Digital Realty Trust, Inc. (NYSE:DLR) Bank of America 2024 Media, Communications & Entertainment Conference September 4, 2024 8:50 AM ET
Everybody, I'm Dave Barden. I head up Telecommunications and Comm-Infrastructure Research for Bank of America. And I'm here with the Chief Financial Officer of Digital Realty, Matt Mercier. Thank you for joining us, Matt.
So gosh, you guys -- what a difference a year makes. It's been pretty strong. I think maybe we're going to talk mostly about the business and operations, but I think it's worth kind of just taking a step back and saying, you've given some guidance that you want to go from the 0% to 1% AFFO per share growth that you're expecting to generate in 2020 for to somewhere in the mid-single digits in 2025, talk us through how that happens?
Matt Mercier
Sure. So I mean, just to give a little bit of the history, right. I mean we've done a lot over the last year to, I think, put us in a pretty solid position where we're in today in terms of -- we've done a lot with the balance sheet in terms of bringing down leverage from 7 times, now we're down to 5.3 times. A big part of doing that was part of an effort to bolster and diversify our sources of capital. And that was a lot of work that Greg in particular and his team did around our capital recycling efforts as well as establishing and growing our joint venture platform which we did throughout last year in terms of our number of stabilized joint venture opportunities and executions as well as establishing some of our first, what I'd say, are pure development-oriented JVs. So that's been part of this journey that we've been on over the last year and half, I think, to really put us in now what I'd call going from a little bit of defense. Now we're a little bit more on offense. And that's been -- all those things have had some impact on bottom line growth in a time where operating fundamentals have been about as strong as I think we've seen in a number of years where pricing's now swung in our favor. We're seeing positive renewal spreads.
So I think that sets the stage now that we're -- we've gone through the majority of that capital recycling effort. We've got positive pricing across the majority of our major markets. That's flowed through now towards our same capital stabilized portfolio growth, which has been positive for the last two years. We were plus 5% last year, guided towards -- we've been guiding towards, call it, a 3% midpoint, which had some power pricing impact that we've talked about. So -- and we expect that to continue as part of that stable base of growth into 2025.
And then on top of that, we've got our development pipeline which -- where we're seeing yields that have continued to increase. Again, part of that due to pricing that we've seen improve given where the supply-demand fundamentals are. So we've talked about all that -- putting all that together, we've in essence talked about a baseline of 5% growth in 2025 and we see an opportunity for that to improve as you look out, and you see not only our backlog of development that will start to come online in 2025, into 2026 at those higher yields. But you also see through our lease expiration, you see, I think an opportunity for improved organic growth into those outer years in light of where the overall pricing dynamic is throughout most of our major markets today.
David Barden
So kind of a mid-single digit for 2025 and kind of an expectation that some of the offense that you're playing today could manifest itself in kind of drifting higher from there into the high single digits?
Matt Mercier
Yes. I mean I think you see us based on where again, supply-demand is today in terms of pricing, where our setup is in terms of where we think renewal spreads can improve in outer years. I think, again, that 5% sort of baseline growth is setting us up for accelerating growth beyond that.
David Barden
Okay. So that was a good kind of setting the table big answer. So let's dissect it a little bit. Equity issuance has been both a defense and an offense it seems like. When you were issuing at 97 to address balance sheet issues that were surfaced by the rating agencies, it was defense, but more recently it's been maybe taking advantage of the strong stock price performance. What is -- what are the criteria for the potential for additional equity issuance or are we done and we're just not going to do it anymore?
Matt Mercier
So look, I think we've -- I think you're right in terms of like we've been now. I think now we're in a position where the balance sheet is in a much better place than it was a year or 18 months ago. And so we now have, what I'd say, is a full -- our full set of available capital services available to us. And that -- those available capital sources has also been growing. Again, back to now, now that the balance sheet is in place, we're going to be able to be back into the debt capital markets. We're -- now that we're out of -- largely out of sort of the JV capital recycling, I mean we'll still do some, but some of the material ones. We'll see improving EBITDA, which gives us an ability to leverage that as well. Growing the cash flow from the business will be another source. And I think then you've got equity as really part of what I'd call and what we've talked about over the last several quarters has been largely demand-driven. So largely looking at that development pipeline that's getting 10%-plus yields today and an option in terms of funding that potential in addition to the array of other capital sources that are available to us.
David Barden
So just to parse that answer down a little bit, there's always this appetite, of course, to say, okay, well, there's this great return opportunity. But if I chase that opportunity, I pushed the growth out another year or two. And that's been a criticism of Digital Realty looking backwards was that there was never a deal that they didn't like to do, sometimes they were dilutive, that pushed out the growth and the growth just never showed up. Is there a commitment of focus on management's part to achieve growth in addition to taking advantage of these developments?
Matt Mercier
Yes, yes. I mean that's a good point, Dave. I mean we're -- one of those other elements on the equity is where we're making sure that whatever we do our -- in essence, our first, second and third priority is generating bottom line growth in 2025 and beyond. So that's going to -- we're going to take into consideration whatever we do to fund our development, our growth is not going to have a material impact on being able to generate that bottom line -- that bottom line growth that we've been talking about, in particular, for 2025 but also beyond.
David Barden
Right. And I think that is important. So you talked about the positive developments on the balance sheet, getting from, I think it was 7.3 down to 5.8 now in terms of net leverage to EBITDA. The tower companies have comfortably lived between 5 and 7. You have a publicly traded peer that's kind of 3.5 to 4. Like where do you want that 5.8 to go, where is comfortable for Digital Realty to live on the balance sheet?
Matt Mercier
Yes, so we're at 5.3. Sorry, I read my 3 and 8 wrong. I didn't bring my reading glasses. And so I mean the short answer is we're -- we've had a long-term target of around 5.5 times net debt to EBITDA. That's been over, I think almost for the majority of the time we've been investment grade. So -- and yes, we've had periods of time where we'd go up sometimes be below. But I think that 5.5 times is -- has proven to be kind of a good place for us on average to be from a leverage position. And part of that's taken into consideration, again, where our discussions in history with the rating agencies, our views to where sort of an optimal capital, where it is from a pricing perspective on debt and equity and kind of looking at, okay and we might have some ambition to get BBB+ and what does that get us. But we think, again, as of today, 5.5 times is the right leverage. You look at our operating business versus -- you talked about some of our peers, I think we have an ability to have slightly higher leverage because we, generally speaking, have longer-term contracts. Yes, longer-term contracts, we own the majority of our portfolio. So that gives us some -- again, also some benefit there from a leverage perspective. And wrapping that altogether, I think we still view 5.5 area to be the right sort of long-term position from a leverage perspective.
David Barden
So I guess my last one on -- specifically on the balance sheet was just you've got about $1 billion and maybe change maturities coming due in 2025 and 2026 each. What do we think about doing about those, just refi them, if so, what do you think the rates we should be expecting?
Matt Mercier
Yes. I mean -- so we've got -- as I say, we've got roughly $4 billion of liquidity as we look at -- in particular, as we look at those maturities, the most likely path to some form of refinancing that as I'm sure most in this room or others, the debt capital markets have been alive and well. And now that we're -- and again, this is coming back to now that we're in a position where our leverage is back in a good place. Our ability to be able to issue, especially in this case, it's going to be leverage neutral. We'll have access to our global markets that we've had in the past across USD, Euro and GBP as well as some other currencies that we've issued in. In terms of where rates are, that's somewhat dependent on the currency that we issue in. But I'd say between Euro and USD, which are sort of the most likely candidates and where the majority of our debt capital is coming from, you are anywhere from, call it, the low 4s to the mid-5s today on a, call it, on an average 10-year type basis, which is likely where we look to issue as we kind of look out over the next couple of years and look at our debt maturity, making sure that, that stays laddered and acceptable over a long period of time.
David Barden
Perfect. Alright. So let's just get to the JV side of things, kind of a big ambition. One of the things you guys executed on and continue to execute on. Where are we in the $7 billion Blackstone relationship?
Matt Mercier
Yes. So that was a relationship that was established at the end of last year. That $7 billion is over, in essence, covers three markets and we -- and it's around -- ultimately would cover around 500 megawatts of capacity and that's going to take several years most likely before that $7 billion is fully deployed. So we closed in the first half of 2024, we closed on call it, Phase 1 of that JV, which was sites in Paris, in Manassas. We expect to close the second phase of that joint venture in the second half of this year, in which case, we have the -- in which case we'd close on the full transaction. Of that 500 megawatts, we probably expect around 20% that would come online in 2025. And then depending on demand, demand-driven activity as well as ability to bring power to most of those sites, which is available today, that will dictate sort of the remaining build-out of that capacity over the next couple of years. So we're well on our way with that transaction. I think we've -- we're in a position where we've got a pipeline across all of those sites and spaces and we expect to be able to have a positive outcome with that.
And again, it's kind of taken a little bit of a step back. I mean, the reason for doing that is the hyperscale demand currently is quite large. And I think for -- the way that we've looked at it is for digital -- a little bit back to your -- some of your equity questions, for digital to be able to take that on itself would mean ultimately be -- you're in the capital markets more. You've got -- that has more of an impact on near-term growth because you do have to fund that. And I think so between that and just the size of the opportunity that we think is out there, we saw there's a prudent way to manage that to bring in additional capital partners to partner on that development activity and that funding across what is a large and growing hyperscale universe.
David Barden
So along those lines, do you have ambition to find new JV partners or maybe the other way around, given the space that you're in, are you getting a lot of phone calls asking to be a JV partner?
Matt Mercier
It's -- I would say, we've done a lot of work. I mean we've cultivated a number of JV partners. Blackstone was one of several JVs. JVs are not new to us in terms of our overall landscape within our organization. I mean we've looked at joint ventures over -- in terms of how we incorporate those into our business from, generally speaking, kind of two lenses. One is call it, financial capital partners, which you could put Blackstone in there, although they're somewhat turning into an operator themselves, obviously, as well. But for this discussion, we'll say from capital partner perspective is we're managing the majority, if not all, of that joint venture activity. And then we have other joint ventures within our portfolio that I would call are more operationally strategic in terms of where the partners that we're looking to leverage. And those examples would be like Ascenty down South America, Teraco in South Africa, and Mitsubishi in Japan. So it's in places where you're looking at local -- need to have some level of local expertise, local knowledge to help navigate those country-specific rules and regulations or they're helping bring some level of customer and local enterprise expertise hopefully as well as some pipeline into that mix. So bringing some additional elements from an operational and strategic perspective into those joint ventures.
So we've done -- we've now done a mix of both of those within our overall company. Both, call it, those operational strategic data as well as financial and capital partner oriented joint ventures. In terms of going forward, I think there's still -- there's definitely an appetite. The data centers have, over the last years, I think you opened what a difference a year makes. The number of, I think, people that are looking to enter this space to put capital to work within data centers has only increased. So there's no shortage, I think, of people and partners that would be willing and able to be -- to put capital to work within this space. We're not in that position where we're in the same place we were last year. So I think we can be a little bit more selective in terms of the types of opportunities that we put to work from a joint venture perspective versus keep on balance sheet. Again, keeping all the things in mind that we've talked about so far about keeping an eye on bottom line growth. So I think joint ventures will always be part of our strategy and part of our capital funding needs going forward. But I would say it's likely that it would be -- we won't have the same likely level of activity that we've had at least in the last, call it, 18 months in 2025.
David Barden
So I guess one critique of the JV structure or strategy is that, well, if you only waited given your balance sheet, you could have done it a 100%. I think the counter to that is hyperscalers and AI wait for no man. These things are going to get built. Could you talk a little bit about the returns that you get from these JVs relative to on balance sheet versus JV because you got the management fee component that is incremental to that?
Matt Mercier
Yes. I mean, so the -- one of the other benefits of JVs is that, one, so we're still participating in these ventures. We're generally, call it -- some are 50%. Those have been more on the strategic side on some of the more, call it, financial or capital-oriented JVs. We're generally in the 20% in terms of our ownership. So we're still able to participate in this growth and development. And again, we're also not JVing a material amount of our overall capacity that's available to us. So using the Blackstone as an example, that was ultimately looking at roughly -- I think it was between 15% and 20% of the capacity that we had available at that time, right. So we're still -- we still have 80% of our development capacity that we're keeping on digital's balance sheet at 100%. So we're not JV in the entire -- in essence, we're not JVing the entire company. We're still keeping the majority of that opportunity for us.
And two, to your point, we're getting management fees on top of that, that's helping to -- that's accreting to our overall return on that capital invested. So those are specific and dependent on each joint venture and what you're able to negotiate, but you're generally getting a couple of hundred basis points, if not a little bit more of incremental return on our capital as a result of the fees that we're able to generate. And in some cases, on top of that, not in all joint ventures, but in some we're getting promote opportunities as well, which are more on the back end.
David Barden
So -- thank you for that. So let's maybe shift gears a little bit now more to the operating side of the business. So just to start off, I think after first quarter, you had record leasing. I think Andy came out the CEO and said, that roughly half of that was AI related. And I think that the market has really gravitated to your company, your business, your stock because there's a sense that you are in a really good spot to monetize this AI training environment. Could you kind of just describe a little bit about what the conversations you're having with the hyperscalers look like and how many people are showing up now that there's all this capital that wants to be part of this industry?
Matt Mercier
I mean the demand environment is about as robust as I've seen in my time at Digital, and I've been here for, let's say, over a decade. So this is about as robust as I've ever seen. And it's coming at a time -- and that demand is not just -- I know AI gets a lot of the headlines, a lot of the attention. So I think it's important to kind of remember that this was already starting even last year where we were still seeing a robust amount of demand, and that was from the more traditional workloads, cloud, digital transformation. AI came along and was adding on top of that. And that was also happening at a time when supply was starting to become more restricted across a number of our major markets. That's for a variety of reasons, but the one that gets the most, we'll call it, attention is around power constraints. So that has led to some of the things which I'm sure we'll talk about. But in terms of the overall demand profile, I mean it continues to remain robust. As you mentioned, as a couple of data points we had, we had a record quarter. And the first quarter was over 250 million of signings. We followed that up in the second quarter where we were a little north of 160. So year-to-date, we were at, call it, over 400 million of signings, which is in essence double what we did the same period first six months of last year. And that has -- the pipeline continues to be robust. We've talked about -- I think Andy was asked a couple of times or maybe like a few times over the last two quarters, what can we expect, do you expect another record, and I think he's used some version of the words like, it's hard to do back-to-back records, but in terms of what we're seeing in overall pipeline, it's not out of the question that we could see another period this year where there's the potential for -- to be able to top what we did in the first quarter. So in essence, just some anecdotes in terms of like the demand environment continues to remain very healthy.
David Barden
Is -- and I'd be weary of kind of setting the expectation to hit the lottery forgetting this one quarter right in terms of leasing. But is 200 million a quarter in terms of new leasing revenues, kind of a good baseline, the new normal, if you would or is it -- obviously, it's going to be lumpy, but is that like the new normal?
Matt Mercier
I mean we just did like -- I mean we just did 160 million so you're already -- I don't know if I would say 200 million is like a new normal.
David Barden
That could be 250 million though. So, I'm just averaging.
Matt Mercier
Yes, got it. I understand. Look, again, I go back to -- if you look at our -- if you look at the kind of two lines of business we have, right, which is we have our 0-1 megawatt business, where we've done, call it, close to $50 million a quarter for the last probably four to six quarters. That -- and that business tends to be more steady, right. So we've got that stable base of signings. We've -- as part of our business, we've continued to, I think penetrate, grow, and take share within, call it, that 0-1 megawatt, which is more of the retail-oriented, enterprise-oriented workloads. And that's still a focus of ours. We haven't taken our eye off the ball in light of what has been a market that gets -- a lot of the discussion gets for the larger deployments.
You then shift over to the greater than 1 megawatt. I mean even with AI, that continues to be -- that continues to be and we expect it will be a lumpy business, but it's also been a lumpy business that's been boosted by what has been an AI demand wave over the last several quarters. So I don't want to -- we don't really give signing guidance on in terms of a quarterly cadence. I mean you've seen, to your point, what we've done the last two quarters. We feel good about the pipeline, and we think right now, we're in an environment where there is very healthy demand for the greater than megawatt segment and steady and growing demand for the 0-1 megawatt segment.
David Barden
So could you talk a little bit about -- I get this question a lot about, okay, two things. One is, how much different is an AI data center to build than a regular data center, if there is such a thing, the cooling, all that sort of stuff? And then the second question is, how long is it going to take for these record bookings to turn into record revenues?
No, I got the -- I mean in terms of, we'll call it, the development design, the big difference, which I'm sure is not saying anything profound is around density. The AI workloads are looking at higher densities in terms of watts per square foot watts per cabinet.
David Barden
So that mean you build smaller buildings or you just have a much bigger power need for the same size building?
Matt Mercier
You generally -- I mean, there's a couple of things there. One is, I mean, we've been able to satisfy AI workloads within our traditional designs today. So we've been able to satisfy that augmenting with things like rear door heat exchangers in places where we do have water, we have the ability to bring in liquid cooling. I mean, liquid cooling is still in -- we expect that to be what will be the go forward, most likely main type of cooling for an AI deployment. But today, that's still -- it's still relatively new in terms of its deployment within most data centers, not just digital specifically. So I think you're going to see a liquid cooling largely more in new deployments, new builds. But within our portfolio, we've also got the ability because we've got sites that have water cooled chillers, but we have an ability to be able to over time augment those to satisfy liquid cooling as that comes around.
I think the other point that's important is not all data center needs, data center applications, data center workloads are AI. So we expect to be able to -- we're continuing to see cloud general enterprise workloads. Not every type of application is going to need a GPU in order to service it. So I think that's part of, again, an overall portfolio strategy and mix that we expect to see going forward. So I think that's where -- and Digital has been in the business and had been dealing with more of the hyperscale customers for a long period of time. We had already been, call it, part of what has been a growing density need. So the hyperscalers were typically the ones that were pushing density envelope, they were utilizing a higher percentage of their workloads for a period of time. So we had sort of a center stage in terms of where densities were already going, having a little bit -- being able to be a little bit more in front of that in terms of we're already building larger facilities for workloads that had densities that were starting to increase over time because it's the same type of customers.
And I think these hyperscale customers are also looking for fungibility in their workloads, to be able to toggle over time between what might be AI workloads and what might be more traditional workloads. And I think we've had the experience and the benefit of being able to design and accommodate those different workloads given sort of our heritage and the amount of data center space that we've designed and delivered over the last, call it, decade plus of our tenure.
David Barden
And then how long does it take to get these things up and running?
Matt Mercier
Yes. I mean -- so that's -- ultimately, that's going to depend on the size of facility, bringing power to the site. But let's just -- if you're assuming that we're -- we start from land today, we have all the power we need to be able to bring that site online. I mean on -- I'll say, on average, we can usually bring on a full data center within again, from land, land to production within 18 to 24 months. If you've already got the shell, it's even faster. I mean generally speaking, that could be the time range between 9 to 18 months before we're delivering the capacities, certificate of occupancy, and we start commencing revenue.
David Barden
So Digital Realty has been in such a good spot and able to raise prices because not just that the demand curve has shifted, but that the supply curve has not shifted. And Economics 101 has led to these price increases. And the supply has been constrained. I want to kind of walk through, what is and what isn't a constraint and maybe they all are. But number one is land. And what I think one of the -- and this goes to the power question next. But one of the things I think Andy has said, that we thought was super interesting, was that a few years ago, maybe 65% of the absorption was in, say, the top five markets in the U.S. domestic absorption. And that what he saw last year was close to 80%. That people -- the computers are pack animals, they like to live, work, and play near each other, lowers latency, reduces instances of fault. Is land an obstacle for others, for you, for everybody?
Matt Mercier
I mean I think that's -- ultimately, that's somewhat market dependent. But we are -- I mean, you are starting to see land in general -- I mean, between land and power, I think those are some of the two main constraints. On top of that, you've got regulatory issues in some cases, which depending on whether you want to put like a version of nimbyism [ph] in there, those are additional constraints. But even in Ashburn, the amount of land that's -- the amount of land that's available is shrinking. You go across Chicago, it is becoming more constrained. That's also a power issue. Silicon Valley's been constrained for a number of years, and that's only gotten worse. That's largely a power-oriented problem as well. You go to other international markets, Singapore, that's a land-constrained place as well, but that's not necessarily new just given the size of that country. And they're also running into some power constraints.
But I would say land has become more constrained, but as of today, I think power is still generally speaking, probably the most constraining factor across major markets. And part of that mix of land and power is where you have seen some demand go over into what, I'll use the term tertiary, secondary, tertiary markets. But back to where you referenced to some things that Andy said, I think our view continues to be that we think that our focusing on the major markets is the right strategic play for us, just given our -- one, of our presence in most of these markets. And two, the diversity of demand that we expect and that we see, I think just will accrue to us over the long term and be able to generate that long-term sustainable growth that we need.
Our history has suggested that going -- while some of these markets -- some of the secondary tertiary markets could become bigger markets over time, that you're generally looking at a reduced demand set that's available for you in those markets which has more pricing volatility over a longer period of time. And so we're more cautious on those markets and sticking to sort of our core global markets. I mean we're in 50-plus markets today. So it's not like we have a limited set of places that we can choose from in terms of where we deploy and where we're able to put customers. And so we think focusing on those core markets makes the most sense for us.
David Barden
So we've talked a lot about this topic. And we've talked about the notion of land bank and then now we've started talking about the notion of Power Bank. And if you have land and power available to you, you maybe have a moat around your business, you have an advantage. And one of the concerns is that a lot of this AI training doesn't necessarily have to happen in a low latency environment, doesn't necessarily have to happen where the power is expensive. And so it could easily get commoditized, moved out to Wisconsin where the first guy to buy a farm next to a utility with some fiber could just be the next guy who gets that business. Are you seeing that happen, is there a commoditization of the industry about to happen?
Matt Mercier
I mean, I think that goes kind of back to the point I was making in terms of like what -- I don't think we're seeing a commoditization at this point. I think what you're seeing is there's a lot of demand out there and customers want access to that now. They -- as I'm sure you've heard and seen from a lot of the commentary around these hyperscale providers, their views that this is, in essence, it's a version of winner take all. They're going after it now and they're looking for whatever capacity is available in large block contiguous capacity. In some cases, that's meant that they've had to go. I think if they had their druthers, which is again, coming back full circle, they would prefer to stick in the major markets where they've already got availability zones, deployments where they can traverse between traditional workloads and AI. But in the absence of available capacity there, they've looked at other markets. And that, again, is going back to -- our view is that I think there's more risk with going to those markets over a longer period of time in terms of the potential for what other demand is out there, when renewals happen, what does that mean for pricing over the long run. So we think there's more volatility potential there. And that's part of why we haven't traversed into a lot of these newer markets head first. We're taking a view that we've got capacity in most of our core markets today that -- where we can -- and especially in light of some of the power constraints that we've got that we're able to satisfy customer demand within those core markets, we're able to, I think, generate what we expect will be healthy returns, which will accrue to our bottom line over time. And so sticking to the core markets as a result of some of those things is what our strategy has been.
David Barden
And, I'm going to run out of time, but I want to ask one more question, but this one is just on another supply constraint. What about the supply chain, even if you're building a data center somewhere in Wisconsin, where maybe power and land are not a constraint, if you don't have the supply chain, if you're not in the line, ready, you're not going to get on time?
Matt Mercier
Yes, it's -- I mean you have to be ahead of the curve in terms of where you are in terms of the supply chain and getting -- there's still -- there's always been long lead times for the major large mechanical electrical equipment. It's gotten better since, call it, pandemic levels but it is still elongated from, call it, historical levels in terms of -- especially if you're talking about transformers, switchgear generators, I mean you're at least 12-plus months out on most, if not longer, when you're talking about, in some cases, transformers and some of the major electrical gear.
So if you're not planning ahead for your -- when your capacity is going to be available, when your build is going to come online, you're behind. And so we've -- again, this is part of, I think, the benefit of our heritage and history is that we've been building hundreds of megawatts typically a year of data center capacity. We've had the experience of being in the queue with the large vendors, planning ahead for when we're expecting deliveries, putting together VMIs with a number of our major manufacturers and making sure that we're managing that supply chain so that we can deliver the capacity that's under development when it's needed and when it's contracted with our customers, given the -- call it 60% plus of our underway development today is already preleased.
David Barden
Right. So I guess my last question, just quick is, if we look backwards 10 years ago, when the cloud was beginning to boom, you guys were super well positioned. Fast forward 10 years, these giant cloud providers had a lot of negotiating leverage, they used it to their advantage. You had negative re-leasing spreads as these 10-year leases ended. Now we're at the dawn of this new boom. Do we need to be worried that 10 years from now we're going to have a problem or have we learned lessons or we know things today that we think history will repeat itself?
Matt Mercier
But -- I think where we're at is -- this -- the constraints that we've talked about in terms of power supply chain, land and others, these are I think are constraints that aren't going to be solved in the near term, probably in the near to midterm. So I think we're in a position, especially, again back to focus on major global markets, where our view is that we're in a place where we expect to see a positive supply-demand situation that should lead to long-term pricing power and that should ultimately accrue to digital reality over that time frame. We -- to your point, we've seen volatility. I mean the majority of the ups and downs that we've seen, I mean, Ashburn gets a lot of the attention just -- and rightly so, it's the largest global market out there. And you've seen the pricing go from, call it, down in the 80s, now it's in the 160-plus. So -- but it also hasn't reached the peak that it was 10-plus or 10 years ago.
So I don't think we've sort of hit the -- I don't think we've hit the high end of possibilities. I don't know but I also don't -- I don't think we'll -- we expect to see the type of volatility just given, again, what we're seeing today and what we see over the near, mid to long term in terms of the confluence of healthy demand that we continue to see across our portfolio and across most of our major markets. And generally speaking, what is an environment where it's harder to bring the supply online.