Digital Realty Trust, Inc. (DLR) REITWeek – NAREIT Conference (Transcript)

Digital Realty Trust, Inc. (NYSE:DLR.PK) REITWeek – NAREIT Conference June 6, 2022 11:45 PM ET

Company Participants

Andrew Power – President & CEO

Chris Sharp – Chief Technology Officer

Conference Call Participants

Jon Petersen – Jefferies

Jon Petersen

All right. I think we can get started. Thanks. Thank you, everybody, for joining us today. My name is Jon Peterson. I’m the Head of REIT, research team at Jefferies. Really honored to have the Digital Realty team here with me today. For any of you that follow my research, this is my top overall pick in the REIT space. So I think there’s a lot of really exciting stuff to talk about today. So you have pretty full room here, so I think you guys all agree.

But joining us today, we have Andy Power, who is the CEO; Chris Sharp, Chief Technology Officer, right? I got it, okay. So I’m going to hand it over to them to give us a brief introduction. I have a list of questions. But if you guys have questions, we’ll also open it up to those as well.

Andrew Power

Thank you, Jon. Again, I’m Andy Power, I’m CEO of Digital Realty for — I have been in this seat now for five whole months, fun time in the world to step into it. But luckily, we’re a fantastic industry with a great company and been making great progress in short order here. For those less familiar with us, Digital Realty is the largest owner, operator of data center capacity, digital infrastructure, spanning six continents north of 50 metropolitan areas, supporting 5,000 plus customers worldwide, with a full spectrum of capabilities from enterprise oriented hybrid, IT, co-location interconnection to the multiple megawatt dedicated data halls for hyperscale customers who we support in 20, 30, 40, 50 different locations around the globe.

Since moving from my prior role as President and CFO in to this role, I’ve had three point plan looking to activate. One was demonstrally strengthened to our customer value proposition. In the last half of the year, not even, we’ve taken some actions from an organizational structure, including the creation of Americas region and the appointment of Steve Smith to lead that region as Head of Americas.

I followed that up with a change to our CRO, an internal candidate who had been worked alongside me for many years, initially leading sales operation. as well as our top global accounts and most recently, the Americas Enterprise business and connectivity, now stepping into our CRO and really trying to elevate our go-to-market in a more data-driven format. And we’ve been benefiting from the last several quarters now of better penetration of less than megawatt interconnection category that represent almost half of our new signings last quarter, and north of 120 new logos added to the 5,000 customers this past quarter.

Two, we have some integration and innovation to do. We’ve been making a lot of progress behind the scenes, going live in terms of ERP systems, portal consolidations, removing the friction for our internal customers and our external customers and bringing more innovation on the connectivity, aside with the launch of our Service Fabric, as well as innovation on the space and power front, in terms of high performance computing and higher densities, and more to come on that front.

And last but not least, my last piece was to diversify and bolster our sources of capital. We believe that the demand environment was going to continue to remain robust, if not accelerate. The demand was getting bigger, especially when it came to the hyperscale arena. This decision was even probably before the advent of artificial intelligence becoming a bingo buzzword. And a part of that was to go out and bring in more partners on the hyperscale front for both stabilized assets as well as development projects in the Americas as well as EMEA region.

And we’ve made some progress on that with our announcement yesterday in terms of north of $1 billion of capital raising to position the company for those opportunities that lie ahead, including less than $200 million of non-core dispositions what we believe an attractive valuation.

With that I’ll turn to you, Jon.

Question-and-Answer Session

Q – Jon Petersen

Yeah. Can we actually maybe start with the announcement you guys made last night. Talk about the data center that you did sell and maybe the stuff that you are also working on right now. I thought the cap rate of, I think, 4.4% on that disposition was probably a lot lower than most of us expected, which is a good thing. So talk about that and then also maybe touch on the equity offering or the ATM issuance, why now?

Andrew Power

So on the non-core dispositions, as a reminder, we laid out as part of our $2 billion funding plan that we initially laid out at the beginning of the year about $500 million of that is non-core dispositions. The first major non-core disposition to get over the finish line was the transaction we announced in Texas. The bucket of non-core, in our opinion, are things that do not have robust diverse customer demand, markets where we have durable pricing power and value add. Runway for future growth, where we operate the capacity. This particular data center was leased on a power-based shell basis.

So encumbered by a lease, where essentially, the juice was pretty much squeezed by the operator for years to come. Hence, they fall in that non-core. And I think it’s a small but relevant data point as to asset valuation and the shift of capital towards digital infrastructure away from legacy asset classes that’s been building for some time and certainly playing out this year as well.

On the additional capital we raised, I can tell you, we started down this path with a preference for those private capital initiatives, both non-core and JVs around our hyperscale stabilized assets as well as development assets as our best source of equity capital to fund the business and align with our strategic objectives. The equity raised under the ATM is in addition to those initiatives that we’re still contemplating, not contemplating, we’re actively working on. So not supplementing one of those.

And quite frankly, we are faced with an environment this past several months with a lot of uncertainty. And we were facing at the same time, uncertainty around capital access and places where we saw opportunities to invest. They’re very time specific that would increase our CapEx, development CapEx load at very attractive ROIs. And while as much as I do not love issuing equity at this current stock price, we thought it was a prudent course to not put the enterprise at risk and raise some capital to essentially be able to maintain the optionality and ability to execute on those investment opportunities.

Jon Petersen

Great. Maybe just one final question on that. There’s been a lot of talk and chatter about timing on these JVs that you guys are working on. Is there any update to give us on what we should expect in terms of timing?

Andrew Power

If I had another JV that announced today, I would have announced already, we are — we have — this is not a linear activity. It’s not like we started on one, we’re going to push it and get to the next. There are a host of different types of projects from stabilized assets to development as well as a small portion of non-core. We’re making good progress. In the current environment, I don’t think anyone would be surprised if it’s taking longer than I would have hoped for. But we are making progress with multiple parties and multiple projects and I don’t think we going to have to wait all too long for the next incremental data point to come to fruition here.

Jon Petersen

Got it. All right. Let’s talk about AI. So obviously, we’re in early innings, but can you give us some context on like what demand have we already seen that maybe we didn’t realize before NVIDIA’s results, right? And then what kind of conversations are you having now in terms of future demand?

Andrew Power

I’ll let Chris GPT over here, chime in along the way. There is AI in our 300 plus data centers in terms of workloads, but it is a minority of the true sense of what AI is today and whereas could be in the future. And I’m seeing small minority. The lion’s share of not only our installed base, but our leasing activity, our new logos over the last several quarters, I would say, has not been AI oriented. It’s the growth of the cloud. It’s going to move from on-prem to off-prem. It’s edge activities. We are, I would say, infancy may be an overestimation of where AI is as a category.

We’re seeing it percolating on the demand side in terms of customer inquiries. And anecdotally, it’s urgency around large capacity blocks with contiguous capacity as fast as possible, that we certainly hear from the customer are AI-related. Certainly more training oriented. And I think that the definition of training and what training is today for artificial intelligence and what it will be in the future is going to obviously evolve. Only way I could think about this, I don’t see how this is not a rising tide lifting all ships in the data center because this AI will live in the data center is my belief.

And we have a very big role to play. We’ve — we didn’t come from occasion cabinet colo-only (ph) company for the existence in this company. We actually went in reverse servicing the hyperscalers at the highest technological environments, power densities, cooling features, and we’re already supporting them with their availability zones across the world. And I think we’re going to have a place to play in supporting the growth of AI. And what we’re trying to do is navigate it to do that in the most value-add to the end customers’ deployments to create a more sustainable business model around it.

Jon Petersen

Got it. Is it — I know it’s appreciating that it’s still in its infancy. What is the demand you’re seeing today more on the hyperscaler side, I would assume? Is it kind of the — how you see the large hyperscale deployments first and then the enterprise follows, how do you…

Andrew Power

First of all, yeah, some of these requirements are so large. You have to be a well-capitalized company to even be in the arena to pay the bill. So the preponderance are hyperscale oriented in terms of customer pipeline demand. But we are seeing names that, quite honestly, I never heard of going months back, pop up with large requirements as well. If you look at our enterprise list of people — of customers without AI applications, it probably is two pages long, but they’re smaller like less than a megawatt type capacity. But I think all these elements will grow over time.

Jon Petersen

Yeah. Okay. We get a lot of questions about data center design and like in the evolution, like whether it’s AI or any other GPU-driven application, it seems to require a higher power density. That’s maybe a question for you, Chris. Like, how are you guys thinking about data center design today on future builds to meet that demand, but then also talk about like what does this mean for legacy, for lack of a better word, the assets that you have today in terms of their attractiveness?

Chris Sharp

No, absolutely, to Andy’s point, right, where we’ve been contemplating and looking at how infrastructure has been maturing over many, many years, right? And one of the things that’s an advantage coming from servicing the hyperscalers and evolving to their architectural requirements where it’s not a matter of if, it’s when it densifies, it’s not a matter of does it need data, it’s how much data. And so the power distribution and cooling technologies we’ve been refining over the years and how you bring data together is that the culmination of what’s going to be required to making AI successful going forward.

And so when we’ve looked at a lot of these pockets of infrastructure coming to market where it could be a GPU, it could be other specific hardware that’s required for, quite frankly, an outcome for hyperscalers, traditional infrastructure, the designs have been contemplating that for many, many years. And so be it old, be it new, we have a theme in our designs called modularity. And so the modularity around the power, the modularity around the cooling, that’s core to a lot of the designs that we have today. So one of the other elements that I think this audience would appreciate is the balance of infrastructure.

To Andy’s point, right, where there’s still going to be a need for CPUs, there’s still a need for storage, there’s going to be a need for GPUs. And so having the ability to fill in the broader architectural requirements and be able to support that in a blended basis where we can align with our overall customer requirements is absolutely top of mind to us. But with generative AI, which I think is the key term and buzzword that a lot of us talk about today, it’s pretty prevalent. It’s still very early. And what’s nice is that, I think to your earlier point, we do see that coming through the hyperscalers initially, where it will be embedded and it’s additive to a lot of the services they’re already delivering, right?

And you can read about all of the online suite of applications. These will have AI capabilities embedded in them, that requires additional infrastructure to support that. And quite frankly, we see it as being a true differentiator where if you reply to an e-mail and it’s not 80% already written, you won’t use another e-mail system that doesn’t have that type of capability. And so we saw that early days. We see that continually evolve. But also, AI is a horizontal, right? It’s not a vertical specific thing. There is not a customer in our facilities that is not thinking or has already thought about how are they going to support their AI aspirations because it’s creating a differentiator. It’s actually creating a state of play that if they don’t do it, they know they will be disintermediated from their consumer.

And so that’s something that we’ve also been looking at for many years and how do we support these larger enterprises coming to market where Andy had alluded to this, these larger capacity blocks in our designs benefit these enterprise customers where it’s a very unique environment, right, where you can do all your broader architectural requirements around networking because we’re all heavily interconnected, which is a key component of this. I think we would appreciate, AI is only as smart as the data sets you feed it. And so having access to these data oceans that reside within platform digital already is a true differentiator as well.

And it’s, quite frankly, a key component in our customers’ requirements, but bringing that all together and being able to support not only the generation of architecture today, but quite frankly, the second and third and fourth generation, because the GPUs that everybody has been talking about for the last seven days, wait, there’s going to be another set and there’s going to be another set. And so our designs have been contemplating that for many years.

Jon Petersen

Got you. And I think you touched on this a little bit, but latency. But with the applications that you’re seeing today, whether it’s AI related or anything, like talk about like how much does latency matter. Obviously, there’s like those interconnection data centers were like super important, but like for applications stay like do you need to be an Ashburn for some of these things or are you okay in wherever, middle or nowhere, like, talk to me about why some of that stuff is important or not?

Andrew Power

Right now, this is just getting off — we’re in spring training or training camp for the training like and it’s a long amount of training just to get these things smart. And you can see this in terms of these systems that the consumers are using being up to date. Like, what can you tell me about last night’s basketball game type of dynamic. As this training and larger models get built, the latency is going to become more important because it’s going to be more real-time interactions, right.

We were chatting on the way down, whether the inferencing happens at the data center or on the device or both, it feels like the train is at least — at the very least it’s going to evolve to more latency specific, proximately data because this is going to be about B2B applications that have maybe not high-speed trading type latency, but some interaction with proprietary private data sets mixed with public data sets to be making informed decisions from a business lines.

Jon Petersen

Okay. All right. Let’s switch to just market fundamentals. I’m a real estate analyst, so I love talking supply and demand. We’ve seen vacancy rates fall quite a bit over the last couple of years. I think in the U.S., it was 10% just two or three years ago, we’re down to 3% today, tighter in some markets. I mean, what are you guys seeing? How is that impacting your ability to push rents?

Andrew Power

The pendulum on the fundamentals, even pre-AI conversations has been moving since last year, since the beginning of last year, you could see this in our guidance for cash mark-to-markets, you can see this in the results. You can see this in this year’s guidance and the results of this quarter. You see this in the face rates, our price book rates and it’s flowing through. That is a combination of demand remaining intact after records upon records and supply from multiple vectors, power availability from transmission is nimbyism (ph), moratoriums, moratoriums on supply. A whole host of supply chain, a whole host of factors just falling behind.

And it’s more long-term frictions in the system where markets like Ashburn is going to be without power until 2026. Markets like Santa Clara is going to be without power until 2028 or 2030. Similar things happening in Frankfurt, Paris, Singapore having a moratorium. And that has allowed us to essentially make sure our value proposition to these customers translate into better economics. And you’ve seen that now in our new rates, on our rate leasing, in ROIs as well as on cash more to march on our releasing spreads.

Jon Petersen

Yeah. What are those discussions like on the re-leasing spreads today versus what they were like two years ago? Like, how did those negotiations start, where do they end?

Andrew Power

We’re not customer business. We’re trying to build broader relationships that span geographies, go deep in terms of the customers’ needs. But we also have to have tough conversations of that there’s a virtuous cycle here. If — this is always a relationship where you want to just treat us like commodity, I’m not going to be able to have the capital to support your growth. I’m not going to be able to add the 56 metro or the growth capacity in this market or the land bank. But if we can have a real true relationship even with big customers that are super sophisticated, have capital and have a bilateral of relationship, then I have a virtuous cycle that I can help future-proof their growth.

And I can hold land for years that are 1,000 megawatts of capacity that can have dedicated supply chains. That I can be going out and procuring green power on their behalf through PPAs, biggest amongst the industry. So that — it’s all about essentially articulate the value proposition and the why of the commercial discussion is most important.

Jon Petersen

Yeah. It seems like a lot of like market rents are kind of set by where the rents unlike the next development or like most recent development was pegged at. I mean where — what are your required returns on a development to start today? I think if we go in your supplemental, your average expected yield is like 11% on the whole portfolio, I think 8.4% in North America. But talk to us about like if you started all of those today, what are those yields?

Andrew Power

The North America is dragged down by an apple to an orange where there is an open book yield on cost deal, we’re developing, but we’re insulated from the construction cost. And it’s triple net leased, credit quality, long-term lease. So I wouldn’t say that’s not — that’s an outlier to the other majority of the 400 close megawatts on that table. We’re pushing towards double-digit returns even on larger transactions in this environment.

And yes, we priced the cost of capital as a hurdle to essentially greenlight. And we’ve been we raised the bar as to what good looks like at Digital Realty, basically at the end of last year. But we’re also dynamically pricing to the supply demand dynamics. And there’s examples, be it Singapore, that had a moratorium for years now or Ashburn that is without power for several years, where we have the last remaining capacity. And we want to be judicious to make sure the right customers and the right commercials fit that mold.

Jon Petersen

Yeah. Okay. And then I guess in terms of power constraints, I think you listed off a number of markets. Are there any other markets where you expect there to be issues coming up?

Andrew Power

You’re seeing a constant theme, whether it’s Dublin, whether I believe it’s a generation issue, whether it’s [indiscernible], whether it’s a transmission issue, whether it’s France, where the nuclear plants weren’t maintained sufficiently during COVID, and that’s going to derail incremental power growth or its certain communities from a governmental range or a social range saying, hey, let’s pump the brakes on a big consumer of power in these times, or you’re putting it up against water issues in certain markets, that’s more seldom than versus the power issues or just still hiccups in supply chains.

And now all in a world where capital and debt is much more scarce and costly than it was most of the prior almost 10 years. And all those things are contributing factors. And yes, you can say some of these are going to run its course, but these are not overnight quick fixes at the same time.

Jon Petersen

I guess, with how strong demand is right now, like what are you seeing from your competition? I know they have the same issues in terms of just power being available. But do you expect more capital to be kind of thrown at this space? There’s been a lot of private equity buyers of your various platforms that I used to host on the stage that are no longer here. How aggressive are those guys?

Andrew Power

I was thinking about this in a different context. I bet all the funds that bought a platform, which they were to save that money and just put that into speculative development by themselves in hindsight when you think of the return equation, but obviously, was crystal ball was going to be able to see where this was going. I think what we’re seeing thematically in the industry is on the enterprise side, ourselves and one other provider are really the two global capability providers. And we don’t see much interest or third or fourth competitor relevance. And that’s — I mean that’s healthy because this is a hard business to become 50-plus metro six continents. And it’s a category that is a large addressable market with a tremendous value add. And it’s a — don’t put all your eggs in one basket type of buyer purchase, right?

One of our value adds is redundancy, resiliency and diversity, right? On the hyperscale side, we’re obviously seeing all those platforms that were kind of mixed models really just go towards hyperscale. And that’s obviously, they’re much — they’re not permanent platforms, they’re finite funds, their IR clocks are taking more ferociously than you could say, maybe the public market. And they’re running the same thing that we’re seeing that it’s not easy just to build data centers anymore. It’s not easy to staff them. It’s not easy to put the land entitles, not easy to get the power. It’s not easy to bring the sustainability efforts and this is not just a throwing capital yield on cost type of business, quite honestly.

Jon Petersen

Right. So we got about 5 minutes left. Do we have any questions from the group?

Unidentified Participant

[Technical Difficulty]

Jon Petersen

Okay. So the question was, widening the economic moat to not give up market share?

Andrew Power

So I slice in the two buckets there. On the enterprise front, we’re doing everything we can to differentiate our value proposition in terms of new technology and tooling, new markets, new runway for growth and do it in a differentiated format. You can grow with us, you can go higher power density with us. You can we really have one major competitor. So I get to draft off of that major competitor and then play to different areas of strength in a category where the customers want multiple providers, quite honestly.

On the hyperscale front, we don’t operate the market share is the key determinant of value add. We’re basically saying we don’t want to go follow you to places where you go to treat out everybody like commodity. We want to be places where we really make a difference to you. And for digital, at least half of my 50 metros, that matters. They’ve already have availability zones with us. We have massive campuses with runway for growth. We’re already servicing their sustainability needs, their operational needs or health and safety needs, their security needs and not being — we’re not just following them for the sake of market share, but where we can actually add about — moving needle difference to them, and be at top of their list of critical providers no different than their insource teams.

Jon Petersen

Other questions?

Unidentified Participant

[Technical Difficulty]

Andrew Power

I think the — we — including me as a management team. Sorry, I’ll repeat the question. What I think in a nutshell question is, why are you going faster? Why are your stock price total return higher given being — blessed being in a good sector — in the fastest growing sector? The — I would say two things happened. One, when you say it’s the fastest-growing sector, that’s from a total addressable market standpoint. But at the same time, this sector has probably had more new entrants and competition coming in to it than any other sector in this building at the same time period, right?

So I chalk that up to industry maturation where this industry, data centers around 20 years ago, but they’re only owned by opportunity funds with 20 IRR aspirations and real estate folks, even though we became a REIT in 2004, I would say, been doing a lot of learning involving their thoughts on this and infrastructure funds and you had a cost of capital flow into the space. So you know what these data centers are maybe have more mortality and longevity to grow it than an office building finally. So yeah, that going on at the same time.

Also as it relates to digital, we — I think this company has got a lot to be proud of where we started being the first REIT in this asset class, having a tremendous runway of growth to create the platform that it has today and becoming global and scaling. But the company — the company was eight, 10 years ago, whatever the thing was needed to continue to evolve. You wouldn’t say it was aligned towards the best and the highest growing customers. It was aligned towards enterprise colocation or interconnection customers. It wasn’t aligned towards the hyperscale customers.

So that was something that needed to shift. The definition of global was London. So going from two or three international markets to 50 metros in six continents had to change. Didn’t have the capabilities in terms of being able to service the interconnection colocation customers. I think I joined the company in 2015, we maybe had 600 customers, we’ve now 5,000 customers. And in order to make this a more long-term durable business, we had to do measures that included inorganic M&A that strengthened the portfolio and the long-term pricing power and growth as well as selling assets that were non-core to us, but depleted our earnings along the way.

So maybe that’s a big excuse of work in progress. I think we’re coming into a better future than what we’ve gone through. We’ve taken our lumps, and I think the platform today, as you’re seeing in the results is starting to see the pricing power, starting to see the increase in occupancy, starting to see the same-store growth. And honestly, if — even with hindsight being 2020, I don’t think if this company would have made most of those changes in the last five, 10 years. I think you would have a weaker hand even today having stated as prior strategy.

Jon Petersen

All right. One rapid fire question. Is there any risk to the dividend?

Andrew Power

The story with the dividend that those a little bit less familiar with REIT tax. So we have — when we sell assets, either outright or joint ventures, we typically have capital gains because we build things to high returns and sell them at lower cap rates than those returns. In order to retain that capital, the dividend essentially absorbs that return. Otherwise, we have to pay a special dividend. So if we believe that we’re going to be successful in the non-core dispose or the JVs that I mentioned, we’re — we should be keeping the dividend where it is to absorb that gain and retain that capital. So — and we believe we’re going to execute on these joint ventures not in core assets, like I said. So hence, right, I do not see the risk of dividend based on that fact pattern.

Jon Petersen

Got it. Sounds good. Thank you for your time.

Andrew Power

Thank you.

Jon Petersen

We’re all out of time.