American Tower Corporation (AMT) Earnings Call Transcript & Summary

Yes. So it's interesting you say that. We kicked off this year, and we're doing some deep work on scenario planning. So what that means is we're working with some consultants and we're working with our internal teams. We've interviewed our internal management. We've also been talking to external experts, thought leaders. What are the key issues facing the world, right? It could be macroeconomic things like interest rates. It could be political issues or conflicts, sustainability, climate change. So all of those are fair game. So we do a pretty good job, I think, of thinking through those issues and say, what are the what ifs from a risk perspective and disruption? And then more importantly, we get into the technical road maps, thinking about drones, like just crazy ideas like what if we put high-powered DC up the tower, and we're supporting drone recharging so they can move from tower to tower at a point where they can deliver a package in a beyond visual line of sight drone network, which is coming, right? Package delivery and those types of opportunities, how do you enable those things? We create creative ideas of opportunities in those scenarios, but we also would look at what are the risks? What does that mean if there's -- we're looking at sub terahertz communications. Is that a backhaul opportunity on power? Or is that something that's going to be a short range, and we need to look at more densification? So yes, scenario planning and how technologies roll out across pretty much the entire set of digital infrastructure is a fair game for that, including how customers and, frankly, the ecosystem is evolving and who's driving what part of the solution. And that is an important element of, I think, any business and we think we're leading and thinking through those opportunities for future growth, but also what are the disruptive elements that we have to worry about -- and that's what we spend a lot of time doing.

And so on this topic, how does -- and you mentioned the technology earlier, virtualization, O-RAN, how does that affect demand for macro leasing? And are the impacts different like in the U.S. market, where you have regulators that have a preference for facilities-based providers versus maybe other parts of the world that have been more flexible around MVNOs and active network sharing?

Yes. So just on the O-RAN question, there's really no correlation at all, right? O-RAN is itself opening up the RAN network. We've tried to do that over the -- I mean, you think about early days of cellular, and I've been in this business for decades, right? So we try to open up just the switch at the old days, the circuit switch and the radio network. And that took a while, like it took a whole generation before we got those interfaces right. And then we've got to 4G that became natural, and we started to look at the radio, the part that would be what they call the radio unit and then there's the baseband unit and trying to be able to open that up with things like let's call CPRI or eCPRI. These techniques still didn't become pervasive because they were challenging to get the management planning and some of the other components to be supported in a plug-and-play way. But certain operators were able to open that up. So when O-RAN comes, it really doesn't affect the tower business at all. It really says, you still need the radio. The physics says you still have to put all your spectrum in some form of transmitting capacity at the site if you want to deliver on the regulated requirement. But then where you process it could change, right, depending on the fiber density, and how you go to cloud native networks. And we think there's a lot of promise in O-RAN. Ultimately, people are doing it today when they think about it is a virtualized RAN. So a lot of the main OEMs have already and the operators have been deploying as the core is already virtualized. The RAN is starting to -- the challenges at the DU, which is the part where the physical layer needs to have a lot more horsepower, so can't use general-purpose compute. But overall, that's not changing. Now the question is, from an overall regulation standpoint, is there anything that's encouraging or discouraging on how MVNOs and others come in? And I think it comes down to spectrum again, right? So when we have places like in Germany, we have location spectrum, in the U.S., we've had CBRS spectrum, there are opportunities for folks to take a spectrum that was generally unlicensed and use it in a way that can deliver solutions in certain hotspots, right? There's also [indiscernible] licenses, which gives you some restriction. So there's opportunities there, but to be an MVNO over the top and to actually own your own licensed spectrum, I think those really depend on the cost of spectrum and the way underlying, let's call it, telecom providers, let's say, 1&1 in Germany is an example where there were an MVNO. They were satisfied at some level of MVNO. But at some point, they decided strategically we want to own spectrum. The cable MSOs in the U.S. I've looked at that as well when they looked at bidding in the power auction and they bought some portions of the CBRS band. And they felt like, "Hey, I can build out parts of that network and support my needs from a, like, say, payment and economic needs, but still become an opportunity for us to support them on our towers and/or other assets that we have." So I think the movement to more spectrum available to more players in limited ways or even nationwide ways as we've seen with new entrants like Rakuten and, obviously, what DISH is doing, that we would encourage that, obviously. We think it's important to have a vibrant and robust competitive market. But overall, MVNOs do drive some innovative ideas, including things like converging wireline and wireless in different ways that are unique to their business, building hotspot overlays in certain areas where they think the demographics support their customer base. And we would be very supportive and encouraging of that. So those are more, I think, market-by-market opportunistic. It's not really being dictated by a regulator. Now when you mentioned RAN sharing, fringes of the network are very uneconomic, right? So operators have -- particularly in Europe, have partnered with regulator observations to say to limit that exposure to super rural areas where they can actually share a common radio and still maybe do things like MORAN where you put your spectrum and you have 2 radios, but you share a common baseband, those things are happening, but in a very limited way. So we watch that and we understand that. We know what the underlying principles are. In fact, when we look at in-building, we think of in-building as a shared RAN as a service. This opportunity for we call it hybrid networks. So it's not private 5G. It's one thing a lot of people talk about, but that tech building one private network for a specific tenant or user, right? That's not our business. We want to be able to have that be hybrid, where the private use cases get built and supported by the owner of that property, but we want to open up that same network to the mobile operators so that they can actually be more efficient in using that rather than try to go from the outside in or build their own solution to get to the customers that would transit that facility. So we see a lot of opportunity there as well, but it's going to be market by market.

And as you look at maybe the global landscape for a moment, it feels like one of the differences between the U.S. market and maybe some of the emerging markets was the relative densification of cell sites relative to the population and the geographies. Is that still the bull case for international markets that densification of cells or macro sites is going to significantly pick up over time? And as you have a whole collection of markets, do you see a tipping point or inflection point in some of these where you might get there more quickly versus others that are taking more time?

Yes. I think each market builds out its network at a pace that's commensurate with the underlying economics. So you think about ARPU, you think about things like what is important to that user base? How much existing infrastructure might already be there that's an alternative or a substitute? And wireless, in some markets, is the only game in town. So they have to deliver that, but then the underlying economics become challenged. And obviously, we do what we can to support that by building unique tower sites. And a lot of our program, too, you guys have heard this from Tom and the team is, we spent a lot of money every year in debt CapEx that supports 5,000, 6,000, 7,000 sites with build-to-suit. So we're investing in those communities to try to help build that level of connectivity. So that densification of sites when we think about whether it's sort of toll population. But when you think about number of users, active users supported on a particular 3 sector side, it's still been almost a rule of thumb for decades where it's like 2,000 users or something like that, that you roughly will get. So when you think about markets where it's oversubscribed, like you have 3,000 or 4,000 or 5,000 users that's saying I got a sell split unless the user experience is going to be grade, or I need more spectrum. A lot of this in the U.S., too, is driven by the fact that maybe the operators don't have as much spectrum as some other markets might have. So when you think about the amount of megahertz they have per POP, that drive -- and then the relative market share will drive how they invest capital. And obviously, we want to be supportive of making it very efficient for them to do that in a way that's consistent with our business model of a neutral host. So I think you see that whole progression. And then in the cities, you have rooftops. In a lot of places in Europe, rooftops are preferred over the small cell because it's providing, let's say, a 500-meter or even 200-meter level, but that's already institutionalized in the thinking and going street level maybe isn't as effective, especially if you're trying to manage your bandwidth and your frequencies from an overall performance standpoint. So I think that there's -- clearly, we watch all that. We think about that. We try to figure out when the densifications are coming, what operates -- they come to us and ask us for help in certain markets. We look at that on a back to the capital allocation of the business model. We'll take a look at does this make sense? Can we multi-tenant it? Can we take -- what we do for one operator, can we add a second operator to that asset, including the build-to-suits that we do? That's part of our longer-term planning. So we see all that and that's -- actually a global perspective is really great from my vantage point and from the company's vantage point because we see these waves of approaches. And not every market is like the U.S. and some are very different, and the economics are very different in terms of ARPU and what you can afford. But people want the latest and greatest technology. They want to have the latest and greatest devices, but it comes back to economics and affordability.

Maybe switching to the Edge and to data centers for a few minutes, and we're also receiving some questions from our audience on this topic. So I'll try to fit all those into this discussion. So can you just describe the progress operationally for the data center strategy? It's been almost a year since the purchase of CoreSite. And if you can unpack a little bit more of the opportunities for American Tower to leverage these assets to monetize, whether it's the metro Edge or the mobile Edge that the company has talked about in the past?

Yes. So that's a great question. I think CoreSite stands on its own merits. And I think we've been consistent with that. It is a very valuable asset from the standpoint of it's a highly interconnected business in terms of the underlying data centers. It's not just a data center that's colocation-only facility. It's really bringing together the clouds and the cloud on-ramp, is bringing together enterprise and it's bringing together a ton of fiber providers who want to be able to peer and offload their traffic to other fiber providers for purposes of their end customers for enterprises or potentially other networks that support consumer broadband. So it's a fundamentally unique asset in that sense, and it has great growth capabilities. And I think we've talked about 6% to 8%. We've exceeded those this year so far in terms of what we've talked about publicly through second quarter. And I think we see what [indiscernible] and the team over at CoreSite, under Steve's leadership, have been able to do. We think that the ability to have American Tower, and obviously, we have Stonepeak as a partner in that now, that combination has given them and unleashed them a bit in terms of their ability to sell through and bring additional business and growth to that asset. So that stands on its own as a wireline Edge. I think of that as a wireline Edge. That's really where the nexus of traffic and a service chain comes together. You want to have a hybrid cloud, well, you're going to put compute resources in CoreSite. And as an enterprise, you want to cut down the latency of your users being able to do transactions between what's important to you from a data sovereignty and securities perspective, and what you're willing to put into public cloud? And those on-ramps become critical. And what we've done is we've extended that with OCX, right? This is the fabric that they've had between their data centers. And we had the data site, which is our facilities that we had in Orlando and Atlanta and our original Colo Atl, all that's been integrated. So when we think about it, those now are fundamentally part of the family. And we announced recently some of the extension of the campus in Miami. We're also investing in our Silicon Valley 9. So the normal capital and cash flow of that business are driving accelerated growth. Now that's sort of on its own. It's valuable, and there's a lot of demand for that business and highly interconnected nature in the cloud or on-ramps become critical because what we want to do is extend those as we move to the next layer of the Edge. And when you think about that, moving from 12 or 14 major peering locations to something that can move to second-tier markets or even just think of NFL cities to start with or MLB cities, and then moving into those areas of population where there needs to be a faster response. There needs to be a way to have compute and bringing the clouds who really want to densify their on-ramps further, we support that. Now through OCX, could be done easily or it could be done if they want to build their own facilities. So we're seeing that aggregation layer be built. The access Edge at the tower sites, we have 6 of those, and we're actually experimenting with them. We're thinking -- we're looking at building sort of an incubation capability so we can bring users and partners together to sort of see what their relative performance is from a latency perspective. So we've been thinking through that plan. We've been getting a lot input from internal and from external parties on how to put the pieces together. But think about it, the real important piece here is what happens in the future as you move the wireline and wireless on-ramps together, we have convergence, right? We have to be able to look at those players that want to be horizontal across those 2 domains. So they're dealing with either consumer or enterprise use cases, and they need to be able to have consistent experience. Today, it's all a patchwork of all these different networks finding their way to CoreSite or maybe other CoreSite like companies, and also mobile operators, who are all individuals. So think about a place where we can do things like we're working on today, which is called [indiscernible] Imagine a future intersection where you have -- operator A has a pedestrian cross industry, operator B has the 5G camera supporting classifying what's happening at the intersection, and operator C has the telematics modem in the car passing through. How do you connect the dots on all 3 of those. You've got to go through 3 disparate access networks in mobile across the Internet, decide what's going to happen and send a lot information back to the pedestrian to alert them that something may be happening at the intersection. We think there's an opportunity to build that layer of Edge. And it starts with the cloud on-ramps being part of that, the mobile operators being able to bring the traffic together from their tower sites, and we think that, that combination of bringing mobile operators and clouds together will drive that next layer of what we call [indiscernible] needed for future workloads as they move with mobility, just like handoff in the old days when you had to move capacity from one cell site to the next as use his hand off. Think about in the future, we need to move those workloads from one data center to the next, and you have to automate that entire fabric to do that. And I think that's what we're thinking through. And we think there's a great opportunity to bridge the distributed real estate we have at the tower sites with those campuses that CoreSite has been developing on their own and growing and bridging those 2 worlds together between the mobile operator at the tower and the cloud on-ramps at CoreSite.

To get the mobile operators in, does that mean you're essentially going to create C-RAN hubs for the mobile carriers, and you want all of them to participate in these around the U.S. Is that one of the ways to do this? Or are there other ways to try to get these different firms together into one hub?

Yes, Michael. So there's several different workloads. So when you talk about C-RAN hubs, you're talking about telco workloads, which are kind of taking cloud native, let's say, network functions and deploying those. And a lot of operators have done that already, right? So they've done that uniquely. What we're talking about here is that next layer up. So we're -- in 5G, they have this thing called a User Plane Function, it's a data path. You could exit that data path way north bound in the network, and it has a certain amount of delay and jitter, or you can start to have it more deterministically closer to clusters of towers or nearby even feeding multiple C-RANs into that location. Nobody's done that yet, right? And then one of the questions that's out there is, "Okay, why do I need to do that? When do I need to do that? What's the use case that's driving that?" So we're planning that architecture. Again, we won't own the whole thing. We don't have that aspiration. We think there's partnerships to be had. But we think we have a leadership role with all the right ingredients to be a catalyst to make that happen and working with key partners, not just domestically, but looking at how does that roll out around the world. And I think that's where we are with that funding. So we've been connected to the data centers, as I said earlier. We put those on the fabric. We have the classic business operating at really good capabilities today. And we're still looking at how do we extend that to what might be the future, which is still several years away. I mean, we're not saying this is around the corner, but planning and understanding that with the partners that want to go first is where we're spending our time.

And have there been any benefits beyond the investment that Stonepeak has brought to the table for American Tower and the data center strategy?

Yes. It's still early, but I think, Stonepeak itself is fairly well understanding of the market. They have their own existing investments. I think they're a great partner. And we'll see how that plays out as we move forward together. And we're looking forward to that, and we're excited about having them as part of the now joint venture with respect to those CoreSite assets.

So as investors try to follow the progress and evolution of what, I think, some are calling this Edge monetization opportunity, what are the things, what are the milestones, the data points that investors should be looking for, what they should be mindful of as they track this evolution?

So a couple of things. One is, I think, on the wireline side, there's going to be more and more AI and ML processing need to be pushed to yet. So it's not -- if it's not on a corporate campus today, and it needs to be near a facility that supports that community that's doing that, we're seeing that edge move out closer to those demands. On the wireless side, the question would be when a devices start to become more real time and immersive, in other words, can I have an AR experience [indiscernible] physical and -- my physical world with digital and be in a position where if that has a lot of delay or jitter, it's not going to be that attractive. So that's going to push what's today, the edge in mobile is probably about I don't know, 10 to 15 or so sites that are all still on the other side of core network locations. How does that bridge and move further to that aggregation layer, we're watching that. That comes with device, watch the device road map, and watch the software development as 5G gets built out, at least for coverage, and we see those applications. There's also a CDN model, right? So you look at the content delivery networks. Those are the folks pushing content, whether it's streaming or other types of interacting, gaming, those things are going to also require those like locations closer to where their devices are being supported. It could be immersive concerts. It could be things like, again, experiences that we've seen over the pandemic, where people are starting to immerse themselves and blending multiple worlds, whether it's the concert on robots or Epic games and those types of companies, they are going to be key drivers as well as users take on expectations of a different type of, let's call it, web 3D, if you want to call it, that type of experience. So those are all leading indicators. And then, obviously, there are companies out there making their bets on metaverse, right? And there's a lot of people who want to see that happen sooner. And there's a lot of people making silicon and obviously, devices. Those are also leading indications. And we're talking to all of those folks trying to understand what can we do to help facilitate their plans as an infrastructure company. And how do we take our tower assets and our CoreSite assets and invest further in that next layer, right? So that's that layer of densification that we think is needed. We still need to do the plumbing right, right? So the plumbing gets into the peering and the fiber network, that's all part of that. And that's a fair game for discussion with various partners.

We talked about a wide range of subjects and different initiatives for American Tower. Are there any others, or any technologies or products or any other areas that investors should be thinking about for American Tower?

I would just say the one thing that's come up recently is satellite, right? We've seen the SOS texting. There's some other worries in narrowband services, the LEOs and the GEOs are looking at ways to offload -- GEOs are actually looking to offload some of their density to tower sites terrestrially. So there are discussions on that. Obviously, satellite as a backhaul, helps with densification and sites in rural areas. So that's an element, and we think it's very complementary, very supportive of all the folks that are looking at that area, supporting the white space, connecting the unconnected, solving some of the digital divide aspects as well. Those are all things we didn't cover that yet, but I'd say that, that's also part of our portfolio of things that we think about.

And as you look at the different satellite technologies that are out there and where it may be going, do you see a risk of cannibalization for macro sites? And how do you evaluate the scale economics of a macro site maybe versus some of these other solutions as you think about that future of the business model?

I don't -- I mean there might be some things around the edges, but I wouldn't say that there's any dominant belief that terrestrial networks and tower sites will ever be subsumed by satellite. Satellites are really their own unique platform. They provide broad area of coverage. They're not capacity drivers and terrestrially radio spectrum that's available and fiber will always be very much at the heart of how users will demand services, especially when you move from rural towards the cities. The main thing is trying to create that broad area coverage so that you can actually identify demand, right? Think about a rural application around a farm or agricultural, there's a lot of IoT devices and other types of things that today may not be supported because it's too expensive maybe to build that. So maybe you create that opportunity through satellite, and then you backfill in terrestrially where you support the offload of that so the satellite can continue to develop new use cases and applications essentially in different parts of the geography. So I see them as complementary. I think that one will inform the other on where demand might be. One may be more economic than another. Obviously, satellites I see as economic for coverage. And for broader coverage then I see towers is economic for more localized capacity. You are on mute, Mike.

Bound to happen, at least once during our session. As you look at the rural use cases and maybe just to finish this topic out, one of the interesting changes that's to come is the amount of funds that the government has greenlit for potential rural fiber infrastructure builds and to bring broadband to these rural communities. And as you look at fiber getting deeper into America potentially with this money, how do you look at that from a build-to-suit and tower opportunity? Is that a net potential positive, neutral? Is it a net negative to the degree that it could cannibalize maybe what towers could have been used for in some of these areas? How should investors look at more fiber in rural communities, and what it means for tower colocation and wireless service?

I see that as a positive overall because what happens is, is driving communications for that to meet that demand. But the last let's say, mile, if you want to use that term, the last hop, it's still more efficient to do it wirelessly. So at the end of the day, the cost -- obviously, I understand the ambition and the support from a subsidy and a finance standpoint. You try to bring everybody on to fiber, but it's kind of a little bit shortsighted in thinking that, that's going to happen magically very quickly, and it's going to be inexpensive. It's going to be very expensive to connect every household, right, with fiber. So -- and the timeframe will take quite a long period. And there are more efficient ways to do that. So we think as fiber gets deeper, it creates opportunities for backhaul so that we can put power sites up and then reach those folks more efficiently. And again, we think some of the improvements in wireless technology will help the data rates for what most use cases and demand will be. So we see them as, I think, positive because it's more of that fiber -- you don't have to put a microwave dish that saves some additional costs potentially. And it may not be supportive of the types of throughput. So when we think about 5G throughputs, and you think about system throughput, having fiber does help offset any limitations you might have on the old forms of backhaul or potentially microwave. So there's definitely opportunity there. And we see a lot of [indiscernible] that have been looking at leveraging different underlying technologies. There's -- they don't even have to be 5G. They can be legacy. There's some point to multipoint solutions that are out there. They keep getting better with MIMO and some of the other technologies in wireless. And then we also see the opportunity for those folks to deliver more value in CBRS when we start to combine the 80 megahertz of GAA or the unlicensed part with the 70 megahertz of licensed PAL, those things can be put together. It's a lot of spectrum. And then if you do it right through the SaaS and organize it, we think that there is opportunity there. But fiber could help connect more of those towers in a more efficient way and deliver a better user experience over the wireless channel, which I think will still be more cost effective and faster.

Great. Well, Ed, thank you so much for sharing all of your insights with us today. It's great to see you, and it's just a great opportunity for us to learn all about what American Tower is doing across the range of subjects that we covered. So thank you so much for your time.

Thank you, Michael. It's pleasure to be here. Thanks, everyone.

Thank you.