CBRE Group, Inc. (CBRE) Earnings Call Transcript & Summary

Hi, good afternoon, everyone. I'm Eric Luebchow, Senior Analyst at Wells Fargo, covering communications infrastructure and telecom. I'm very pleased today to be joined by Tim Huffman, who's the Executive Vice President at CBRE's Data Center Solutions Group. So Tim, thank you for joining us today.

Great to see you again, Eric.

So, Tim, sorry, we couldn't do this in person, but obviously this is the next best thing. Maybe you could provide everyone just a quick background of yourself and what you do within CBRE's larger organization, that would be helpful.

Sure. Happy to be with Eric and team. So our Data Center Solutions group is a global organization, and there are a number of things that we do, and in some cases, we help a client in an outsourcing process where we do market searches and negotiate colocation agreements on their behalf. In other cases, we might represent a landlord on a leasing assignment. Some members of our group kind of specialize in, what I'd call, agency work. There is a significant part of our business, that is, what we would call, capital markets, Eric. It's basically the notion of monetizing data centers, so we're helping clients buy or sell data center assets, some with income, some without, land, kind of all the elements that kind of go into it. And our Facilities Management business manages about 100 million square feet of data center space globally. So there's a part of our efforts that's kind of mechanical, electrical expertise. But big picture, we're really an advisory and transaction services shop that focuses on helping clients bridge the gap between IT and real estate.

Okay. No, thanks for that background. So we'll kind of dive into the meat of the discussion now. I think we're pretty fortunate to cover and you work in an industry that's fared remarkably well during very highly uncertain time in the broader economy. So maybe just kind of recapping, from my perspective, what we've seen this year, in particular, is a real acceleration in demand from some of the hyperscale players in the U.S. and, obviously, in Atlanta, but also in Northern Virginia, several other markets. So what are you seeing now activity-wise? Are you seeing continued demand for hyperscalers? Do you see signs of enterprise picking up? Maybe that would be a good place to start.

Yes. I would -- to your point, Eric, about hyperscalers, I would describe it this way. It's a significant uptick in activity domestically from that category of group. And it's the public cloud players, for sure. It's also some Asian cloud and social networking folks, who are creating the demand domestically. But what we see unfolding is an expansion of those hyperscale and cloud providers into colo facilities and a significant uptick in activity, where they're buying land and preparing to build buildings in multiple markets. So they've been very aggressive in the COVID season, expanding on multiple levels. On an enterprise level, kind of in their first few months of the COVID season, we saw in companies, who were in verticals, that took revenue hits being required to reduce IT spend on the back side of 2020 in part just because it's a cost center. And then I think in the reality of how critical IT is into a company's ability to function, especially when almost everyone's working remotely, that thought process evolved. And so what we saw was, perhaps, a catching-up, if you will, of that spend threshold, so perhaps not all of it landing in 2020, but a bit of a recovery around that. And so we've seen enterprises, again, active in expanding data center footprints and rethinking their strategy, especially those who had primary data centers in corporate campuses or in office buildings that they didn't own and control, needing to get environments that could scale from a network perspective and then just from an access perspective, having their people be able to get in and out. In the case where you couldn't get in and out, if your data center was in an office building, and you weren't a landlord, it's not like you have remote hands and eyes people the way you would in a colo facility. So there's just a lot of kind of operational functions that evolved and drew attention to, perhaps, weaknesses in strategy for enterprises. So we see them kind of in a healthy growth -- return to growth pattern. And then in that same category of enterprise, I would say the notion of sale leasebacks continues to grow, especially in the scenario where the client has a fractional use model, where they say, "We're using 1/3 of this facility, and we'd like to sale leaseback it and essentially bring the colo to us." Not all cases are that model, and some of it is a single-tenant model that drives it for all the right reasons, but especially in markets where there's demand that colo providers could leverage, come in and do an acquisition and then take the balance of the facility and make it multi-tenant, continue to see uptick in that as well as just straight-up, single-tenant investment sales of data centers.

All right. That's helpful. It's helpful color. So I'll unpack some of those things that you just talked about. So particularly on the enterprise side, it's such a broadly used term because there are so many different verticals within it. So if we think about the highly impacted verticals, we can probably guess what those are, hospitality related, oil and gas. Which verticals have been pretty strong this year and kind of moving forward with digital transformation? We've heard financial services, health care companies have actually done pretty well this year in terms of new sales activity. And then what are the industries where you're starting to see activity pick up that was a little slower in the depths of the pandemic? And now that we have a vaccine, hopefully, coming in a few months, they're taking this whole kind of hybrid multi-cloud IT outsourcing journey a little more seriously.

Yes. I mean you hit it, Eric. A couple of fin services in health care, I think, continued to grow, even though, in some cases, health care organizations were challenged just by all of the constraints and overloads that they had. We represent a number of health care systems and hospitals, and we see them aggressively growing as are the fin services clients we serve. When I look at, say, content providers and media companies, they have also thrived. For example, the platform we're using right now, Zoom, grew several footprints. And then interestingly, rideshare companies that perhaps suffered a loss in business during COVID but anticipating a return to normalcy in expansion did some very large multi-megawatt projects in colos. They just basically use this window to take advantage of how they might manage, operationally, a project without a huge client demand. So media and others, again, the kind of things that would make sense when you think of millions of people working from home, and et cetera. So those industries, they have been growth driven. When you look at REIT market caps by industry, so if you think of the top 5 companies, for example, in the office or industrial, the top 5 data center REITs in aggregate are 25%, 20% larger than the top 5 industrial REITs, and they're 60% greater than the top 5 office REITs. So as an asset class, it's -- data centers have performed very well. And investment in the space is growing in part because of investor lack of confidence in office and retail and wanting some safety and quality. And so then, hence, pursuing the data center market as an option.

Fair, fair, fair. So related to that, on the hyperscale side, obviously, this year, the numbers we've seen have been pretty eye-popping in terms of new lease activity. I think Northern Virginia has been -- is on pace for another record year. Do you see that kind of waning next year, as they kind of absorb capacity from these largest hyperscalers? Or do you think this is going to continue into 2021, where there's going to be takedowns of large chunks of space, multi 5-, 10-plus megawatt type deals, from not only of the large cloud companies, but social media and ride-sharing and a variety of other kind of, I call them, second-tier or third-tier hyperscalers? Like how do you think that evolves over the course of between now and the end of next year?

Sure. Here's my take, Eric. The -- if you think of the public cloud providers as a segment of hyperscale, they're going to continue to grow opportunistically in markets where they need inventory, and they don't have time to get buildings stood up. But I do think that they're aggressive pursuits of build sites perhaps could change a little bit of the demand that they might place on colo providers. And in other words, there's going to be some aggressive construction projects that will -- that are taking place and will continue to take place over the next 2 years. And that may change a little bit of the demand cycle that they would have intentionally to outsource in the colo facility. But the other components of hyperscale like media companies, et cetera, I just see it continuing to grow, candidly.

Okay. Got it. That's helpful. You mentioned, too, that you're seeing a lot of enterprise sale-leaseback activities. Where do you -- What do you think enterprises are doing a lot of that with a lot of that -- with a lot of their IT loads? Do you think it's going into third-party colos? Is it going straight to the public cloud? Like do you think there's going to be a big ramp in third-party colo leasing among enterprises, just given the fact many of them are looking to exit their on-prem facilities, and maybe COVID is only going to force that to happen faster than it would have otherwise?

Yes. My take, Eric, is that the hybrid IT wisdom has taken root. And so if you flash back 5 or 6 years in a huge rush to private or public cloud from IT organizations that had their workloads in their own facilities, then there's this pendulum swing where they begin to rethink workloads that don't eloquently transition into a cloud framework and maybe their legacy applications that are still really important to the business. So this notion of some workloads in public cloud, some in a colo and some on-premise is pretty strong, especially in the scenario where an enterprise might own their own data center and then pursue a sale leaseback. And so there's this remnant of equipment and applications that may be important from a proximity perspective. They're too risky to move or for whatever reason, they're willing to allow those to run if they can -- in their own facility if they can monetize it properly and then perhaps gain a better OpEx model, if, for example, a fellow operator were to buy it and then make it a multi-tenant. But I -- my instinct is, is that we're going to continue to see momentum in the sale-leaseback world, momentum in colo, momentum in public cloud. I don't know that any of them hinder the other. They're just complementary, and they solve unique problems that each organization will have.

Fair, fair, fair. And I guess this kind of dovetails into the next topic, which would be supply across the various U.S. markets. Obviously, Atlanta did very well, but you guys track, across CBRE, all the major markets in the U.S. So given that we had such tremendous absorption this year of new capacity, are you seeing some signs of tightening supply where vacancy rates in some markets are dropping a bit? And then what impact is that having, if any, on pricing? Obviously, it depends on what type of deal we're talking about, but at least what some of the public operators are talking about are for the larger footprint deals that there's been pressure for years, but maybe it started to stabilize a little bit this year, and that could be a result of potentially some supply constraints in certain markets. So how do you see those 2 dynamics playing together?

Yes. My thought, Eric, on the pricing model is that the hyperscale tenant in a multi-tenant colo can often be a breakeven proposition for an operator. It's a way to establish an anchor, and it generates cash flow and it might attract other enterprise users just from a proximity perspective. And hyperscalers have been very effective at the price beat-down, if you will, because they're generally good credit, and they're growth-oriented companies. So when you see pricing sub-$100 a KW, if you go do an 8- or 10-megawatt deal like that as a hyperscaler in a colo building, it doesn't necessarily mean that the balance of the tenants are all going to ride that same pricing wave. I think the nonhyperscale tenant pricing model is solid and has stabilized and, as inventory tightens, even more so. That being said, we see the primary colo operators carefully expanding in markets where they have primary demand. And so if I use Atlanta as an example, T5 for -- has a 175,000-foot shell that they're building out as a spec in a new campus in advance and anticipating demand. Stack has been expanding and acquired some land. QTS is building buildings and then preleasing them, and they continue to grow. There's just a lot of market dynamics that change the rhythm, but my distinct is that the supply, it feels a little tight in some markets, but it's all -- it's a cycle where it's tight, and then it gets overbuilt because you've got to -- takes a while to build them and a while to fill them up. The horizon for these things, I think, remains very strong, Eric, and you guys know this so well. If you look at the Data Center Index, it outperformed the S&P by 2,400 basis points so far this year, and it outperformed the REIT benchmark by 4,900 basis points. So we see the sector and its view by the investment community as being a really safe, strong bet.

Yes. And that's even with the recent pullback in some of the stock prices. It was up even more a couple of months ago. So yes, no, it's been a really good year for the stocks, which we're encouraged by. I guess the other factor that's had a big impact on the data center space is this private capital that has really flowed into the sector. So we've seen a lot of joint venture activity, M&A activity. So what are you seeing in terms of -- where are private capital players trying to put their dollars today? Is it doing maybe joint ventures with some of the larger players for stabilized assets? Do they want to do M&A for big development projects where there's potentially a lot of growth opportunity? Where do you think they're putting their money today?

Yes. There's about a dozen, what I'd call, kind of domestically focused investors that are pension fund, private equity, institutional, that are very passionate about the space. And they like it because of the infrastructure component. It's not just a building with some dirt that serves a certain purpose. And so -- and then there's probably another half-dozen sovereign wealth and foreign investment entities that pursue the space with some passion. And in some cases, they're looking to buy income. They may want to buy an Amazon data center. They may want to buy a sale leaseback that's a single tenant from a corporate entity. But in other cases, they're investing aggressively into colo operating companies like Macquarie does in a line. And then I look at what happened earlier in the spring where, to your point, I think, took about $300 million in. So lots of momentum to not just invest in the asset part of the data center business, but in the business itself. And I think that continues in 2021.

Got you. And do you have any sense -- I mean we talked about pricing, but that's just really one input to get you to a return -- that return hurdle that you have to hit. So some of these private players, do you have any ideas like what kind of return hurdles they're underwriting to for a lot of these deals? Like my sense is it's come down over the years, particularly with these longer-term investors, like infrastructure funds in the space. And it's probably only been further exacerbated by the low interest rate environment. We've seen a lot of deals get priced -- we've seen some private players do securitizations at like 2%, so like really, really cheap cost of debt. So has that had an impact, do you think, on the competitive intensity on the willingness of some of these providers to really compete and lower pricing for some of these hyperscale tenants because they can still get a pretty good spread even at a, call it, 8% yield?

Yes. Exactly. I mean, historically, I think the operators love being in a double-digit return world. And then kind of what's influenced an appetite that might accept something less than that has been these infrastructure funds, who are patient capital and pretty inexpensive, allowing data center operators the ability to aggressively price. So some of the recipients of, what I'd call, infrastructure fund investments have been pretty successful at winning deals at lower price points in part because the rate of return requirement of their capital has changed in a very good way.

Yes. Yes. That's fair. One thing we've heard from the data center operators, the ones that we cover, is the value of incumbency. So when you have an existing customer in a lot of locations, you're able to kind of get premium pricing, whether that's an enterprise or a hyperscaler, and that frequently, a lot of those deals don't go to the market. They're not as competitive. So you can usually extract a premium over some of the low prices that we hear about for some of these new build-to-suit or new logo acquisitions that -- in the market. I mean do you think that's certainly the case where the a lot of the embedded base of customers, the data center operators are a little more leveraged than some of these RFPs we heard about where pricing is at, call it, $70 per KW in some cases?

Yes. I mean the churn component, that customer retention element, still remains pretty strong because if your competitor is the recipient of low-cost capital and can undercut your pricing model, there's a competitive conversation that begins. But the cost to migrate and the pain and risk is so much that, in most cases, it doesn't result in an exit. It may result in a rightsizing of the current agreement. So for the most part, we see the operators able to balance the notion of a 20% or 30% write-down on the rent of a client being able to keep them because of the new clients that they continue to bring in. So it's a little bit of an onset. But the notion of very competitive pricing is appealing, especially on net new projects. But I'm not saying that our clients are not willing to move because we have a dialogue all the time with landlords on their behalf. It's just you have to count the total cost of it and the risk and the timing.

Yes. That's always been interesting. I guess it's different for every company how much it would cost to move, left the data center deployment and move it to another location, but I have to imagine it would, be pretty substantial. I don't know if you have any numbers you would put around like the cost of exiting or moving. I mean it seems not to mention the operational complexity and the risks that aren't -- maybe aren't fully quantified enough.

It's millions of dollars, Eric. And it's the physical move, the data move, the network swing that has to happen. And in some cases, a client can achieve it in an incremental reduction. They'll just start -- they'll give notice and say, "Look, we're going to terminate in 18 months," and then they begin to incrementally move, and they may still have to pay rent in that window. But it's just a way out. I think more often than not, the data center operators are generally successful at looking at market dynamics and rightsizing a client's deployment because part of what happens is the tenant's paying a dollar, for example, and maybe they should be paying $0.75. But the reality is they're only using half of what they bought to begin with. And so when an analysis starts on utilization of power and space, then in the notion of the rightsizing of the rent, there's a rightsizing of the resources the tenant absorbs. And so they may give back some space and power, and the landlord just re-rents it at a different rate.

Fair. Fair. Yes. Yes. So I wanted to switch because I know you're an expert of the Atlanta market. I wanted to get your perspective on what's happening there. So from what we can see from the outside, it's been a really strong absorption year. We've seen some large deals get signed. QTS has, I think, done an 8-megawatt, one in their last quarter. So they announced -- Switch announced that they've had a couple of customer wins at their new data center. I know there's a lot of activity around the market as well with cloud and social media companies building our own campuses. So maybe you can give us a little bit of overview of what you're seeing in Atlanta.

Sure. Multiple cloud providers and hyperscalers are actively acquiring land and planning buildings. And so those include some who are already here and several that are not here yet. And so -- and they're talking about multiple facilities, so not just doing one. There was an announcement a week or so ago about Microsoft doing a $400 million project in South Fulton County, which is the city of Atlanta, but it's just south of the city. There are many others that are kind of in the process of acquiring land and with the intention of building significant facilities. So I think what that says, Eric, is it's just an example of intention. And it sort of changes the landscape from a cloud node perspective. Today, we have a couple of nodes here with one of the public cloud providers, but the notion of freestanding large buildings that they own and operate is powerful. And I think what it does is it's that rising tide floats all boats, as the cloud nodes grow just like they do so beautifully, Northern Virginia, colo providers and other people want to be close to them from a proximity perspective. So it's very positive. The other thing I would say is our incentive package on a state level is amazing, but we're seeing aggressive local municipalities offer big benefits for real and personal property, in addition to the state's relief on sales and unit stacks.

Fair. Fair. Yes. I guess related to that, I mean even when a hyperscaler builds a self-build facility in that region, I mean are they still -- will have like edge node needs within the metro market? Like I think of something like a Google does a lot of, call it, a couple of megawatt type deals even in markets where they have a large self-build facility. We've seen Facebook do that in Fort Worth with QTS. So just because they're self building doesn't necessarily mean that they're out of the market for third-party colo. Is that right?

Exactly. And I totally agree. And they tend to be certain use driven or more network-centric reasons why they would go stand up a kit in a colo.

Fair. Fair. Okay. Related to that on the network topic. So I think one of the other beneficiaries has been interconnect dense ecosystems this year appear to have done pretty well. Equinix, CoreSite, Digital's interaction asset over in Europe. So what are you seeing? I mean do you see a lot of customer migrations to kind of the most network-dense assets in the markets that you look at? You see pricing remaining pretty firm and stable from what we can see it has been for the Equinix' of the world? And do you see a real acceleration in demand happening?

Yes. I think that the network-centric colo operators were big winners in the COVID season because network and the power and need for it became ever apparent. So we see their value prop as stronger than ever, and their ability to charge a premium for it as being consistent in the coming year for sure. In terms of growth, one of the things we see, Eric, is our clients saying, "We may be in multiple colos, and we may have our own facilities, but we want a strategy where we set up 2 to 10 racks in multiple cities in an interconnect hub like an Equinix or a CoreSite for -- kind of for network proliferation reasons and benefits and not just for running typical workloads like you would in a colo.

Yes. Yes. No, it makes sense. Okay. So I only have a couple of minutes left. So I'll ask a couple of just rapid fire ones. So in your crystal ball, do you think that kind of across the U.S., at least, I won't ask you internationally, but do you think leasing volumes will be higher in 2021, equal or lower?

My instinct is they'll be higher, and part of the reason will be there will be -- as the economy scales, and perhaps it's wishful thinking, but as, in general, as a population, we become more free to travel and do business and operate in a what was some sense of normalcy from what we knew before. I think that, that release creates momentum and drive. And my view is that the demand drivers are still the same, the constraining factor being COVID. Assuming it gets better or gone, I think 2021 can be very, very strong for the sector, yes.

Got you. And I guess one more for me real quick. Do you think there, 5 publicly traded REITs, and you have a company like Switch that's not a REIT. And do you think we'll see consolidation in the next couple of years among those players? Or do you think they all have enough of a niche that they can continue to do well as stand-alone companies?

It's a great question. I think the private players, it's all how you want to spin it, but the reality is they don't have to perform to the Drum Beat of The Street, but they do have to perform on a return level and an exit level to private equity money. And so you could argue them as a benefit either way, private versus public. I think the tenant can win either way. It's just that there may be some risk in a privately owned enterprise that you would have a different ownership structure at some point. But again, a minimal risk, in my view.

Fair. Fair. Fair. Okay. Well, I think we're running out of time. So Tim, thanks for joining us today. Glad to see you doing well, and look forward to staying in touch.

You, too, Eric. Thank you.

All right. Take care.

Bye.