Welcome to Citi's 2026 Global Property CEO Conference. I'm Nick Joseph here with Michael Rollins with Citi Research. Pleased to have with us Digital Realty and CEO Andy Power. This session is for Citi clients only. Disclosures have been made available at the corporate access desk. To ask a question, you can raise your hand or go to liveqa.com and enter code GPC 26 to submit any questions. Andy, we'll turn it over to you to introduce the company and team, provide any opening remarks, tell the audience the top reasons an investor should buy your stock today, and then we'll get into Q&A.
You have to press the button to turn it red.
Thank you, Nick and Mike and Citi team for having us back here. With me is our head of public and private cap investor relations, Jordan Sadler. Digital Realty supports nearly 6,000 customers in 55+ metropolitan areas on six continents with their digital transformation, cloud computing, and artificial needs through our data center and connectivity infrastructure. Solely focused on that for 20+ years. Think we're a differentiated opportunity supporting the full spectrum of our customers' needs. The three reasons I'm very bullish in long-term Digital Realty are, one, we are at the forefront of not just one, but numerous secular tailwinds of demand. Digital transformation changing everything we do and how we transact and operate in life. Cloud computing still rapidly globalizing and expanding.
Half a trillion dollar market for our customers growing at nearly 30%, some would say restrained in today's AI era. At the very earliest innings, or maybe not even out of the dugout on artificial intelligence and demand trends that are about to come to market. Two, we at Digital, I believe in both the public and the private markets, are the only one uniquely positioned with the full product spectrum. We are rapidly supporting nearly 6,000 customers, which are growing at double digits year-over-year with our colocation connectivity enterprise needs for hybrid cloud, all the way to hundreds in megawatt buildings for our hyperscale customers for both their cloud computing and now AI needs. The full spectrum is how we're doing it.
Lastly, we're coming off the heels of a call it, 9% constant currency compounding in our bottom line with a guidance close to 8%, and a tremendous backlog records for the history of the company, that we think will portend a multi-year consistent compounder at our bottom line, for our shareholders.
Great. Thanks for that. You know, as you think of delivering on the objectives for 2026, you know, what are the key assumptions that are, you know, underpinning that in terms of the types of sales and bookings growth and participation in both your under 1 MW retail-centric business and in your over 1 MW business that leans to the hyperscale?
Sure. On the less than a megawatt or enterprise colo connectivity, we're coming off a landmark year with multiple record quarters in the year. In that category, we put up 35% more new signings than we did in the prior year with pretty much the same amount of quota-bearing reps. Massive productivity step in that category, building on the investments we made in prior years, and really, taking share in that market. We're off the heels of our sales kickoff, setting quota double digits higher than that for the year we're going into. We've got tremendous momentum in that category, and we're off to a good start in just the first few months of this year.
That continued focus execution, that's a category where we were doing call it $50 million-ish of signings a quarter for a long time, $200 million-ish a year. We're on a journey to double-digit in that category, take it to $400 million, and we've shown we can do it in the last two years. On the larger capacity blocks, that's an area where we don't solely judge ourselves on market share. We pick our spots with our hyperscalers, maybe 30 out of the 55 markets where we have a competitive edge, something that our customers truly value, be it our large capacity blocks, our operational relationships, places where it's tough to do business, and that's where we deliver and deliver outsized returns on those capacity blocks.
Fortunately, we've been doing the hard work in spades to line up the capacity blocks with the large suites of megawatts and on gigawatt campuses where our customers need them, lined up our supply chain, pulled in our deliveries, and putting those in the sweet spot. We don't have to do a tremendous a lot many in there to deliver on our 2026 plan, but we are working on a big, big pipeline in that category, and I think it's gonna fall into some very attractive capacity blocks, especially here in North America, but also some outside the U.S. as we start to see cloud continue to globalize and AI start to more rapidly globalize.
Maybe starting backwards on the hyperscaler demand front, can you talk a little bit more about the pipeline and what you're seeing in the interest levels and how they're evolving both in the North America markets as well as Europe and Asia? You know, is there something that's, you know, as AI use cases are starting to get absorbed, you know, by enterprises, are you seeing any, you know, change in the speed or needs of your customers?
In the hyperscale arena, we're talking well above single megawatt type transactions. I think two data points we experienced in the last year are very relevant. One, we set out tremendous diversity of hyperscale demand. From investment grade hyperscale customers, top three, top five hyperscale customers, those types of names were our largest given signing in, call it, five or six of the last prior quarters, not the same customer over and over again, which is part of our strategy. We want numerous customers landing with us and growing with us and thriving on our campuses. Tremendous diversity of demand. Two, we started to see the seedlings of more international growth. If you saw in the fourth quarter, EMEA had an outside contract contribution in the less than a megawatt or excuse me, greater than a megawatt category.
You're starting to see the seedlings of globalization of these AI trends, not to the extent that the U.S. has to date, but moving in that direction. I think Latin America is gonna be a strong contributor to that in the first half of this year. We move into 2026, albeit we're only two months or 60 days into it, I think that diversity demand theme is repeating itself. I think some of the innovation that's just been launched in preceding weeks with new releases from the likes of Anthropic or OpenAI are spurring incremental confidence in this technology and incremental demand for our hyperscale customers to seize on capacity as we speak.
I think our conviction around the demand pipeline has only increased since we last spoke with any of you at an Nareit or third quarter or fourth quarter earnings call.
That's really helpful. As you think about that demand side of the equation, can you discuss the supply side in terms of the types of inventory that you can sell or the quantity of inventory that you could sell into the market? When do those get delivered? Are, you know, 2027 into 2026, 2028. Like, where are you in terms of where new leasing gets delivered?
Our 2026 capacity blocks are the most precious capacity blocks in the planet, on the planet. 2027 is shortly next in line. 2028 is we're starting to come into vintage, but I'd say those first two years are the most sought after. In those, call it, delivery time cycles, and call it the nearly 50 or larger single buildings, we're talking multiple places in Northern Virginia, with that in 50s and 100s type capacity blocks. We're talking same sizing roughly in Dallas. We're talking 200 to 400 in Charlotte. Similar kind of capacity sizes in Atlanta. Closer to 50 in Santa Clara. Those are all certainly on our customers' radars with active discussions and proposals underway.
Outside the U.S., as I mentioned a minute ago, Latin America is certainly beginning to light shine on it in the, call it, 20s to 50s or more type capacity blocks. Asia, there's conversations around what we have in Tokyo, Osaka, and Seoul that are very much lend itself to that. I think our Amsterdam, Frankfurt, and Paris markets, where the concentration of those larger capacity blocks are being most actively discussed.
One of the questions that we're getting from a number of clients is just how to think about growth and returns for Digital Realty as some of the leasing goes directly on balance sheet, you know, 100% owned by Digital Realty, and some of the leasing is going into joint ventures or your you know, fund strategy. Can you share with us how you think about the contributions to profitability and returns, you know, in each of these buckets?
We essentially made a strategic pivot in our funding model to tap into private capital initially through one-off joint ventures, both stabilized portfolios and also development, really with a move to, call it, bolster our ability to fund this tremendous growth in hyperscale that we saw. Really allow the tremendous growth in our colo business, mark-to-market, EBITDA all flow to that bottom line, which you've now seen come through in 2025 and now continue on into 2026. That next step of that evolution was to evolve into building out our own strategic private capital initiatives. We closed on our inaugural data center fund, which was upsized to, call it, $3.25 billion of equity, so $10 billion of total potential spend.
We're gonna build upon that arena as a funding tool for digital, but also an incremental arrow in our quiver as we think of external investment and growth, be it M&A or large portfolio acquisitions. What that does is essentially shares on the development vehicle, shares some of the non-cash producing years. All those projects I mentioned, 2026, 2027, 2028 deliveries, billion-dollar buildings don't show up on one day. They don't get built in one day. They don't get NOI in one day. Maybe the first megawatts of a 100 MW building could start in those years, but it could take the entirety of the next year, if not longer, for the building to be fully completed. We're sharing the CapEx load with our private capital sources on the development.
We're recycling stabilized assets, harvesting development profits on the stabilized versions of that as well. We're moving into fee models, be it development fees, asset management fees, property management fees, that's really pulling forward, call it monetization of these asset opportunities. Yes, in 2029, we may have a lower ownership of said project, but we've garnered a tremendous amount of near-term, call it FFO to our bottom line along the way. These were all incremental steps onto a funding model that's gotten our virtuous cycle and our cost of capital spinning in the right way that has allowed us to grow faster and do things in a bigger fashion. I think as we look in the future, it's gonna be more of this fund management version of what we're doing and less of these episodic joint ventures.
We had a question come in, through live QA. What's the opportunity to aggregate capacity in well-located data centers upon tenant renewal and get a much higher return on the bigger blocks?
We don't have a lot of churn is the issue. I don't think there's a tremendous opportunity to essentially kick our customers out. We are obviously moving folks to much higher cash mark-to-market or market rates, as you've seen flow through our mark-to-market. Even as these customers are not saying, "I don't want this capacity if I have to pay the higher rate," they're saying, "This capacity is precious, and I'm more than happy to pay the higher rate to keep in that capacity." What we are doing is when we look at strategically growing the business when it comes to hyperscale, which I think is the root of this question, we're not buying land or powered sites that are the last to be developed in our portfolio.
I look at what we did in Charlotte with this 400 MW site adjacent to the airport, worked with our utility partner, got creative. I think we were talking about Charlotte in maybe 12 months or less year's time. Now talking about 2027 deliveries of that capacity. Really rapidly seized upon hard-to-deliver capacity markets. I think that's an important nuance. We don't pick these harder markets, be it Atlanta, be it Charlotte, be it Northern Virginia, be it Dallas proper, because we like to frustrate ourselves. We do that with a view that in our business, location does truly matter for the longevity of the infrastructure, the customer base, and the cash flows that we're driving. We're talking about the cloud that doesn't live everywhere. It lives in availability zones with radius restrictions, in set metropolitan areas.
Our strategy has been widening out to new versions of that and adjacencies, but we stuck to where it's harder to do business, places where we deliver a lot more value for our customers and helps us drive those outsized returns for our investors.
Building on that point, in your latest investor presentation, it looks like you've once again updated the capacity that you're planning on in the future in terms of gigawatts. Can you share with us the update of, you know, what's evolving in terms of the capacity that you have a pipeline to? You know, how much of that might be secured or committed from the utilities? How much can come online on an annual basis, like on average? You know, is it all available or is it, you know, in chunks? How should investors just think about that?
I think the total's We operate 3 GW today. We've got another 5 GW from nearing delivery, under construction, shells, pad ready to land. All of that is owned on our balance sheet. All of that has a path to power, and I think the tail end of that, the last projects in that 5 GW. I can think of, maybe 500 MW of that is a 2030, 2031 type timeline. This is not Field of Dreams speculation and throw out a gigawatt that may never manifest itself here. These are real type deliveries of capacity.
Does that mean that you're gonna take additional steps to accelerate capital availability to try to fulfill this supply with demand sooner?
We are the farthest thing from capital constrained, right? We're constrained by finding good opportunities and the timelines our customers needed because of the markets we focus on, right? I'm not saying we're not capital constrained in jest or in a reckless fashion. Between a very under-levered balance sheet versus our leverage targets, tremendous liquidity, and a deep well of capital we have from private capital sources, we're able to activate more of our development pipeline and also keep replenishing it in an arena where the stakes are getting bigger and bigger and bigger. I mean, land parcels by themselves could be $250 million, $500 million for the land. Listen, that scaling laws that we're seeing accrues to our benefit, right?
Because our utility partners are realizing they want real partners across the table that can deliver, that have the financial wherewithal, that are committed to the projects, that have done this before. That lends itself to our 20+ year track record, and the size and caliber of business we are.
You mentioned the evolution of stabilizing assets and recycling capital. As there's more private capital partners involved with these assets as well, does something need to evolve in terms of the ownership of data center assets to find homes in the future for more of the stabilized assets, you know, to free up even more capital and keep this cycle going?
I'm not losing a lot of sleep that there's gonna be homes for what I value as an attractive asset class if you call it prosecute the execution the way we're doing it. That goes back to not all data centers are created equal, and I don't think all data centers are gonna be treated equally in this, right? What we've done and shown with our strategy is tremendous resiliency in the markets and customer diversity. We've gone through cycles in data centers. We've had churn in our data centers and release capacity, right? We've done this multiple times, and we've proven successful due to our strategy of where we invest in our platform. A lot of folks in data center land today are not doing that, right?
Those, I believe, will be at the end of the queue for the valuations and the access to the capital they need to recapitalize. Where we are essentially investing, and others are investing, we're not solely in this strategy. There's many others doing this the same for hyperscalers. I think there's gonna be numerous homes, as capital, quite honestly, probably rotates into this category given there's the growth trends, not only stability, but growth that we're seeing in it. We think we're gonna be at the forefront of that in our private capital initiatives.
There are clearly more headlines from federal and state participants about utility costs and the impact on households. We had a question come in specific to that. On your views of any potential policy or changes just requiring data centers to build their own power plants or bring your own power broadly.
At the risk of doing a public service announcement, I'd like to bust a few myths for the industry. It was important. We at Digital are actively strengthening the grid, providing resiliency to power. Those hot summer days or those cold winter nights, we can turn on our generators as our prime power source and provide needed relief where other constituents or customers of the grid can use it. The studies from Berkeley can show that states with data centers have had better power outcomes than those without it. When it comes to water, I think the data center uses the same amount of water as a McDonald's. We have 300+ data centers around the world. That's less water usage than 18 California golf courses. There's 16,000 golf courses in just the United States.
That's half there are in the world. For every job we're creating, in our data centers or building our data centers, plumbers, electricians, construction workers, after all that, there's six more jobs being created somewhere in the community. I think I pride Digital as being a leader in that category. Not everyone's doing it, but a lot of us are doing it the right way, and we're gonna continue doing it the right way. Now, we live in a world where folks don't let the facts get in the way of a good story when it comes to political vibes, and we are working on making sure our voice is heard. I put this as it's getting harder to do our business, and that means our value to our customers is increasing, and that continues to strengthen our moat, in our business.
How much can temporary power solutions, bridge solutions, or, you know, behind-the-meter solutions accelerate your ability to bring inventory to market?
We are actively working on numerous projects, but the key to that is bridge. The duration of that bridge could be a couple years to several years, but it's always getting back to that grid as the end state for the infrastructure. That's a product of the markets we operate in, and where we see the long-term customer preference in the resiliency of our workloads. We're supporting mission-critical workloads, supporting hospitals, pharmaceutical medical research, health and safety, financial, buying your coffee, your Charles Schwab account, everything that's running society today and for your grandchildren. The mission-critical element I cannot stress enough, hence, the having resiliency in our infrastructure and the grid is a piece of that resiliency in the long run.
You discussed the differentiation in markets and the customers that you're pursuing. What does that mean in terms of pricing and development yields that you're achieving maybe relative to what you see in the industry? Are those getting better, staying the same, or maybe coming in a bit?
I think your question is more towards the hyperscale piece.
I'm sorry. Yes.
...piece of the equation.
Yes.
I would say we've led the way on yields, but that's a product of our strategy, where we didn't just chase volume for the sake of volume and accept price, and we didn't just go to where it's easy to offer customers a business and have commodity-like competition, and we're able to drive outsized outcomes for our customers that came with outsized returns for our shareholders, which you can see on our development schedule. We're living in a world where supply chains are tremendously tight. People are moving at, quote, "the speed of light" to build infrastructure, and there is inflation hitting our category. I think Digital stands shoulders above many of our competitors, given our track record, consistency, our supply chains, who we are, our say-do ratio.
We're probably leading the way in shielding those cost impacts, but they're not, we are also absorbing them. At that same time, we've been able to activate rate outcomes that have maintained, grown our rates last 24 months, surely, but in the current backdrop, maintained our returns that are hundreds of basis points above many of private competitors.
Pivoting over to the under 1 MW business, you know, how does Digital view this business strategically in terms of the value that it contributes to the overall company? What are the ways, in addition to the organic growth that you talked about earlier and the sales momentum, that you're trying to make this a bigger part of the company?
We are long-term believers in the durability platform effects, necessity of call it hybrid IT for cloud, private and public cloud, same will unfold for artificial intelligence. That has been our guiding light of the last several years. That's what our double-digit and our colo or enterprise bookings has been about. That's what brought our customer count to nearly 6,000, compounding double digits. We see that as a virtuous cycle with our major hyperscale customers. Because whether you're a major financial institution, pharmaceutical company, manufacturing company, retail company, your architectures are gonna have workloads with CPUs and GPUs that sit inside one of our 300 + data centers that are rapidly growing, and you're gonna be consuming numerous cloud service providers and likely AI solutions as well.
We've had tremendous success in not that long of time with, I would say, the wind in our face still, despite the odds. There's three things that I think are gonna propel us, the wind at our back, and take us well above the goals we're seeking. One is partners. We ended the year on a fantastic note on partners. You heard some of the anecdotes in our prepared remarks throughout the year. We can grow that to a much bigger sales force enabler for our customers.
We made some moves in our partner organization this past year in terms of leadership and others, and I think that's gonna be a big lever of growth for us that's gonna allow us to continue to accelerate our bookings in that category, deepen our relationships, multi-site, multi-geo, multi-product, in a category where the partner's been yearning for another option. Two is our technology. We are very fortunate, or serendipitously blessed that, in the days of AI, we started getting ready for this many years ago. Many years ago, we had a fragile IT house 'cause we came from more of a traditional real estate, call it business processes and technology stack. We stacked upon it acquisition to acquisition, globalization scale.
Then we didn't have the time until the last few years to go back and streamline our business process, drive consistency, set up data governance and cleanliness. In our digital one journey, we're essentially getting AI ready, and we're gonna be removing friction from our sellers, friction from our internal customers, friction from our external customers and partners. Technology's gonna certainly infuse that and be incremental, called gas on our fire here to go faster and do more, and scale our resources. Lastly, it's about growing what we can do for our customers. It's not just doing more of the same and billing faster. It's not just saying we're gonna do our day jobs, which we should have been doing already, 'cause we're already doing that, a great job of that. We're expanding our addressable market.
We're expanding into more colo into Atlanta, Charlotte. Just this week, the most highly connective business in Sofia, Bulgaria. We're bringing on capacity in Crete, Rome, Barcelona, Lisbon. Just recently entered Malaysia. Last year, entered Malaysia, Indonesia. We're incrementally helping our enterprise customers in more places they need to be, so expanding our addressable market at the same time.
When you think about the momentum that you've been generating in the under 1 MW category, at times I think you referred to it as taking share. Where are you taking share from? You know, how should investors think about the growth of this overall TAM in the under 1 MW category that you're going after?
The TAM in that market is very, very large, but it's growing high single digits as a market today. I would say it's still yet to be infused by supercharging of AI growth, although we're coming off two consecutive quarters with the data points reporting to the beginning of an inflection. It is a massive addressable market. If you add ourselves up, and the company ahead of us, we're, I would say, call it less than 30% of the market. There's a long tail of competitors. Enterprise customers, global businesses are consolidating. They're moving from legacy system integrators, telcos, whole host of places, and other regional smaller scale platforms. The good thing is they want to not put all their eggs in one basket.
We are one of two in that category, and I believe in that category we are taking share from everyone behind us and the one group ahead of us. You can see that in our numbers, right? Look at our bookings acceleration, up 35% on a base in that category, just the 0 mega interconnection. That is much smaller, so we're essentially accelerating in that category and taking share from both.
Another question that we get in this under 1 MW realm is the role of inference. You know, what you're seeing in terms of AI as a % of bookings, how it contrasts with your hyperscale business, and, you know, any predictions or things that investors should look for to kind of, you know, better appreciate the inflection that may be coming for your business from that.
So it's a category where we're up 35%, adding call it 500, 600 new customers to nearly 6,000 in total. The last two quarters, we did see the starts of data points of that inflection. We'll call it 18%-19% of that category. Of all the leases we signed it in that, the nearly $100 million of leases we signed, all the leases, 18%-19% AI related. We assess that via power density, liquid cooling. We see the GPUs. We know what they're using it for. I wouldn't say two quarters makes a long track record, but we're starting to see it move that way.
I think there's gonna be a long, long tail of this because I know how hard it is as enterprises move to cloud, move to multi-cloud, adopt AI, think about privacy, think about data sovereignty. I think we're very much benefited from that we're playing this across the entire spectrum. Training is landing in our facilities for sure, but it's in our facilities in Northern Virginia that has all the clouds and enterprises and service providers there as well. Inference is gonna happen in those facilities. Enterprise cloud's gonna happen in those facilities. Enterprise digital transformation and enterprise AI are gonna all happen in these facilities. I think it's gonna come back to the latency, and location's gonna matter in the end of the day. As just like cloud, the service level agreements that nobody's gonna want their AI to go down.
When the hospital room surgery is going on and AI is being part of it, you can't have the AI go down just like you can't have the cloud go down. I think it's gonna be a large addressable market and you're starting to see an inflection, but I think we're setting up for a multiple year growth in that category.
Can you give us an update what you're seeing in terms of renewal spreads in the under 1 MW versus what you're seeing in the over 1 MW? For the hyperscale, are there any significant constraints because of the contract languages from maybe older dated contracts?
In less than a megawatt category, which we've done a great job growing, and it's consistently in the front end of our expiration curve. One, two, three, four, five-year type they're always coming up. We've been consistently pushing that toward call it the four and a half, 5% mark-to-market consistently and pushing rate. The pricing on that is very focused on, where we are in the market, where we are with the customer. Are we trying to grow wallet share? Are we trying to get them stickier to our platform? Are we in a market where someone is ahead of us in price and we're peer on capabilities? We narrow that gap right away.
Is it a market where they're ahead of us on price, and they're a nice moat to stay a little bit further behind as we try to scale and build more customers, build better ecosystem in that market? Consistently 4% or 5% type growth. On the greater than megawatt, you can look at our lease expirations, and the market rates have inflected tremendously, continue to go up, and the expirations keep going down. I think we're at $130-ish or something weighted average in total, lower than that in some of the nearer years. We did have legacy agreements that were signed prior to this time of supply demand fundamentals that had some renewal caps.
I think as we first started estimating what was gonna actually be activated versus not, I think we're seeing more and more of those caps not being activated because the infrastructure's changing. The customers are mixing up GPUs and CPUs in the same data halls. They're pushing power densities. They're moving to liquid cooling. They need to change service level agreements, and all those things open up a contract and essentially have renegotiating the opportunity on the caps.
Very quick, rapid fire. Same story on OI growth for data centers overall sector-wide next year in 2027.
Go ahead, Jordan.
5%.
Market rent growth , did you say?
Same story.
Okay, same story. Yeah.
Yep. Same number of public data center companies a year from now. More, fewer, the same?
More.
Thank you.
Thank you.
Thank you very much.