comparison. At the end of 2024, our backlog would have been about $7 billion, right? We transferred that to $8 billion, obviously, which we sell work in the current year that we also put in place.
If you look at next year and the forecast, it's a little bit more to do with the fact that some of these projects are getting bigger. They're big numbers, but they're stretching out because there are multiple phases. It used to be we would build 50 MW, 100 MW. Now we're building a gigawatt or 2 GW, right? You'll see big chunks of revenue get booked, but then it's going to take longer to burn it off, right? Yes, we're seeing backlog grow, which is good because our revenue, obviously, as that grows, we're eating the backlog. The reality is these jobs get bigger and therefore they're a longer run, right? Do I think there's a higher, could the number next year be a little bit higher? Sure it could. The backlog is less 100 MW jobs and more gigawatt jobs.
Therefore, the run rate of that revenue is longer.
To add to that, one very, very, very important aspect of the projects we are dealing with is we need to achieve excellence in construction. We cannot fail. Private clients are very, very strict, and we need to make sure we achieve a rate of coming back clients above 80%, 85%, 90%. Every time we communicate, we are basing our communication in our capabilities, capacity, and making sure that we achieve excellence. When you look at last year, Turner gave a guidance, or we gave a guidance between 17%-32%, right? That was at the beginning of 2025. We are achieving 60%, right? We pushed the entire guidance above. Why? There was a point that we were comfortable that we had the capability and the ability to deliver at the right standard, right?
That's very important. We need to make sure that we assess that as we go. We cannot just go crazy thinking this is not just revenues and EBITDA. This is the consequence of making sure we achieve a certain level of a standard every time we build a job. That's why we've been outperforming our guidance repeatedly over the last three years, right? People say, "Why are you so conservative?" I'm not conservative. We're not just trying to play. We want to make sure that what we announce is what we can build, right? If it goes better, great. That's the first one. The second one is you're right. The first 1.7 GW is a little bit unusual because it has a big weight on Spain. When you look at the 11 GW pipeline, it's almost the US, right?
Because that's where the growth is. That's where the AI training is. That's where most of the processing is going to be. It was a little bit unusual that we started. We did secure a few sites in Spain. We will continue doing a lot in Europe. Most likely we'll increase a little bit on the large ones. It is very likely that the small edge will grow into higher megawatts, right? Europe is not facing the 1 GW, 2 GW, 3 GW that the U.S. is facing. That is why the 11 GW is mainly in the U.S.. It is unusual, the first 1.7 GWs. More importantly, the platform with GIP is not just for the first 1.7 GW or 1.4 IT. It is for more.
There's a percentage of projects that were not included in the initial platform, but there's no right of first offer for the platform, which is the 1.3 GW that we're showing on the screen. That's out of the platform scope. Certainly we will discuss with the platform when the time comes.
Any other questions? All right then. Thank you so much. We appreciate your time. Yes. We appreciate your time, the Q&A, the interest, your support. Happy to answer any question in the calculator right now. Of course, as I said, we will be available for any follow-up in evaluations, modeling, etc. Thank you.
Thank you.