Pure Data Centres Group Limited (SGRO) Earnings Call Transcript & Summary

Good morning, everybody. Thanks for joining us this morning at short notice today. Let's move on one slide. So some quick intros before we get started. I'm delighted today to be joined by Bill Davis, Chief Investment Officer of Pure Data Centers. And for the Q&A session, we'll also be joined by Andrew Pilsworth, our Chief of Staff, who, amongst other things, leads our Data Center Strategy; as well as Soumen Das, our CFO. Next slide, please. Before we get into the transaction we announced this morning, I'd like to quickly recap some of the things that we told you with our full year results presentation just a few weeks ago. Firstly, a reminder of our track record in data centers. We've been operating in this market since 2005, and we have a good understanding of the business as well as strong relationships with the major participants. It represents about 8% of our current portfolio by rent. We have 34 data center assets in the U.K., mostly in Slough, which, as you know, is Europe's largest hub of data centers. Our strategy to date has mostly been to deliver powered shells and our announcement today marks the evolution of that strategy, the rationale for which I'll come on to shortly. On to Slide 4, please. We also told you about our 2.3 gigawatt land-enabled power bank focused on key availability zones, which we believe is one of the largest pipelines in Europe. And we gave you a rough time line for the delivery of that power capacity. Our focus on major urban markets means that we are well placed to benefit both from cloud-driven requirements as well as the inference aspects of AI growth. And this opportunity will likely be accelerated by the emergence of technologies like DeepSeek, which will democratize and accelerate the adoption of AI. We believe that demand for data center space in these core markets will outstrip supply for many years to come. And so we're confident that we have a significant and deliverable value creation opportunity with our exceptional power bank and related land positions. On to Slide 5. As we've also discussed before, there are a number of routes that a real estate data center strategy can take, and they're illustrated on this slide. Our focus to date has been on powered shells, although we've also sold data center land. Given the expected market growth, plus the rarity of the land and power positions that we control, we've been considering how best to capitalize on the opportunity in front of us and specifically whether SEGRO can efficiently capture a larger slice of the value creation pie. The announcement today formalizes the evolution of our strategy to include the development of fully fitted data centers on certain sites. We believe this will enable us to deploy capital at scale and generate returns over and above most other investment opportunities available to us. There are some risks involved in moving up the value curve, but we intend to carefully manage and mitigate these risks through initially at least working in partnership with experienced operators as we're doing with this first project. So today's announcement does not mean that everything we do from here on will be fully fitted. We intend to remain agile, and we'll pursue whichever model offers the most attractive opportunity and risk-adjusted returns for each site. So you're likely to see us deploy a mix of powered shells, fully-fitted data centers and in some cases, powered land sales. Next slide, please. Let's move on to this morning's announcement, the formation of a joint venture to develop our first fully fitted data center. It's a 50-50 JV with Pure Data Center group, a company with over a decade of experience and a track record of delivering data centers to the most sophisticated users. The innovative partnership combines a 10-acre site that we own in Park Royal with 70 MVA of power available today that Pure have secured. Through this partnership, we intend to develop a data center with a 56 megawatt of IT load. And we expect to pre-lease it to a major hyperscaler as this site is in a key London availability zone, where land and power are severely constrained and where we know hyperscalers are focused on growing their capacity. This partnership provides us with the technical expertise to embark on a fully-fitted project, and it will help us further strengthen our own data center expertise and knowledge. Now before going into more details on the project, I'm going to hand over, first of all, to Bill to tell you a little bit more about Pure and to provide some color on the market.

Terrific. Next slide, please. First off, thank you, David. I just want to say how excited we are to announce this joint venture and excited we are to be here with you. So just very briefly about Pure. We are a hyperscale data center owner, developer and operator with a specific focus on core cloud infrastructure and in particular, fully fitted and operational build-outs. As David mentioned, the company has been in existence for approximately 10 years. We have 500 megawatts of capacity either online or in development in core markets, primarily in Europe with additional focuses on the GCC and APAC. We are a portfolio company of Oaktree Capital Management, one of the world's largest alternative asset allocators with approximately 250 people across the world. Next slide, please. So fundamentally, this is the backdrop for why we're so excited about this. London is the leading data center market within Europe and one of the very largest markets globally. The market for public cloud has roughly tripled over the past 5 years. And frankly, it's not slowing down at all. We anticipate it to grow another roughly 2.5x over the next 5 years. It is -- as I'm sure most of you have read, it is a key growth sector being backed not just by corporate interest, but by the U.K. government as well. And it was just worth really pushing on data centers are a form of real estate, and so the specific location and submarket really matters. While there are a lot of themes in the sector, this site is beautifully positioned for core cloud infrastructure, where there is a very significant supply-demand imbalance. And so again, it gives us a great deal of conviction about the execution. And just by way of example, you can look at some of the quotes from some of the biggest technology companies in the world, the level of commitment they're publicly making to this market that this site is, again, exceptionally well positioned to capture.

Okay. Thank you very much, Bill. Let's go to the next slide then, please, Slide 9. Thanks a lot, Bill. So given those attractive market dynamics that Bill has just described, we're really excited about this opportunity that we've created together in West London. And we're thrilled to be working with Pure, a business that we've known and admired for some time, which has the approach, capabilities and experience to be a perfect partner for us. The trigger for this first project together is that each partner is bringing a special ingredient. We have the land and Pure has the power. And both of these are rare commodities in a tightly supplied availability zone, such as the Park Royal availability zone. So the 10-acre site on which the data center will be built is owned by SEGRO and sits on the edge of our Premier Park estate. It currently houses an older warehouse recently vacated by Matchesfashion, which filed for insolvency last year. Premier Park sits within one of London's key availability zones, Park Royal. And this is a target growth market, as Bill just said, for a number of hyperscalers. But the site doesn't have sufficient power for data center development anytime soon. And if you have the power lined up -- if you don't have the power lined up today, it's a long wait, which is where Pure come in. They've secured 70 MVA of power capacity in West London, which is available today, but which will need connecting to the site over the next 18 to 24 months. Over and above that, and as Bill has just explained, Pure also have the technical expertise and the experience of doing the data center fit-out for the most demanding users as well as strong customer relationships with some of the largest players in this space. So together, we intend to develop a 30,000 square foot fully fitted data center with 3 stories, capable of delivering a 56-megawatt IT load. The building will be sized to accommodate a further 22 MVA or 16 megawatts of IT load, which will be provided by a second phase of power due after 2030. As Bill said, given the known customer demand in this area, we expect to pre-let the entire capacity to a single hyperscaler. Next slide, please. The total investment in the project is expected to be around GBP 1 billion, much larger than we would see for a powered shell. And this is due to the cost of the fit-out, which includes generators, cabling, cooling systems, security and fast suppression systems, et cetera. The funding structure will also be different to that which we would use for a normal project. Once we've secured a pre-lease with a hyperscaler, the joint venture will seek to secure non-recourse bank financing on the project, which is a very well-supported funding model given the expected quality of the covenant. And then the balance of the financing will come 50-50 from the 2 partners in the form of land and power contributions and a cash equity contribution. We expect SEGRO's share of cash equity contribution to be roughly GBP 150 million. Now whilst the total investment cost is substantial, the returns are attractive. We expect the unlevered net yield on cost for this project to be between 9% and 10%, leading to a significant development gain and an attractive overall IRR. The other point to note here is that we're not paying any development fees. Both partners are providing their expertise and input at no additional cost to the JV. On to Slide 11, please. And then finally, a few words on the time line. Now that the JV is signed, site clearance will begin, a planning application will be submitted and premarketing of the site will commence. Financing and the construction contract will be arranged in parallel with lease negotiations. We hope and expect to be able to start construction in 2026, which means the shell should complete in late 2027. The fit-out would then be undertaken and will be delivered to the customer in phases, most likely in 2029 and 2030, at which point it will start to generate income from that stage. On to Slide 12 then. So to conclude, we're really delighted to be working with Pure to deliver our first fully fitted data center in what is a creative partnership that is expected to deliver GBP 1 billion fully fitted data center, producing strong returns for our shareholders. Our 2.3 gigawatt land-enabled power bank located in core availability zones across Europe offers the potential for further significant value creation in this fast-growing sector. And as we move forward, we'll assess each and every opportunity to decide on the approach that optimizes the value upside and delivers the most attractive risk-adjusted returns, whether that be further powered shells or fully fitted data centers, perhaps in conjunction with partners like Pure or through other routes. So thank you for listening to us. We're very happy now to take questions. And I think we're going to go to the conference line first of all. So Drew, if you'd like to open it up for questions, please.

[Operator Instructions] Our first question today comes from Rob Jones from BNP Paribas.

The first one was on the value of the land. I appreciate you haven't disclosed it. If I think about -- or if I assume a yield -- sorry, if I assume an LTC rather LTV of, say, 70%, then implicitly, that would be GBP 1 billion of total cost, GBP 700 million of debt held by the JV and GBP 300 million of equity. Now I appreciate you've already said that you're contributing GBP 150 million of cash equity and Holdco 6, I guess, also puts at GBP 150 million of cash equity. Some of that obviously already adds up to EUR 1 billion, so that would imply a 0 value for the land. I'm just wondering if you've got any comments around that or maybe the answer is the value of the land plus site clearance on your -- in terms of your cost equals the implied value of the 70 MVA that Pure are contributing? And that was my first question.

Okay, Rob. I'll let Soumen comment on the detail. There's probably a bit of precision in your calculations that might be leading you to the wrong conclusion because we value the land at quite a bit more than 0. Well, Soumen, do you want to make a comment on that?

Yes. Look, I think the point to make here is there's quite a lot of moving parts here. We're entering into the joint venture today. We've set out the plan to kind of create our first fully fitted DC, but there's quite a lot of water needs to pass on the bridge between from where we are today to securing the pre-let, then building it out and having it fully operational. The reality is power -- land in the Park Royal area, particularly in Prem Park, you would typically value in the order of GBP 5 million to GBP 10 million per acre. And so as David said, we're clearly not valuing it at 0. We're saying obviously, our equity contribution is that plus a further, call it, around GBP 150 million of equity. And the gross capital spend on the whole project is around GBP 1 billion. But I would just draw your attention to the circa that precedes the GBP 1 billion to just say there's a little bit of -- we're just deliberately approximating around it. The debt package for something like this with a hyperscale lease in place, if you go to market, you get somewhere in the 60%, 70% loan-to-cost area is very doable. So it's just -- you're not kind of -- the land value is not 0, but I think the other things you said are not far off, but just be slightly careful around being too precise on a project that's got some way to travel.

Okay. Clear. And then the other one was in terms of yield on completion, if I was assuming, I don't know, 5.5% yield, that would imply roughly GBP 350 million of development profit to come per JV partner? And I guess linked to that, I guess, a, do you have any comment on that? And then secondly, if we were assuming, say, an LTV at the end of, say, 60%, potentially that could mean that you could get all of your equity back out, should you wish to, and then use that to fund other projects going forwards. I wonder if that was a consideration on your part post completion...

Yes, Rob, I think -- I mean, that's interesting speculation on the likely shape of the numbers. And I'm not going to comment on those because as Soumen has already said, there's quite a lot of moving parts. But some of the numbers you've come up with there are kind of in the right ballpark, but we're not going to give you the exact specifics.

Our next question today comes from Marios Pastou from Bernstein.

Just a couple of questions from my side. Can I just check in terms of -- you mentioned your planning applications to be submitted. Is there any issues here you could foresee that potentially could delay any of the project time line depending on the change of use or also your plans for the future time line of the project?

Well, not specifically. I mean it's -- we never take planning for granted. We've got a number of important stakeholders to inform and take with us on the journey. Both [ Brent Council, the JLA ] are very important stakeholders in this. We've had some early conversations with them. The project is compliant with the planning guidelines. And also, as you'll probably be aware, the U.K. government has identified data centers as strategically critical national infrastructure. And you'll have seen also the quotation in our release from the Secretary of State. So I think we won't take anything granted, but we'd like to think that there's a very good chance this will be supported at planning stage.

Okay. Very clear. And then in terms of the discussions progressing on a potential lease, how secure is this in terms of the achieving a lease ahead of construction?

Yes. Look, I'll let Bill comment on it. But as we've already said, this is in an absolutely critical target availability zone for a number of hyperscalers. We know that the power and the land availability is incredibly scarce for any hyperscaler, who want space in this area. So as I said earlier, we're confident that demand is going to outstrip supply. But Bill, maybe you could add a bit of color.

Yes, happy to. And completely reinforcing what David said. We're not going to overly specifically comment on customer negotiations at the moment, as I'm sure you'll appreciate. But given the location and supply-demand imbalance, we're exceptionally confident of the ability to let this space in line with the time frame that's been laid out.

Okay. Very clear. And just finally one more from my side on Slide 11 of the presentation. Can I just confirm the exact point at which the income starts getting produced on the lease to the JV partners. So is it at the point where the construction and fit out commences? Is it the kind of plus 18-month period where the first fit out completes? I just want to just make sure that the kind of income from the lease when it kicks in is aligned with expectations.

Yes. I mean I'll let Bill comment on specifics. But again, we've given you an indicative time frame. There are lots of moving parts that all need to come together. But Bill, maybe you can comment on that.

So I'll just make a conceptual point rather than a specific quarter X or quarter Y. But large projects like this come online in phases. And so you'll see the first phase, the second, the third, the fourth cash flowing while future -- while the remaining phases are fitted out.

Yes.

Our next question today comes from Zachary Gauge from UBS.

A couple of questions, one more specifically for Pure DC and then one for both of you. But yes, for Pure DC, it says in the press release, you have 500 megawatts under construction and available power that you've developed. On your website, it says 200 megawatts, which is quite a sizable difference. So if you could just touch on why sort of the gap there is so wide and also perhaps give a few examples of hyperscalers that you've worked with on a pre-lease basis and delivered fully fitted units for globally. And then second one, sort of for everyone and kind of following up on one of the earlier questions. I think we know sort of globally, the challenge with this model has been securing the pre-lease. That's obviously the really valuable part of this development project, and that's where you get sort of the uplift on value. What sort of really makes you confident on this specific location? And perhaps if you could touch on why Park Royal is seen as slightly different to Slough, where obviously SEGRO already has a platform and has quite a lot of power availability. Why is Park Royal different in that sense? And again, why you're sort of very confident that you will secure a pre-lease? And could you give any examples of hyperscalers, who have recently pre-leased fully fitted units in and around the London area, give us some context on that.

Bill, do you want to...

Yes. Look, a couple of things to hit on there. So one, you have correctly noted our website needs a refresh. There's nothing I can do, but apologize on that. I'd tell you that is in flight, and so we expect to launch that shortly. The figure given today is the correct figure. And I hope you'll appreciate, again, as a private company, we're not going to overly disclose specifics about our portfolio, but we are in a number of major European markets like London, although, of course, none is as big as London in Europe. And our customer base is the very, very largest customers in the world, the AAA and AA-rated American technology hyperscalers, of whom we have multiple deals with multiple of them across our global footprint. In terms of the pre-let, this is fundamentally, again, a function of location, right? You noted Slough and Park Royal. So Slough is the single largest submarket within what is broadly defined as London. The way that -- without going too far into specifics, the core cloud infrastructure works separate from things like artificial intelligence is through what are called cloud regions that need to be balanced. So for the major hyperscale customers, they have big footprints in Slough. They have matching footprints in Park Royal. And so Slough can't run that far ahead of Park Royal in order for their infrastructure offering to work as they've promised it to their customers. So this is fundamentally a function of supply and demand. There is enormous demand for growth, and there is an extreme constraint on supply in this submarket, and there is a lack of flexibility. Our customer base cannot put this capacity in the Docklands. They cannot put this capacity in Waltham Cross and have it work the way they need it to. So you look at the overall trend line, the supply and the demand, again, what makes us so confident about the ability to pre-let this.

So maybe just to add, Bill, the -- in terms of that London and West of London market, there are effectively 3 key availability zones, Park Royal, Hayes and Slough. So it's all about having the ability to replicate and have resilience. If you're a hyperscaler, you don't want to rely just on one data center cluster in one availability zone. And that's why you see demand for space in all 3 of these areas.

Yes. And you mentioned have we seen examples. Again, I won't cite specifics because they're confidential, but we are just within the last, call it, 24 months aware of at least 5 or 6 specific examples.

Okay. Great. And just maybe a very quick follow-on. I appreciate you're a private company, so I didn't mention anything. But have you delivered fully-fitted operational data centers to these large hyperscalers that are currently up and running?

Yes.

Or whether is it...

Yes. No, I was at our data center -- operating data center live with a major hyperscaler in West London on Friday. We're just doing a little walk around following the Heathrow scare, making sure everything was good and talking with the team. Happy to say, no impact to operations at all.

Our next question today comes from Max Nimmo from Deutsche Numis.

Just a quick one from me. You've been pretty clear about the fact that you're not going to be taking on IT equipment, racking servers, this kind of stuff. I guess one of the big questions that the sector still has as a whole, though, is just on this kind of obsolescence point. And as you said at the top, this is not riskless. So within that sort of M&E element, sort of long lead equipment, where do you see sort of the biggest obsolescence risks are? Obviously, you're trying to mitigate those as much as you can, but just to try and get a bit of a feel for where are some of those obsolescence risks that sit within this over a 15-year time horizon.

So look, it's a really good question. It's one we think about a lot as well. And again, it's worth really driving on the fact that while we are an enabler of high technology, we're providing infrastructure. And so fundamentally, power and cooling are not significantly rapidly changing concepts. Now you do see some evolution in the specifics, but the relative scale of the refreshes at the end of these leases are relatively modest. And while obviously, the sector has grown significantly, there were a number of these contracts signed now 10 years ago, and those contracts have rolled. You probably see this a little bit more in the U.S. But again, the scale of the refresh and the re-leasing is entirely consistent with what we'd expect, which is the demand remains robust, and there is a relatively modest requirement for refurbishment. I would also just add that the lease structure that's being pursued by the joint venture is designed to further insulate from this in terms of the responsibility for maintaining the equipment.

Our next question today comes from Jonathan Kownator from Goldman Sachs.

Two, if I may. The first one, I think we discussed in the past about sort of rule of thumb of about EUR 1 billion for 100 megawatts, excluding the cost of land, I guess, for all the equipments and MEPs, et cetera. I just wonder if that rule of thumb still works or if you've seen increases in costs more recently or if you're building, I don't know, for instance, for higher spec. That's the first question, please.

Andrew, do you want to comment on that?

Yes. Good morning. I mean, obviously, we've built in what we think the numbers here are indicative, as Soumen said, that's the first point to make. We've built in what we think are very sensible cost estimates in line with market. And when it all triangulates when you look at the expected yield on cost, noting that, that is indicative. It's -- we believe it's bang in line with market at a yield on cost of 9% to 10%. So I think probably those numbers you quote there probably have moved on a bit. But as Soumen said at the start, I think we need to be really careful with this given the numbers are indicative not to go too precise on dividing one number by another to get to the cost per megawatt. We're very comfortable that the estimates that we've used here, whilst being indicative, are in line with market. I think the point is that -- the other point is that Park Royal is not the cheapest location in the world. You wouldn't -- if we're building this in some remote location in Spain, for example, you'd expect the cost to be considerably lower. And so the costs are higher, but so will the revenues be higher because of the scarcity of the location.

Okay. Very clear. And the second question was, if I look at the JV contribution, so you're saying that on the SEGRO part, you're contributing the land and GBP 150 million. I just wanted to understand on the Pure DC part, there's the power contribution and are you also going to contribute the same amount of EUR 150 million of equity? Or is that a different consideration of...

Yes. I mean both parties are contributing some, if you like, tangible asset. In the case of Pure, it's the power. In the case of SEGRO, it's the land, and both will be contributing a cash equity to fund the development that's not covered by the project finance. That's kind of all the detail we can share with you on that.

But both parties ultimately are putting 50% of the equity in whether in kind or in cash.

Okay. So the contribution on the DC part is not GBP 150 million. It's not the same.

I'm not going to comment on the valuation of the power. But the overall -- I'd say, the equity that both parts are putting in is 50-50. So you can come to a reasonably good view as to what the land value being contributed in is. And we're saying that our cash equity is a further GBP 150 million. The total equity contribution from our partner will be of the same magnitude in total.

Yes, it's a 50-50 joint venture. If it ends up being GBP 150 million of equity from SEGRO, it will be GBP 150 million of cash equity from Pure as well.

Our next question today comes from Ventsi Iliev from Kempen.

Just a quick question. Is this a onetime JV? Or are you actively exploring further opportunities together? And then actually, just one clarification question. Does the EUR 1 billion capital commitment include the value of the land and the power?

I'll just deal with the second piece. Yes, the GBP 1 billion does include the value of the land and the power. On your first question, is this a one-off? I hope not. I mean this is the first deal we've done with Pure. We -- as I said earlier, it's something we've known for a long time. We've been talking specifically about this project, but our partnership and our intent is not limited to a single site. So I wouldn't be at all surprised if you find that we announce further deals with Pure or indeed other potential partners in the future.

Our next question comes from Callum Marley from Kolytics.

Just a couple. On the fully-fitted model, given the significant CapEx and equipment required, how sensitive is the 9% to 10% development yield to CapEx overruns in the construction phase? And then what are the kind of new delivery or operational risks that you're now exposed to in this model versus the powered shell?

Look, I can take that. As I sort of say in the beginning, you've got -- the priority of the joint venture in the next 12 months is to get a pre-lease secured. Alongside that, we'll be negotiating a debt package and a construction contract. And all those 3 things will come -- the plans they come together. And therefore, the construction contract will match the specification requirements under the lease. Now clearly, in any development project, there is a risk of overrun. But look, given the Pure's track record in this particular space and our wider track record across our wider pipeline, I think that's a manageable risk, frankly, when we know the specification of what we're building. You wouldn't -- it won't surprise you that within the GBP 1 billion that we're talking about, there's a reasonably decent contingency in there for the unknowns that are sitting here today, we're not aware of.

That's clear. And then just a second one. It sounds like the net lease structure is the preferred method. If you can't achieve this over the next couple of months, would you -- going into the operational model or a hybrid between the 2? And then would that materially change anything with yields or profit on cost?

Yes. No. Look, at the end of the day, we're here to serve our customer and provide the product they want. Our preference is to go for a fully fitted model where it's a net lease structure. And those are the numbers we've quoted on that basis. If the customer, for whatever particular reason requires it to be a, if you like, a fully serviced operational data center, then the good news is Pure have tons of experience of doing that and providing that capability. And the economics on a net basis would still work out to be similar. So the gross income will be higher, and that would cover the cost of providing the services required. So either way, we're comfortable with the overall net return we're showing here.

I'll now hand over to Claire to run through the webcast.

We just have one question on the webcast. With the change here to a kilowatt per month pricing model versus the usual square foot model, how do you expect this to change the income profile of the lease, please?

Yes, that's pretty straightforward. It's not a variable price is the point to say, it's not like a turnover lease according to the power. So it's just the way that the lease is negotiated. We -- the data center market talks about prices per kilowatt of capacity. Once you negotiated and agreed it, that's fixed. So the income profile is exactly the same as any other real estate lease. It's just the way it's -- the way the headline pounds and millions are calculated is just done differently. You end up with the same profile.

It's likely there will be some form of inflation linkage going forward...

That's a good point...

You wouldn't have in a -- unlike a typical U.K. lease structure, we have a 5-year upward-only review. But look, everything is -- I'd say we're sitting in today to launch of this joint venture, we need to go and negotiate...

That's a good point that. There will be inflation protection.

Yes.

Okay. On the webcast, thank you all very much for sparing the time this morning. We -- as I said earlier, super excited about this first step into the fully-fitted market. We're thrilled to be working with Pure. A lot of work to do, but this is a great first step for us. So we look forward to updating you all in the coming months. Thank you.