Right, good morning everybody, and welcome to our full year 2024 results presentation. Thanks very much for all of you joining us in the room on a Friday morning. Good to see some people still come into the city on a Friday. But also thanks to everybody joining us online. As usual, I'm going to make some opening remarks to set the context, and then we're going to go through the presentation and do Q&A at the end. As you know, we've been following a consistent strategy for almost 15 years now, founded upon disciplined capital allocation and operational excellence, and underpinned by an efficient capital and corporate structure, while taking account of the needs of our wider stakeholders through our Responsible SEGRO commitments. Application of this strategy has continued to deliver for us in 2024, and it's positioned the business to perform well in the future.
Our teams have been driving rents. We've signed up GBP 91 million of new commitments during the year, which is our third best year on record, and I think that's pretty impressive when you consider the macroeconomic environment we've been operating in. We've continued to invest for growth, both through executing on our profitable development program and leveraging our local knowledge and strong relationships on the ground to buy some very good assets at attractive prices. In parallel, we've crystallized profits and created extra liquidity through selling carefully chosen assets to motivated buyers, and our balance sheet is in great shape. With the additional firepower provided by our equity raise in February last year and the disposal proceeds we've generated in 2024, we're very well positioned to invest at a point in the cycle when we think having access to capital is a competitive advantage.
The hard work in 2024 added to our strong track record of delivering growth in rents, and that's fed through into an 8% average annual growth in earnings and dividends over the past eight years. We've also continued to deliver on our Responsible SEGRO commitments across three main areas. We've made great progress in cutting our carbon footprint, not least through the doubling of our solar capacity in the year and reducing the carbon intensity of our developments, and we've updated our science-based net zero targets in line with the latest SBTi methodology. Investing in our local communities also remains in focus. We now have 14 community investment plans designed specifically for our larger asset clusters across Europe, supported by our own employees, our suppliers, and our customer base.
And our nurturing talent, sorry, I remembered to click on, and our nurturing talent efforts, which are key to maintaining and improving our operating platform, not least through the continuous development of our bench strength and the improvement of our people policies to be more family-friendly, are helping us to achieve high levels of employee engagement and create a diverse pool of existing and new talent in the business. So 2024 has been a year of further strong delivery by our teams in what we know was a challenging environment. And in a moment, Soumen will take you through some of the more detailed financial and operational figures behind that statement. We're also feeling confident about 2025 and the opportunities in front of us. Our business is in fantastic shape, and we're primed for further growth in earnings and dividends.
We're super excited about the significant additional value creation opportunity that we now have in front of us with regard to data centers and the growing 2.3-gigawatt land bank that we've created. Thank you. It'S gives a battery issue. We're going to take you through each of these elements in turn. So let me hand over now to Soumen, who will talk about the first chunk.
Thank you, David. Good morning, everybody. So let's have one of these work. Starting here on slide seven, and the usual slide with the key financial metrics. Now, look, the key takeaway, as David said, it's been a really busy, really active year on the leasing front, which has fed earnings, it's fed dividend growth, while asset values have been stable after what was clearly a roller coaster 2022 and 2023. So we delivered 5.5% growth in earnings and dividends per share. Now, it's just worth noting that the additional rent and income that we earned on the proceeds of the equity raise is the reason the profit before tax is up 15%, but EPS is up 5%, as that extra income is offset by the higher share count. The equity raise was fairly EPS neutral in the year, but clearly it's accretive once it's fully deployed.
We're recommending a final dividend of 20.2 pence, which makes a full year dividend 29.3p. We saw a small increase in the like-for-like valuation of the portfolio, but due to a higher level of disposals in the year, the total remains at GBP 17.8 billion. NAV per share is unchanged at 907 pence. The balance sheet is in good shape, with loan-to-value down to 28%, providing considerable opportunity for growth. I'm going to write in a minute. Right, so turning now to the income statement on slide eight. We saw 8% growth in net rental income in the year, which I'll break down for you on the next slide. Before we get there, just a couple of things to note. Firstly, capitalized interest was GBP 69 million in 2024, similar to the level of 2023 and likely to be around these sort of levels in 2025.
And on costs, we saw an increase in the cost ratio due to increases in both our property and our admin costs. Now, the key items were higher vacant property costs, some additional tech-related spend, and some one-off abortive transaction costs. Now, we anticipate the ratio will move back below 20% without those one-off costs and as the vacancy is leased up. On slide nine now, and turning into that net rental income growth in some more detail. So we delivered strong growth in net rents, up GBP 41 million in the year. Now, as in previous years, there are two main contributors to that growth. There's first two pink bars. Rent on the standing portfolio grew GBP 29 million, with very strong like-for-like growth of 5.8%. And that's driven by the capture of reversion in the U.K. and indexation and leasing activity across the continent.
Development completions added GBP 32 million in the period. The investment activity, the buying and the selling, resulted in a net loss of GBP 9 million of income last year. And you should just note that due to the timing of that activity, the full year impact is GBP 6 million higher than that. And slide 38 in the appendix has got some more detail on that. The other box at the end is a mix of take-backs for developments, FX, and some non-recurring items. Looking ahead from here, we will continue to drive rental income strongly, with like-for-like growth, especially from reversion and new income coming through from our development pipeline. Turning that to the valuation, and we're on slide 10. 2024 saw the first positive valuation after two years of declines.
Pleasingly, there was a positive, albeit modest, revaluation in both the UK and on the continent in the second half of 2024, which suggests we could well be past the inflection point on values. The portfolio is still at GBP 17.8 billion, net of the revaluation, capital recycling, and CapEx. Yields were fairly flat, especially in the second half. So the capital value growth will increasingly link to ERV growth going forward. We saw a healthy 3.2% ERV growth in the year. That was 3.7% in the UK and 2.3% in the continent. And again, that's weighted more towards the second half. Turning now to the balance sheet, which remains in great shape and provides with significant firepower for growth with over GBP 2 billion of available liquidity.
Loan-to-value has reduced to 28%, and our credit rating is stable at A-minus, which continues to be a differentiator versus the vast majority of the real estate universe, which is rated mostly BBB. Our debt metrics are in really good shape. Net debt to EBITDA is down significantly from 10.4 times to 8.6 times, as EBITDA has grown and net debt has been reduced by the equity placing and through disposals. and I turn to slide 12. We've got a diversified long-duration debt profile with an average maturity of seven years, which, as you can see on the graph on the left, has debt stretching out till 2042. We've got low refinancing requirements in the next few years, reduced further because we've tapped the debt capital markets in the past six months out of both SEGRO and out of SELP, with bond issues with a 3.5% and 3.75% coupons respectively.
But to help you quantify the impact of what's to come in terms of interest costs from refinancings, we've added the graph on the right-hand side. So this assumes that each year, as we refinance the debt that comes due that year, using today's indicative rates in both sterling and in euros. And you see the overall cost of debt moves up only 10 basis points this year and by 50 basis points through to 2027, but up to a level that's still lower than the level of 2023. And in terms of quantum, we're talking about around GBP 25 million. So look, it's not immaterial, but it is much, much lower than the reversion that we hope to capture in the period over the next three years. So it's very manageable in terms of the earnings impact.
So moving on now, and having provided you the 2024 financial overview, I'm going to take you through the highlights for our operational activity for the last year and how we've created value through both asset management and development. Now, I said earlier that 2024 was a really active year. And on this chart, you can see why. We signed GBP 91 million of new rent during 2024, and that is our third best year on record. Now, our asset managers and leasing managers have been super busy finding opportunities to lease space whilst capturing record levels of reversion. Now, you can see on the graph here, the contribution from the existing portfolio that's shown in red has grown enormously over the years. There is still a very healthy and strong contribution from development lettings, including one of the largest lettings last year at our site at Northampton.
It's fair to say the overall volume of pre-lets in the market is lower than in previous years. The big change is simply that there are fewer mega box pre-lets around. There are still deals to do, and occupiers are still engaging with us in conversations around expansion. In the middle of 2024, converting those opportunities into deals was slower, as we found occupiers were taking longer to take decisions. Pleasingly, we saw a very pronounced pickup in the activity levels in the final weeks of the year across the business. The momentum has continued into 2025. This chart, page 14, the chart on the left illustrates some of what I've just talked about. This chart on the left summarises literally the hundreds of leasing transactions that we undertook in 2024. Every dot represents a single lease event.
You can see it was a really busy year. There were over 400 individual transactions in there. You can see how the momentum in volumes grew in the first half of the year, fell away in the summer, as you might typically expect, but it stayed there in the early part of the autumn. Then in the last two months of the year, it saw a real pickup. Simply, you can see from the density of the dots in December that we had a particularly busy month. Now, on the right, you can see how all that activity has kept the customer retention rate high at 80% and occupancy strong. Now, 94%, it is at the lower end of our target range, and it fell 1% during the period as our leasing progress was offset by a couple of specific items.
Firstly, we completed on some speculative space, for which we're seeing good levels of interest, and we've signed some units already since the year-end, and secondly, we had an occupier in the global media sector who chose to leave us in North London, and that accounts alone for around 50 basis points of group vacancy, but for us, it's an opportunity to capture that reversion earlier than we expected and potentially to move that rent on ahead of ERV, so on slide 15 and diving deeper into the reversion capture, which I've said was a really major driver of that net rental income growth, that we capture a record level of uplift from rent reviews and renewals in the year, 34% for the group and an extraordinary 43% in the UK. Now, you've got to understand, these are really quite remarkable uplifts. They're not one-offs.
We had 170 across the portfolio, and 69 of those showed uplifts over 60%. Now, you realize it requires us having the best space available and great customer relationships to achieve that level of uplift without adversely affecting the retention stats I showed you on the page earlier. It also demonstrates that our customers are able to afford these higher rental levels. As we showed you at the investor day in June, and particularly within our urban portfolio, businesses are providing high-value goods and services and have the pricing power to pass these on. And importantly, we are not done. If you look at the chart on the right-hand side, it shows you we've still a lot more to go for. There's GBP 118 million of further potential reversion to capture, and GBP 71 million of that comes up over the next three years.
So on this slide, having talked to you about lots of numbers across the portfolio, I wanted to highlight a few examples of how we do this on the ground, around our approach to customers and on leasing. So starting on the top left with our German portfolio, and we've been seeing some really strong demand, particularly for our urban space, maybe counter to what some of the macro headlines might suggest. We set a new record warehouse rental level at SEGRO Park Düsseldorf. And in Cologne, we set the highest industrial rent in the city at the estate that you see pictured in the top left. Secondly, our leasing activity helps us set evidence for rent reviews and helps to capture the reversion that we've just talked about. On our East London portfolio in the bottom left of the screen, rents are up 100%.
They were GBP 9 per sq ft five years ago and now close to GBP 20 per sq ft. Thirdly, we use refurbishment to help reposition assets to drive rents further. We're doing it a lot in London at the moment. One of our most striking examples during 2024 was at our flagship estate, Premier Park. We took back a unit in May 2023 when a customer upsized to a larger unit at our recently completed SEGRO Park Hayes. We carried out a high-standard refurbishment on that space and signed a lease to a new customer, which was 13% of ERV, but it was more than double the previous passing rent. And finally, an example of the importance of supporting customers and benefiting from their growth. HG Walter is a butcher providing high-quality meat to some of London's finest restaurants and hotels.
It's been a customer of ours at Tudor Park in West London since 2017. Now, unsurprisingly, their business was badly hit during the pandemic when we locked down. We offered them financial support, but we also pivoted their businesses to start delivering—they pivoted their business to start delivering products to people in their own homes. That helped them not just to survive, but to create a new growth opportunity. So their business is now thriving, so much so they need some new space. We refurbished the unit at SEGRO Park Rainsford Road, which they took occupation last year, taking three times more space. Finally, on this section, this slide looks at our development activity during 2024 and looking ahead to 2025. Now, we completed space equating to GBP 37 million of headline rent. 84% of that is leased and delivering an attractive development yield of 6.9%.
97% of that was BREEAM Excellent rated, reflecting our efforts to develop the most sustainable energy-efficient space for our customers. Looking forward, we have £46 million of headline rent in our current pipeline, almost all of which will complete this year in 2025. The pre-let percentage is 50%, which is a reflection of the lower level of new pre-let signings in 2024. But also, we've just commenced schemes in some of our urban markets where we always build speculatively in responses where we've seen some really strong demand. As I mentioned earlier, one of the regions we've seen this is in Germany, where our newly created flexible modern urban space is being snapped up by mostly small and medium-sized businesses who are servicing growing populations in nearby cities. Today, we've got schemes underway in Berlin, Cologne, and in Düsseldorf.
25% of that space is already under offer, and a further 15% of that is in negotiation. And these schemes don't complete until the second half of this year. So summing up from me, we had a really busy and active 2024 with positive financial and operating metrics, record reversion capture, profitable development completions, and 5.5% earnings and dividend growth. Importantly, we saw a pickup in occupier market activity in the last quarter and continued that into 2025, all of which lays the foundations for continued delivery and growth. And for that, I'll hand you back to David.
Okay, thank you, Soumen. Now, I want to talk about the future and how we are positioned to derive further growth in rents and performance from our portfolio.
As you will know, we've positioned our business to benefit from a number of enduring structural trends, which continue to support occupier demand and feature in many of the conversations that we have with customers regarding their future plans. All four remain very much intact. E-commerce continues to take market share from physical space. The digital economy is becoming an ever more critical part of our business and personal lives. Major cities continue to grow at a faster rate than the wider economy. Our logistics customers continue to reconfigure their supply chains to optimize efficiency and reduce carbon emissions. And most of the corporates we talk to know that improving their sustainability credentials are not just good for the planet, but they're also good for their businesses.
We expect these tailwinds to continue driving demand for well-located and modern industrial and logistics space, especially as and when the macro environment improves. Meanwhile, competition for other uses of brownfield land and tight greenbelt planning restrictions will limit the availability of land in most chosen markets, which will keep new competitive supply in check and maintain upward pressure on rents. We have a fantastic portfolio of assets concentrated in Europe's most attractive markets and a market-leading operating platform dedicated to making sure that we remain close to our customers, offering them great space and service while ensuring that our portfolio delivers attractive returns through the cycle. We also have an exceptional land bank to support profitable development. But it's important to remember the unique composition of our portfolio and our 65% weighting to urban markets.
We have conviction about the attractions of big box logistics, and we believe that we have one of the best, if not the best, logistics portfolios in Europe, along with some fantastic rare land holdings. But as we covered during our investor and analyst day last June, we believe the urban part of our portfolio has some particularly special characteristics. The space we provide is used by an incredibly diverse and dynamic customer base, often providing value-added goods and services to businesses and consumers in some of Europe's largest, most congested, and densely populated cities. And as those urban populations continue to grow, land supply in these locations is not just limited. It's shrinking, which puts even more upward pressure on rents and underlying land values and opens up opportunities for alternative high-value uses such as data centers.
This makes our portfolio unique, and it would be very hard, if not impossible, for anybody to replicate what we have, so we believe we've assembled an irreplaceable portfolio of assets, and this chart illustrates the returns we expect our portfolio to deliver. We do look at the running or the initial yield on an asset, but we're much more interested in where that yield can get to over time through active asset management. In other words, our primary focus is on total return that we expect to earn typically over a 10-year period and the risks to achieving that return, so if I just walk you through what that means on this chart.
If you start on the left with our existing portfolio of standing assets, you'll see that the, well, you'll know that the initial yield topped up is currently 4.4%, and the equivalent yield, which assumes that we cap to the reversionary potential, averages 5.4%. We've guided to recurring rental growth expectations of 2%-6%, depending on the asset type, so let's say in the 3%-4% range on average, and all else being equal, therefore, it should lead to an 8% or more unlevered return. Clearly, some assets will be high-yielding, reflecting lower growth prospects and vice versa, but overall, we're thinking in terms of at least a high single-digit unlevered return from standing assets. Our development pipeline throws off a 7%-8% yield on cost.
Depending on the volume of CapEx and the particular specifics of the projects being developed, it should add another 1%-2% of return at portfolio level on an annual basis. That results in a total unlevered return expectation at portfolio level of 9% or more. If you add the benefits of leverage with our 30% assumed LTV, that translates into an expected leverage return over 10%. Of course, that's assuming a stable yield environment. Yield compression were it to come on the back of future interest rate cuts would be additive to these returns. Plus, as I'll come on to later, we have significant further value upside from data centers.
Now, our job, beyond all the asset management, leasing, and development activity that Soumen referred to, which is needed to actually deliver these returns, is to keep actively managing the portfolio composition so that it continues to deliver. That means adding assets that contribute positively to total returns and being ruthless in divesting assets which are likely to underperform. So let me bring that alive by describing our capital allocation decisions in 2024. Let's start with asset recycling. Every property in our portfolio has an asset plan, and we constantly review those plans to identify potential underperformers in terms of future returns or risk profile. We factor in location, rental growth, covenant quality, future CapEx requirements, and a bunch of other things to rank them based on risk-adjusted returns. And the weakest ones become candidates for disposal.
We believe all our assets are good, but it's a very important discipline to continually look to bottom slice the bottom, the weakest 2% of the portfolio each year. On top of that, we always have an eye out for special situations or motivated buyers who may well pay more for an asset than its value to us, which enables us to crystallize gains and generate funds for investment into better returning opportunities. During 2024, we sold GBP 786 million of built assets and GBP 110 million of land. This was higher than in recent years, partly because the softer investment markets in 2022 and 2023 meant that we slowed down our usual disposal program. Our 2024 sales included big box assets, urban warehouses, as well as two powered shell data centers adjacent to the Slough Trading Estate, sold to an occupier and attracted premium to book value.
And although this collection of assets that we sold had delivered for us in the past with an average unlevered IRR in excess of 10%, looking forward, we expected the returns to be weaker. We also sold a plot of powered land with planning in place for a data center development. And we did that because we were offered a very, very attractive price by a hyperscaler that allowed us to record a significant surplus over book value and a 600% profit on cost. With acquisitions of standing assets, we're looking for investments that are additive in terms of quality, expected rental growth, and total return. Again, we were more active in buying standing assets during 2024 than we have been for quite a few years, as we thought pricing was attractive and there was less competition for prime, high-quality assets in this environment.
competition, so staying disciplined on pricing and return requirements is key. And the Tritax EuroBox potential acquisition, which looks to have ended well, was a good example of our discipline in that regard, and the Tritax EuroBox potential acquisition, which looks to have ended well, was a good example of our discipline in that regard. Beyond that, though, we leveraged our strong relationships and local knowledge to create some excellent buying opportunities which might not have been available to others in the market. Acquisitions totaled GBP 431 million and included four prime assets in the Netherlands, all of which have reversionary potential, which also then helped us to create a stronger geographical presence in a key target market, which has shown some of the strongest rental growth we've seen in recent years. We also acquired two urban estates in the U.K., including what we'd call a crown jewel asset in North London, neighboring our existing estate at SEGRO Park Enfield.
This is also highly reversionary, with rent reviews starting this year, which should enable us to record some early wins. The average forward-looking unlevered IRR of around 9% that we expect from the assets that we acquired in 2024 compares favorably to the expected returns from the asset disposals, and the quality is higher. And then finally, as I've mentioned, development is returning a very attractive 7%-8% yield on cost. And that yield on cost allows for all the costs of, obviously, the land, rolled up interest during the build period, and of course, the construction costs. Once the assets are completed and let, they'd typically be revalued to a stabilized yield of about 5%, which equates therefore to a 20%-30% profit on cost.
We would expect them to deliver an unlevered IRR above 10% over a 10-year hold period, which is typically how we look at it. Development is undoubtedly the most accretive use of our capital. As you can see, during 2024, we invested GBP 471 million into development, slightly lower than we'd originally planned due to the quieter pre-let markets. We would anticipate the rate of spend picking up as development volumes increase. Then looking forward, we have GBP 51 million of potential rent in our current and near-term pipeline, and we'd expect most of that to become income-producing within 12 months or so, with just GBP 190 million of capital left to spend on those completed projects. On top of that, we have a further GBP 371 million of additional rent in our remaining land bank, which we'll build out over the coming years.
Our teams have been working hard during 2024 to get these sites ready to go, progressing some large infrastructure schemes and taking projects through planning. We're expecting to spend about GBP 500 million on development in 2025, with a likely acceleration thereafter. Bringing all of these opportunities together, this is an update of the usual chart, which sets out our pathway or our bridge from today's cash-passing rents to almost GBP 1.5 billion in the coming years. You can see that on a three-year view, we have the opportunity to grow our rent by 50% through burning off rent freeze, leasing vacant space, capturing the near-term reversion, completing our current and near-term development projects, along with some others that we expect to become income-producing during that period.
If we look at the longer-term opportunity, we can more than double our rent roll over, say, the next decade, capturing the longer-term reversion and building out the rest of the development pipeline. There's additional upside from this in the form of redevelopments of existing assets. The chart doesn't factor in any further ERV growth. It doesn't include the accretive effects of our acquisition and disposal activity, the recycling I was talking about earlier, and it doesn't allow for what I'm about to tell you with regard to data centers. So let's now turn to the significant additional value creation opportunity that we have with data centers, a market where there are very significant growth forecasts for the next several years. Firstly, I want to give you a quick update on our existing data center business.
As many of you know, we've been operating in this space since 2005, which means we have a strong understanding of the market, good relationships with the major data center players, and we had a head start on most others in terms of securing power to expand the opportunity set across Europe. Today, our data center pipeline represents about GBP 55 million of headline rent, which will grow to GBP 61 million with projects under construction, so about 8 or 9% of our rent roll. The majority of that capacity sits in Slough, where we are proud owners of Europe's largest data center hub. Most of it has been developed on a powered shell basis. Quite a few of the early ones were leases of standard industrial buildings, and some were offered as ground leases of powered land.
Accordingly, the income flowing is not reflective of the underlying value of the associated allocated power, which stands at approximately 0.5 gigawatts. In other words, there's a lot of latent value buried in the ground. During 2024, we progressed our data center development program. We completed another powered shell on the trading estate, and we're on site with a further one, which we'll complete later this year. Throughout the year, we've been progressing conversations about other potential pre-lets on the estate. But most importantly, we secured a new and even more favorable simplified planning zone on the Slough Trading Estate, which gives us a blanket approval for the next 10 years to construct data centers of up to 36 meters high. As I mentioned, we sold two powered shell data centers as well as a powered land plot to hyperscalers, in both cases crystallizing some attractive profits.
But we've also invested a significant amount of time and energy into increasing our data center capabilities and our power bank across all our key markets. And we now have data center specialists in each of our geographies who are well placed to progress land sourcing, power procurement, and planning consents, the three critical ingredients for data center development, which brings me on to the next slide. Our power bank now stands at 2.3 gigawatts on the land positions that we hold in key availability zones across Europe. This is split into our existing capacity of 0.5 gigawatts, mostly in Slough, and we have a further 0.4 gigawatts of opportunities available to pre-let by 2027 and a further 0.3 gigawatts by 2030. On top of this, we're working on another 1.1 gigawatt of opportunities, mostly in the form of well-advanced grid connection applications.
There's a bunch of other earlier stage opportunities which we're working on, which will no doubt see this figure grow in the future. This power bank aligns to our existing urban footprint, and it means that we're really well placed to benefit both from cloud-driven data center requirements, but also the inference, or if you like, user interface aspects of AI-based growth. For every single one of these opportunities, we also have the land. In many cases, including obviously Slough, we already have the planning permission or at least a high degree of confidence that it will be secured in the foreseeable future. And as I mentioned at the start of the section, we have strong relationships with the world's largest data center players and an excellent track record of delivery over the 20 years that we've not been operating in this sector.
Back in June 2024, at our Investor and Analyst Day, we told you about our data center strategy for the first time. We mentioned the 1.2 gigawatts of future opportunity that we had at the time, which of course has grown significantly since then. And we explained the different ways we could approach it: powered land sale, powered shells, or fully fitted. And we explained that the powered shell model would likely be our approach to the majority of opportunities, but that we were keeping our options open. So what's changed since then? Well, the growth of the data center market and expectations of future growth have expanded significantly.
While there's been much talk about generative AI and whether DeepSeek is a good or bad thing, one aspect that we're clear on is that all of it is good for demand for data centers near to the end users, in other words, inside those core Availability Zones close to urban centers. Secondly, the battle for power and sites in the right locations has intensified, and we've come to appreciate what a rare and special opportunity we have in front of us. And thirdly, while we believe that the profit margins from building powered shells are not too dissimilar to the returns available from a fully fitted model, the likely magnitude of value creation opportunities on our powered land is just simply too great for us to ignore, bearing in mind that the volume of investment could be eight times the capital involved in a typical powered shell.
So putting it another way, we can potentially create the same value out of one fully fitted data center as we can with several powered shells. Accordingly, we are now exploring opportunities to create fully fitted data centers on at least some of our sites. We're aware that such an approach involves more operational complexity. Accordingly, it's likely that in the first instance, our foray into the fully fitted model will be done in partnership with others who bring the track record and experience of doing so. We've got some very interesting and active conversations going on in this regard at the moment, and we'll look forward to updating you in due course. But it's a very exciting opportunity. So to conclude, SEGRO has delivered further growth in rents and earnings during 2024 and has made good progress against our Responsible SEGRO commitments.
We're primed for further growth from our existing business in 2025 and beyond, and we're encouraged by the pickup in occupier activity that we've seen of late. And we're incredibly excited about the additional opportunity we have to create value in the data center sector. So thank you for your attention. We'll now take questions. I think, as usual, we'll start in the room, and then we'll go to the webcast. But remember, if you're in the room, you need to pull out the microphone and press go.
All right. Can you hear me? Bart Gys, Morgan Stanley. Big announcement on data centers. Could you please help us understand what that means for CapEx, what that means for rents? How much of the GBP 500 million a year do you intend to spend on that and kind of help us translate a bit kind of these ambitious plans into the financials?
Then secondly, as a follow-up, some of your peers in continental Europe that have embarked on a more fully fitted model seem to suggest this is largely a speculative exercise where you can take bookings, but you don't really sign a pre-let, and therefore it's riskier from that perspective as well. Your slide at the end seemed to suggest this is a pre-let strategy. Can you provide color on that? Thank you.
Yeah. Okay. There's quite a bit in there. We'll show now and try and cover that between us. First thing to say on leasing strategy, I think we said back in June last year, you can do spec or pre-let for those fully fitted data centers.
It's very much our intention to be pre-let-led because of the location of our sites in really high-demand core availability zones, so close to the urban centers where there is tremendous demand for that space. We would expect most, if not all, of our developments to be very much pre-let-led, and there's good precedents and examples of that being the case. In terms of the capital, I mean, it's really difficult to give you some very precise numbers. I mean, firstly, because we haven't yet signed up our first fully fitted deal. And frankly, every situation is going to be very different. The cost of the land, the cost of getting the power, the nature of the building, what you're trying to get out of it, they do vary enormously, but it just gives you a very, very broad guide.
If a powered shell, let's say a 50-megawatt powered shell, could be anywhere between GBP 50-100 million of construction cost on top of the land, you can multiply that by six, eight, ten times, depending on the particular configuration in terms of the amount of capital for the same data center if we're doing fully fitted. We've indicated that the yield on cost is 8%-12%. Again, it will vary according to the particular situation. So were that to apply, if you take that to apply to any one opportunity, it's a significant ramp-up in income and CapEx requirements. But time will tell as to the particular characteristics of the first one we set out. So we're generally not giving a lot of specific guidance. When we announce our first project, we will share more details about what that entails.
But in terms of your question around CapEx, and Soumen will correct me if I'm wrong, the GBP 500 million guidance we've given for 2025 does not make any allowance for a fully fitted data center. That would be incremental over and above that. And look, we haven't done our first one. If you extrapolate and do it over the whole portfolio of opportunities we've got, it's a huge amount of capital. But we'll cross that bridge when we get there. In the meanwhile, we've got plenty of capital and liquidity to do the first one or two, and we'll update you in terms of funding thereafter. Thanks. Rob. Thanks. Two and a half questions, if that's okay. I meant to say you're only allowed one, but we'll let you go with the first two. We'll go down to two.
The pickup in occupier market activity or occupier activity that you talk about in the last couple of months of the year, I wonder if you could quantify that. Is that number of leases signed, number of conversations that the leasing teams are logging in their CRM system internally? It would be helpful to kind of get some kind of quantitative thought around that, if that's okay. It would. And I'm not sure I can give you the very precise quantitative numbers, but I think the way you characterize it is interesting, actually. So what we saw, and you saw it on Soumen's chart where he showed all those dots, what you can take away from that is there were a lot of deals done. A lot of things got pushed over the line in the last part of the year.
I would say particularly in urban markets, London and Southeast was very active. We were quite active on the continent with some of our light industrial schemes in Germany, for example. So there were actual deals done in the last part of the year that gave us quite a, well, that's quite interesting, that people, having general theme of last year being people sitting on their hands for a long time before taking commitments, people actually put pen to paper in numbers in the last part of the year, and this year, it's more around conversations, anecdotal feedback, general sentiment, but it's interesting that we're not just hearing it in one or two places. Everywhere across the business is seeing a lot more activity in terms of looking at deals, whether they're leasing up existing space or pre-let.
It would be silly given the macro environment we're in for us to say, therefore, it's onwards only upwards from here because you just don't know what the macro is going to throw at us. But at the start of the year, we're feeling pretty positive about what we're seeing. Thanks. That one or two. And then the second one was you rightly talk about selling assets that don't meet your forward IRR requirements. I don't know if you could share of the assets that you sold, say, last year, for example, what you thought the forward IRR of those disposals on average would have been, just so we can get a flavor of how that compares to, say, the 9%. Yeah. I mean, in broad terms, and I talked about the average portfolio looking for about 8% plus on standing assets.
We acquired what we think was 9 plus. The disposals, they delivered, I forget the number, about 11% on average to date. We thought the forward look was more like 7%, 7%-8%. Thank you. Great. Thank you. It's Maris Passeau here from Bernstein. May I ask two questions as well? Yes, sure. Maybe we start with the new order. We maybe start with the data center again. Could you maybe give us a bit of detail in terms of the partnership model you're looking at or thinking of? I appreciate it's probably very early days, but is this a partnership model in the sense that you're using expertise for the fit out, or is it a partnership model where you could potentially share the CapEx requirement? Yeah. We will answer that question in more detail when we get there.
But broadly, yes to both could be the case. We're partnering with someone who brings expertise, but also the ability to co-fund. So that's most likely the way we'll go, but we haven't absolutely definitively decided that's the only way to go. Definitely, we want the expertise. I mean, you can look at what's involved in fitting out a data center. It's not that complex. It's M&E, and it's pretty, there's a very, very good supply chain of people that can do this. But for us to go to, for example, a hyperscaler or a major data center operator and say, "We've hired some people, we can do it," that'll take a while to get people over that confidence level. So that's why doing it in partnership with someone who's tried and tested, got a track record, we think is a quicker route to the opportunity. Okay. Thank you.
And then just following up on Rob's question on kind of occupancy levels and demand, I see your urban portfolio in the U.K. is now close to a 10% vacancy level. But you're mentioning that you've got a bit more demand there and have done some active leasing. Can you maybe break that out into kind of what's been done at the start of this year, what's been taken back for refurbishment, similar to the pie chart you gave at the update last year? Thank you. Yeah. I mean, so James Craddock, who's our U.K. MD, is here. Why don't you give a bit of colour in terms of what's in that U.K. urban vacancy? Yeah, no, sure. You'll need to pull the thing out. I don't know how much you might tell. Oh, okay. Yeah, no, absolutely.
As David's already mentioned, I mean, managing the urban portfolio is different from our big box. It's more intensive from an asset management perspective. And that means that obviously, we do see periods of elevated vacancy, particularly as we bring assets through the refurbishment and the redevelopment cycle. So as it stands at the moment, our vacancy is higher than we would like, but the teams are working hard to bring that down. And I think on a positive note, if you look at our urban vacancy, about 25% of that is either under offer or in active and advanced negotiations for leasing. And also, we have about 30% that is either about to go into or is in refurbishment at the current time. So obviously, that's going to come through.
As you picked up from the presentation, we still remain very positive about the structural drivers which support urban in the medium and longer term. I mean, and a good chunk of that UK urban vacancy is going through refurbishment or redevelopment as well. So it's not really effectively available for lease until we get further down the line. What was interesting, and I think Soumen used that example of where we've taken some space back, we've refurbished it. Obviously, it's non-income producing for a while and now it's a vacancy, but the uplift in rents when we can create modern, sustainable space in the right location is pretty substantial. Obviously, proving those higher rental levels is very important. Yes. Zachary. Morning, and thanks for the presentation. A couple from me, but they are linked, so I'll ask them at the same time.
You showed a chart earlier in the presentation that showed, I think, the leasing done by existing space versus development space or new space, and the proportion of development space was, I think, the lowest since 2016. Obviously, your retention rate is still higher, 80%. So I sort of take that to mean that actually occupiers are happy in their existing space and are staying more than necessarily moving to newer or even larger space. If that is the case, when do you see that potentially changing, and where do you sort of want or expect to see your CapEx guidance get to as and when that demand for new space picks up?
And then sort of related to that, do you see any potential impact from tariffs, which I know is very unpredictable, but let's say we did enter a worst-case scenario and a more sort of full-blown trade war impacting exports? Do you see any impact on your logistics occupancy, but particularly around Heathrow Airport? Gosh, there's quite a lot there. So just deal with the tariff point, first of all. If you look at our portfolio, we are not massively linked to global trade. Remember, two-thirds of our portfolio is urban. And most of our logistics space, with a few exceptions, is around supporting inward consumption. You chat to, maybe sadly from a U.K. perspective, you chat to some of the rail operators at somewhere like East Midlands Gate, where we've got a rail-connected facility. They talk about the containers.
They come in full, and they leave empty because we're not shipping stuff out. And that is true for most of our portfolio. It's mostly around supporting inward consumption by UK businesses and consumers in and around the most densely populated urban areas. We've got a little bit more trade flow dependency in some of the European markets, but it's a relatively small part. And we're not massively exposed to traditional auto, the auto sector in Germany, for example. Again, it's mostly about small or medium-sized businesses, very diverse, supporting these, I would say, wealthy, congested population centres. So we're not complacent and don't ignore it, but I don't think we're not particularly concerned in our case about exposure to tariffs and potential trade wars.
In terms of, I think, your earlier question about pre-let and development volumes, where we would like to get to, there definitely was a case, by the way. You're right. A lot of occupiers were sitting on their hands rather than moving. We like to have a bit of churn. We want churn, particularly in our urban portfolio, and the concept you sort of touched on in that slide that I only skipped through briefly, but big box tends to be longer let, low asset management intensity. People stay there, and so the returns from a big box are much more development oriented, plus if you like, low-cost income on those boxes. Whereas the urban piece is much more intensively managed, lots more churn. We like to have some churn, not too much. We've got a big portfolio, so we've always got opportunities to move things around.
I think we're happy with the churn. We'd like to see, as James said, we'd like to see the vacancy rate reduce and fill up some of those units as we particularly bring that refurbished space back onto the market. But I think what's really missing to sort of, if you like, even out those two parts of that lettings graph is just a bit more volume in the pre-let market. And that's mainly going to be around logistics and possibly data centers. And that's where it's too early to say because these projects, they're massive capital investments for the occupier. If you're building a million square foot, which is not uncommon, it's a huge technology investment as well as a real estate investment. It's usually part of a much wider reconfiguration of a supply chain.
And so I think what we've seen is during the last couple of years, even though these are long-term strategic projects, a lot of occupiers have just been a little bit more cautious about putting pen to paper and committing to those projects. What I hope we're going to start to see soon is that there's this pent-up demand building, and therefore it gets released into 2025 and 2026 as people say, "Okay, the wheel's not going to stop turning. We do need to put in place these plans. We've held back long enough. Now's the time to start committing." So hopefully, we'll see that pick up. We've slightly anticipated that with our own spec development. Soumen mentioned it.
We're doing a bit more spec actually in urban markets, but also a couple of spec schemes in big boxes in Germany because we just think we've got a really good product and very little supply and good underlying demand. Sure. Thanks. Morning, it's Tom Musson at HSBC. Just a couple of questions. First one on the new data center fully fitted ambition. The 8%-12% yield on cost is sort of the same as you reported versus the powered shell approach. Should we expect capital returns and therefore total returns to be higher from the fully fitted ambition? Because I suppose otherwise it's optically a very similar return, but I suspect that the fully fitted is clearly higher cost of capital. Yes. We're obviously using the same. We're using our cost of capital.
The main attraction, you're right, and it's hard to generalize because 8-12 is quite a range. As I said earlier, a great variety of different projects. But fundamentally, 8%-12% on a much larger amount of capital is whatever you capitalize the finished income at. And probably a fully fitted income-producing data center will be a slightly higher yield when it's finished than, say, just the core shell real estate. But nonetheless, the volume of capital deployment and therefore the volume of capital upside, the value creation is just a very much bigger number if you do fully fitted. Do you want to add anything to that, Soumen? The only thing I was going to add is, and I was going to be very careful because a lot depends on the specific situation, depends on the cost of land.
And particularly, we're talking about doing these in the core availability zones, which means kind of in and around the urban areas where the land price is going to be higher than trying to do it in parts of Europe where, frankly, there's very few people live. But taking all those caveats into account, you're looking at profit on cost north of 50% and probably north of 60% or 70% in some cases. So that's very, very firm. When David talks about 50, maybe 100 on a powered shell, but 8-10 times that on a fully fitted basis, that's a phenomenal absolute profit and percentage profit if we can deliver it on those cases. Makes sense. Thanks. If I could just ask a second one.
Just within the property disclosures, I think there's GBP 85 million or so of rent due to expire or break within the next 12 months. If we sort of broadly assume stable retention rates, does that create a slight occupancy headwind in 2025? I think as I hopefully was trying to say during my presentation, I mean, David, do that. One of the things I think this business has done terrifically well over the past 24 months is capture that reversion whilst keeping the retention rate really high, and where there's been opportunity to relet, either potentially directly or through a refurbishment and then reletting, we've done that terrifically well as well, so in our mind, we've got terrific real estate, as David talked about. It's part of our model, which is to own really good estates, really good space, and capture the like-for-like growth.
And I think we can see really strong growth coming through here. So I've got to say, it's been a concern that has been voiced by people to us. But I've said, if you look at the metrics, you look at our like-for-like capture and all the things that kind of go around it, I don't feel particularly exposed. I don't know if Marco, James, you wanted to add anything? No. Any more questions? Right. We go to the conference line next, I think. Oh, sorry. I was wanting to go back, but we'll come back. We'll come back to that in a second. Okay. One more in the room. Sorry. Hi. Good morning. This is Suraj Goyal from Green Street. Just one question from me on the occupier market.
So it looks like the development margins are pretty reasonably attractive across the board and maybe suggest a slight pickup in development activity in the sort of near midterm. Then if you couple that with sort of vacancy potentially having peaked, would you say that if we do start to see that pickup, the market would be well placed to absorb the development? Would you still sort of expect sort of ERV growth of 3%? Yeah. I mean, we've for quite a while had this guidance range of 2-6% for ERV growth, which, let's say, averages 3-4% depending on mix. Slightly stronger in urban than in big box. We're comfortable that is a very long-term sustainable number. Very difficult to predict in short-term periods what it's going to be.
We clearly massively blew out those numbers in the pandemic and slightly softer since then, but we're still in the right range. We're comfortable that we should at least do those numbers because, as I've been saying pretty consistently through the presentation and previously, they're just not making a lot of land. Land is not easily released, whether it's greenbelt or even in urban where you've got these competing uses. So yeah, we're pretty comfortable that the run rate of take-up of new logistics space will support a return to some much higher development volumes for logistics. And urban, it's all around, can we get the sites? Can we position them? Can we make the product available? Because in many of the urban markets, particularly in Europe, we're creating a product that just doesn't exist. We're introducing a new product.
A lot of German businesses own their own very inefficient space. So when we go into market, we get a piece of brownfield land, we redevelop it, we create a secure, well-run, well-laid-out warehouse park, they're getting a much more efficient operation from it. So that's why we're very confident that there's lots of untapped demand still to go there. I hope that covered them, hopefully. Thanks. Right. Let's go to the conference line, please. The first question from the phone comes from the line of Frédéric Renard. Please go ahead. Hi. Good morning, guys. Thank you for the presentation. Just to come back on the data center opportunity. So you showed a fully fitted model, of course, and the yield is the same than what you could achieve on a powered shell model. But my question would be, what about the depreciation for the fully fitted model?
Should we look at the 8%-12% as a gross yield versus a net yield for the other model? That would be the first question. And then on the second question, you still depict a very good picture for London, but as was pointed out, your vacancy rate is around 10% now. What does it impact in terms of your rent incentive? Do you give a bit more rent incentive to your tenant? Thank you. Soumen, maybe I'll just answer the vacancy one and maybe an incentive. Maybe you can pick up on the first bit. I think if you look at our incentive packages, they haven't really moved very much. I think it's moved up less than 1%. So we're still seeing incentives around 6% or 7%. It does vary a little bit by market.
But no, one of the pleasing things is that where we have space and we've got an occupier lined up to take it, people are generally willing to pay the asking price. So very, very slight increase in incentives, but nothing very material. And encouragingly, we've been able to prove and push on rental levels with the deals that we've been doing in 2024, and I'm sure that will continue in 2025. Soumen. Yes, Fred. If I try and answer what I think was the question. On the fully fitted model, what we're talking about doing is fitting out things like the chillers, the backup generators. We're not buying the tech. We're not putting the tech in. And as one of the earlier questions, I think, asked, we're looking to then lease this or pre-lease this in its entirety on a triple-net basis.
So we are underwriting a wider yield than a pure powered shell to take into account some of that slightly different risk profile. But it's still a net yield to us. There's nothing else that we'd be taking into account. Was that the question? Yes, that was the question. Thanks. Thank you. The next question comes from Jonathan Kownator from Goldman Sachs. Please go ahead. Good morning. Thank you for taking my questions, too, if I may. Just to go back on DCs as well. On the fully fitted model, obviously, you've highlighted on slide 32 that you had three pools of assets you had where those are secured with power before 2027, the reserves until 2030, and then the ones where you have power later than that. Can you help us understand where the fully fitted fits in in terms of timing given those timelines? Is it before 2027?
Is it where you have power between 2027 and 2030, or is it post-2030 that you expect to do these fitted? What is the first project in fitted? Well, in broad terms, we are looking at current near-term projects for fully fitted. So it's not a question of waiting to 2030 and beyond. We've got some particular projects and opportunities we're looking at now which would be similar to the timeframe for a powered shell. The only thing I would just highlight, just to mention and give a little more color on that is clearly, if you do a fully fitted, it's a longer journey from when we start construction to actually completing the fit-out as well as the building, and therefore when the project becomes income-producing.
So you're going to be adding 12 or 18 months to the construction time, at least, I would say, before you get your income. But again, this is all about scale of the value creation opportunity rather than the near-term income impacts. And does that mean that it's a way for you to accelerate the CapEx that you're putting into data centers, or is it just a question of demand for powered shell that is perhaps not as strong as what you thought initially? Sorry, say that last bit again. The question is, is it a way for you to accelerate CapEx, or is it you're shifting to fully fitted because demand for powered shell is not as strong as you thought initially?
I think, as I said during the presentation, the real reason why we're looking at this is that it's an opportunity to create a lot more value in totality. It's fair to say that fully fitted is a more flexible, I'd say, doable model across more European markets. We've found that powered shell demand is exceptionally strong in Slough because, frankly, we've got such a unique replaceable asset there that anybody who wants to plug into the network needs to be there, and powered shell is all we've been willing to offer. In other markets across Europe, it's a bit patchier. Given a choice, most operators would rather own their facility. And there's probably a bit less demand for the powered shell model. It's not impossible, but there's less demand for powered shell.
So it's another reason why the fully fitted model may well be the best route and the quickest route to extracting value from the opportunity set we have. Okay. Very clear. One follow-up just on the sort of take-up and its absorption. Obviously, you had strong take-up this year, but there was also space return in equal number to take-up on the existing space. Can you perhaps give a bit of color on the space being returned? Is it from 3PLs? What type of operators? And I understand also it's not from space taken back from redevelopment, if I read your footnotes, right? Yeah. Sure. No. Soumen covered that in part of his presentation. But why don't you just mention that again, the take-back? Yeah. Sort of.
I mean, look, if there was one particular example in terms of the take-back, which I referred to a global media company that chose to give us back their space in North London, which actually we see as a really interesting opportunity. It's three buildings. We've had multiple viewings and would hope to set some interesting rental levels on it this year. But that, in particular, accounted for about 50 basis points of vacancy. So almost over half the step-up in last year's vacancy. Otherwise, if you kind of get beneath the kind of special cases, there was nothing in particular that was different about that. But there was one special case that affected last year. That was particularly big. Yeah. And I mean, there's also the MatchesFashion unit that came back in Park Royal. So Premier Park Matchers Fashion went insolvent.
That unit's come back to us. That's a great opportunity for redevelopment. So if you actually took out that Matches Fashion unit and the studio operator that Soumen was alluding to, I think you'd say that the level of take-back was pretty normal and very similar to last to the year before. Okay. Very clear. Thank you. Okay. Two more. The next question comes from Zheng Wang from Kempen & Co. Please go ahead. Hi. Good morning. Thank you for taking my questions. You were just talking about the strong sentiment and leasing activity. At the same time, your balance sheet has much more headroom now. But your expected development CapEx is not materially changed compared to, say, last year, despite the relatively short lead times in developments. Could you perhaps elaborate on that? Soumen, do you want to cover that? Yeah.
I think we tried to talk about, the balance sheet has got the capacity to grow. We've got GBP 2 billion plus of available liquidity. In terms of the development pipeline, as I think I talked about, it's a mixture of things. There are fewer large, very, very large prelets around. There are deals we've done. We've, say, we signed one of the largest prelets in the UK last year at Northampton. And we're looking at several now. The investor committee in January, we looked at several. But we have to say that the time from inquiry through to signing is longer than we might have liked. And it is just a feature of the market over the last 18 months or so.
We have started spec schemes because particularly where we see pockets of demand on our standing estates, which makes us feel very confident about starting, particularly on the urban side, but a couple of big box schemes as David mentioned, and so where we are, we are taking that risk-ground spec to kind of keep those development volumes high, but really, we do want those to convert more of the inquiries into actual prelets through the course of this year, but we're positive around the sentiment change that we saw two to three months ago. I think the reality is, even if the sentiment converts into more prelets that we sign this year, which we hope will be the case, the likelihood is the amount of time on site this year will be relatively limited, so the ability to spend significant amounts of additional capital this year, it's limited.
So we think hopefully demand will ramp up. We'll sign more pre-lets this year. But the CapEx will start ticking in soon after then, and we'll most likely impact 2026 rather than 2025. Okay. That's clear. And then just one more on data centers. I appreciate the difference in absolute investment volume for the fully fitted approach, but with similar returns. It suggests that there's not much of a difference in risk premium assigned to, say, execution or obsolescence risk of the fit-out. What's your thought on this? Well, my thought is that we're going to approach this the way we approach every capital investment decision, which is we're looking for an attractive risk-adjusted return. And without getting into the specifics of a deal that we can't talk to you about just yet, there's no point in speculating about that.
But what we can say is we're looking at the returns. We're looking at an evaluation of the risks involved, very clear what sort of return we want to make. And the overall value creation opportunity is sufficiently large that this is getting some serious energy and attention. And we'll definitely share more details when we can. But I don't think there is a one-size-fits-all answer to that question. Any more questions? The next question comes from Paul May from Barclays. Please go ahead. Paul, just a couple of ones from me. It's a bit similar to what we've had. But a lot of the commentary around the outlook seems very positive, but some of the overall operational figures, for example, vacancy higher since year-end 2021, over every period, and is lower than Q3 2024. Like-for-like growth has been slowing. Pre-lets are at the lowest level.
I think I can recall them maybe ever, probably point to not being quite as strong. Given market vacancy rates are expected to grow and take-up is weak year on year, I think CBRE has prime rents in London warehouses flat since Q2 2023. Just wondered, what should we take away from the numbers? What numbers should we be looking at to give us confidence in the operational outlook that you have? Because some of the data would suggest potentially slightly otherwise in terms of the overall numbers and the headline numbers. I've got a second question after that. Thank you. Soumen, do you want to? I think so. Well, let me start. Look, I think we've sort of talked a couple of times around the occupied market this morning. The reality is we've had a normalization in the market since 2021 and 2022.
I think regarding that, the benchmark is the wrong thing in our view. When you kind of look at the conditions in 2024, I showed it on the chart where I showed you the kind of difference in activity levels by month. We had months that were really strong, months that were less strong. I think we take it actually as a positive. In a year like that, we delivered very healthy levels of leasing activity, really healthy levels of reversion capture. And importantly, in terms of momentum into 2025, we finished the year really strongly. So all of that, to our minds, shows, look, it's a market that is kind of working through some of what happened in the pandemic, but it's in good shape to be able to kind of recover from here.
And as I've said a couple of times, if we can convert some of the inquiry levels into deal activity, we can actually start to drive kind of those development volumes pretty strongly. And if we performed as well as we did last year, it doesn't take a lot to change it. So I think we just could be realistic in terms of it's a market where there's things to do, and we've just got to make sure that we convert those opportunities that we can find. Yeah. I mean, and the only other thing I'll just add, Paul, to that is, again, remember, most of the market data you look at will be market-wide big box logistics. Two-thirds of our business is not that. And I don't know whether you've got better sources. We find it very difficult to get market-wide data on urban markets.
So what we're telling you is what we're seeing and what's coming from our conversations with occupiers and local agents and our teams on the ground. The proof of the pudding will be in the eating. We'll see what happens. But our expectation is that occupancy will improve as we go through this year. Okay. So just to say that last bit, some occupancy rates improving through the year and prelets also improving through the year. Would that be the expectation? We're not going to give you a forecast of our year-end vacancy rate. Good try, though. Okay. And sorry, just the second question was on the leverage and the balance sheet. I think you've worked hard to get your net debt to EBITDA down towards a reasonable level in a European context or high on a global context.
Just wondered, what gives you the confidence that this should move higher? Because obviously, if you are investing, that will move higher. And just wondered what your thoughts were around that and how much the net debt to the EBITDA is a focus, or is it all about LTV, which has become less relevant, at least in our conversations with investors? Thank you. Sure. Look, I'll give you an answer that I think I've probably given for several years, which is there is no single one debt metric that is the right one. You've got to look at them all. You've got to look at loan to value. You've got to look at the debt to EBITDA, which effectively is the debt yield. And you've got to look at the interest cover, which frankly, people stopped looking at five or so years ago. Our covenants are all set against LTV.
So frankly, that is the relevant one to kind of focus on. I think if you look at kind of where we are rated, you look at kind of where we're rated against the European population. I've said this before. You can take your pick. You can have a high loan-to-value and a lower net debt to EBITDA, or you can have our balance sheet, which is 28% levered and an 8.5% net debt to EBITDA. I know which one I would pick every day of the week because a high yield is no sign of a good set of assets, and it's the sustainability of that leverage in the balance sheet that matters. David talked about kind of the approach we have around the portfolio management. It's the same thing on the balance sheet.
And so just having lots of high-yielding assets that artificially depress your net debt to EBITDA doesn't really get you very far. And the companies in Europe that have, outside of our sector in fairness, that have lower net debt to EBITDA, I would question whether those loan-to-values are sustainable or not. So I think it's not a simple answer, Paul. We look at all of them, but I'm pretty comfortable as to where we sit in terms of the balance sheet today. And it's got more than enough funding available for what we've got in front of us. And hopefully, as David says, hopefully, things do start to pick up. We've got the firepower we need to capitalize on that. I agree. It's multiple factors. The question was, at 8.6, you mentioned that you're happy today.
Would you be happy seeing that 8.6 move to 9, move to 9.5, move to 10 again? Or would you prefer to keep that around 8.6, in which case it implies there's actually not that much funding available or leverage available to invest? We've been pretty clear that we like our A rating. And that requires a net debt to EBITDA around the nine times area, plus or minus half a turn. Now, that's quite a lot of flexibility, particularly when you've got the EBITDA growing through that period as well. So I think, as I said, you've got to kind of look at the whole thing and flex every piece of it to really understand what the firepower and the opportunity set looks like. Yeah. And I don't talk to the rating agency the way Soumen does.
So if you were close enough, you might kick me under the table here. But I think just to add a little bit more color on that, we're certainly not planning on having our net debt to EBITDA go in the wrong direction by, for example, buying a whole load of land that's non-incrementally producing and sticking it on the balance sheet. If it were to go up temporarily because we happen to have signed up a lot of prelets and it caused a temporary increase while we're waiting for those prelets to become incrementally producing, that would be a high-quality problem to have and not something I'd be too worried about. I'm not sure about you, Soumen. No, no, exactly right.
As I said, the EBITDA is the thing to keep an eye on here, which is the thing that, for all the things we've talked about, is a huge opportunity to capture. Yeah. Okay. Any more questions? And the last question for today's call comes from Paul Gorrie from Citi Please go ahead. Hi, all. Just one more follow-up on the data centers from me. The slide 32 with the breakdown of the gigawatts is very helpful. Can I just ask on the split between secured, reserved, and applications in progress, how much of the kind of reserved and applications in progress is dependent on, say, the Iver upgrade or other kind of infrastructure products or projects, I should say? And how much is just kind of you going through the process of applying for power?
I guess I'm trying to gauge how much is kind of in your control and how much there's maybe external factors that impact delay, make it more difficult on the power side. Yeah, Paul, it's a great question, and the answer is there's a whole mix. I mean, most people who know in this who are in this sort of power business, it's always a moving target. To your specific point, some of that capacity is related to Slough and the Iver upgrade. If you're looking for something more encouraging on that front, we know that this week the planning permission for the Iver upgrade was granted by Buckinghamshire Council, which is the first time we've actually had a firm date in mind. So that will be some of that power is related. So it's secured. We have it.
We're right in the right place in the queue for that. We now need that upgrade to be delivered, which will happen before 2030, is our current understanding, and then there's a whole bunch of other ones that are at various different stages of certainty around the group. Okay. Yeah, that's helpful. Thanks. Okay. Claire's got some on the webcast, I think, or on the webline, web chat. I'll try and group these in the interest of time, so starting on the occupier market, how do you explain such a high level of reversion? What do you expect for the next year? Could we expect to see the same reversion recapture level? Well, we've sent out a slide showing the breakdown by year, so that is what's available to us. It will potentially slightly increase as ERV growth continues.
But we've given a slide which gives you the breakdown of ERV available through reversion capture. Thank you. And then I think that's working out. On development CapEx, can you give the split between on-site and new constructions on the GBP 500 million of development CapEx? So of the GBP 500 million, about GBP 150 million of that is infrastructure, we think, this year. And of the rest, it's probably half and half in terms of stuff we already had on-site and half that's started. Thank you. A couple more questions on data centers. Would you be looking to transfer the data center assets into a separate subsidiary or fund something like you've done with SELP? Possibly, but we'll cross that bridge when we get closer to it. And then another one, DCs. Can you explain a little more about what you mean by the increased operational risk associated with data centers?
We didn't talk about operational risk. I think what we're talking about is just the additional complexity associated with installing all of that equipment, and that's something that we don't think is too difficult, but we do think in terms of credibility with customers, we're right to be doing that with people who have got a tried and tested track record. Thank you. On solar, what percentage of the roof space is covered with solar at the moment? When you enter into it, when you install it, do you enter into agreements with customers to buy that, and what's the yield on cost for those projects? Look, it varies enormously. We've got all kinds of things. I can't tell you. It's quite a low percentage of the total roof space that we have, probably 10% or 15% in totality.
What we've been doing, because we've got a very large established portfolio, we've been doing over the last few years is retrofitting solar where we can. That's not always straightforward because in the UK, the occupier is responsible for the roof, so you can't interfere with that. On the continent, it's a bit easier. Some markets, it's very easy to supply the excess power you generate into the grid and get a sensible price. In other markets, it's not. So it's very hard to give you a one-size-fits-all. But we would say we'll be looking for a high single-digit yield on cost where we do this. We're looking to do it more actively on all new developments, again, where there's an attractive feed-in tariff available from the local marketplace. So I think you'll just see it continue to grow, but it's offering a sensible return.
Some cases, we do a power purchase agreement, and the customer pays for the energy. In other cases, we simply charge them an additional rent, a flat rate for it. So it's a very wide variety of different practices. But we're very pleased that we were able to double the amount of solar we had installed in the portfolio during 2024. We've now got 123 megawatts, and we'll be looking to add to that as we go forwards. Thank you. Two last ones. One for you, Soumen, on capitalized interest. Can you confirm what rate you are using to capitalize interest? Yeah, of course. So we use a variety of rates between kind of 3.5% and 5.5%.
As we talked about probably two years ago now, you capitalize the marginal rate at the point at which you start a project, and that's then locked in for the duration of that project. As we go forward from here, obviously, with rates kind of, well, in euros certainly are a bit lower, I'd expect that number to come down over the next couple of years. Thank you. And one final question. Given the amount of cash and capacity on the balance sheet, would you consider doing share buybacks? Look, I think David's talked about our approach to disciplined capital allocation. So obviously, we have to have all options on the table at all times. Having said that, David's also laid out, I think, a really exciting array of opportunities both across the development book on industrial logistics and a phenomenally exciting opportunity on data centers.
I think with that in mind, I think we find that using capital to buyback shares would be the wrong option to take at this point in time. Investing that capital into those opportunities, we think, will create the better value for shareholders going forward. Okay. Well, thank you again, everybody, for joining us here or online. Have a great day and enjoy the weekend when you get to it. Personally, I can't wait.