All right, let's get started. Thank you very much. Good afternoon. My name is Greg Williams. I cover cable, wireless, and comm infra here at TD Cowen. I am joined in this session by Rich Rossi, the Executive Vice President and President of the U.S. Tower Division of American Tower. Rich, thanks for joining us.
Yeah, thanks for having me.
Before we start talking about the U.S. carrier activity, et cetera, the elephant in the room from earlier last week was about SpaceX. They said more specifically, we can sort of figure out a network topology using femtocells, putting them around homes and their dishes, or putting them even in Tesla cars. I just wanted to maybe vet that a little bit with your opinion on what that would mean. Would that disenfranchise any towers? Just your general thoughts on that. I know there is a lot more questions than there are answers to that.
Yeah.
Your thoughts would be appreciated.
Sure. I think that when you look at build-outs in the U.S., macro sites have always been the backbone, and we think that they're going to remain the backbone. While there may be other types of infrastructure that are used to help get networks into targeted locations, urban areas, or really fixed sites, macro sites will always be that backbone. There may be ambitions or plans to do something that pivots in a different direction, but at the end of the day, it comes down to deployment time, cost, efficiency, and that's when you're going to come back to companies like ours. We think ultimately if there is a large network deployment in the U.S., that we're going to play a big part of that and—
Sure.
—ready to do it.
Absolutely. Just transitioning over to the U.S. carriers, how would you describe the current activity environment across the big three? As part of that, are you seeing any signs of acceleration or deceleration in your application activity?
Yeah. We're still right in the thick of things with 5G, and I would describe it as consistent. If you were to look at 2025 versus 2026, our new business growth was about 2.5%, that component of our algorithm. We're seeing that again in 2026. While you do have a couple of carriers who are working their way towards the end of their mid-band deployment.
We are seeing densification pick up. It takes a different mix of activity, but you're largely seeing the same output as it pertains to our growth. I think it's consistent, and it's where we thought we were going to be at this phase of 5G. Moving out of that initial paint the map coverage phase and more into how you solve for capacity. Obviously the capacity demand continues to increase dramatically year-over-year on mobile networks. Depending on what you look at, there are predictions it'll double by the end of the decade, maybe sooner. We're seeing the carriers aggressively respond to that type of demand.
Got it. Some of the carriers are moving from coverage to capacity. I imagine then the way it looks from your revenue perspective, it shifts from amendments to colocation then. Is that on a carrier-to-carrier basis or market- by- market basis? Is that broad? How would you describe the mix shift?
It's a little bit carrier to carrier because they're all in different phases of their journey in terms of how they're going to get 5G rolled out. What you see is the amendments still drive a lot of the activity just based on sheer volume of amendments compared to how many new locations you see. You start to see an increasing share of new co-locations, and that really is the best indicator for densification given that a lot of those co-locations occur in markets where you have mature networks that have been rolled out for years.
You know it's a capacity play versus just a straight coverage need.
Got it. So maybe you are seeing a little more colocation in the NFL cities first.
Yeah, you are going to see capacity addressed from the urban areas moving out. So you may see different types of infrastructure, rooftops, building facades, small cells in the urban areas. Then you get to our sweet spot, which would be suburban and moving out from there. That is where the macro towers really play the big role.
Got it. Maybe talk about organic growth. I think you grew 1% in the second quarter year-over-year, but 5% if you exclude the Dish churn. New business, I think it was contributing 250 basis points as I recall.
Yep.
To what extent do these colocation trends hold? I guess the question there is how long is the legs for the densification or the capacity upgrades?
Yep.
How long does that last in terms of organic growth?
I'd say overall we have our organic growth algorithm that we look at, which is you've got a 3% escalator, you've got your new business that's somewhere between 2% and 3%, and you've got your churn that pulls that back by about 1%- 2%. We think that that's going to hold true over the long term. We think that it is a durable growth—
Sure.
—vehicle.
So 3 plus 3 minus 1 is 5?
Well, it has proven out that way so far over 5G if you take out t he one-time events consolidation, things like that. Obviously there's a range on that, right? So it can go down, can go up a little bit there. I think densification is you look at the new spectrum that's coming to market over the next decade, a lot of that's going to be at higher bands. By its own nature, as you deploy higher bands of spectrum, you need more dense networks for that to mesh correctly. Then as you have that in parallel, you have that increasing mobile data consumption model, things like AI coming along. You're going to see more and more data consumed, more of a capacity crunch, and then architecting networks to address the way that the frequencies are going to propagate from those higher bands.
Sure, yeah. Higher bands, more towers. That's it all.
Yeah. To answer your question, we see there to be a really long sustained period for densification from 5G into 6G when you're going to see that new spectrum released.
Got it. I want to talk about service revenue guidance. You held it at $245 million for the full year, but you noted a step down in the back half. A lot of times we look at service revenue as leading indicators for leasing. Is there a read-through here that is still valid? Or is there a mix of the service work that is changing?
Well, so I think the tough part is 2025 was our best services year ever.
Okay. Tough comp.
Yeah, exactly. Tough competition. 2026 is going to be a great year. We think it is going to probably be our third largest, so a really healthy business. But compared to last year, it just may not match the peak of what we saw. We have worked hard over the last five years or so to really diversify that U.S. services business, and our bread and butter was always the AZP, the acquisition, zoning, and permitting, for our customers. Now we have introduced more construction work, which higher revenue, a bit lower margin. So it is a mix in profile. But we have also developed a product that we refer to as end-to-end, which picks up the customer's needs at the scoping, so around the time they are applying for sites and goes all the way through construction.
That program management has been an offering that is very well received, especially given that some of the carriers have scaled back their resources in-market, working on deployments. We are able to help fill that gap. Nobody is going to perform better on our sites than we are in terms of getting people out there quickly.
Sure.
We are feeling good about services, happy with where our attachment rates are with the major carriers, and still very healthy.
Great. I want to talk about MLAs.
Yeah.
As we start to look towards 2027, are there any comprehensive MLAs scheduled to roll off through the balance of the year through 2027 that you can help us with? How does American Tower typically approach the conversation around the next holistic agreements?
Yeah. Always talking to our customers, right? When the ink is dried on an arrangement, you are talking about what is the next opportunity, how can you upsize what was agreed to, how can you figure out how to keep continuity when something comes up for expiration. So those discussions are always ongoing. It is not necessarily a sign of what is to expire, but more so what can we help to do to extend out having a good, healthy relationship with that customer. So always ongoing there. We are somewhat agnostic as to whether we do a holistic agreement or whether we do something that is more pay by the drink.
At the end of the day, we think the bottom-line result in terms of the growth numbers would be somewhat the same. It may translate into the financials over a different period of time.
That is one of the nice things about the holistic agreement on the tower owner side, is the predictability of it that you can forecast out very clearly. But ultimately, those arrangements really are about the value creation for the customer and how much efficiency you can give them when it comes time for them to deploy. So you have to have the demand for the customer who needs to have those economies of scale and the speed to market on deployment. And then for us, it just has to be a match on what we think the addressable market is—
Yeah.
—with that particular build or, over longer term, multiple builds, and—
Yeah.
—you kind of see where it goes from there.
Yeah. A lot of times, I think of MLAs, it's like when an upcycle's ready to happen, so it'd be a good time—
Absolutely.
—for the carrier to sign an MLA. What would be those upcycles? Is it safe to say that maybe it's Auction 115? There's some FAA restrictions there, so it's going to be a little bit pushed out. Would that be an accelerant or a catalyst for an MLA? Can you talk to the 2.7 GHz spectrum? Are those the factors you think that'll create or drive maybe an upswing for in turn an MLA?
There are different factors, right? There are going to be the frequency auctions, and when the carriers deploy capital to acquire the spectrum rights, they're looking to do it and aggressively roll it out. Nobody wants to spend at those levels and then sit on their powder, so to speak. So I think that that can be a catalyst if they're going to go out and touch a number of sites. Also, the natural evolution into 6G is going to be another thing that may help drive some of that. When you have multiple of those factors happening, that tends to be when there's a lot of demand forecasted.
Yeah, those absolutely are catalysts, right? When someone shows up at a car dealership, they're usually there to buy a car, right? I think when someone comes in and says, "Hey, I've got a lot of activity that I'm looking at," it means it's probably a good opportunity to maybe talk about getting something done.
Right. Wanted to talk about churn a little bit.
Yeah.
Because you guys noted in the second quarter you absorbed what's characterized, I think you said, as the final wave of churn from a former anchor tenant, and none was expected next year. I don't expect you to give any guidance or anything in the ongoing process, but having that headwind behind you, does that change the way you think about your growth algorithm?
For us, it really just clarifies for folks who are following along with our results and trying to predict our growth. We thought it was important to de-risk the business and just make it really easy for people to follow along with what's happening. Canceling the Dish business and taking that out of our numbers was the way we could be the most transparent to make it simple. I gave you the walk on the algorithm, and we think that outside of those one-time events, that is going to be our long-term outlook on things. For us, it just makes it easy for people looking at 2026. We've talked about numbers with Dish, without Dish. It helps for people to understand the magnitude of what we're taking out, but also be able to track what those numbers look like compared to, say, 2025 or prior years.
Got it. What typically happens when a tenant's winding down their network in terms of, well, I guess it's not a typical situation, but when a typical situation, how long do they leave the equipment up there? It will stop there, and then I have a follow-up.
Sure. There's usually a contractual protocol for it. You work that out as you're leasing somebody's space, what happens to the equipment at the end. Generally, your customer will come in and take the equipment down as a surrender type arrangement there. As you've seen consolidation of networks over the last 20 years, there have been some commercial arrangements where the acquiring party says, "Hey, I don't want to take down this old network. Why don't we work something out where you keep it, and you progressively take it down over time?
Yeah.
Sometimes it's removed, sometimes there's a negotiation for it to not be removed, and then there are other cases where maybe a bankruptcy court's helping to decide what happens with the equipment that's up there.
Right. I guess you harken back to Sprint, too. There was some decommissioning as well.
Yep.
Just thinking out loud, do you believe that if there's infrastructure up there, could that be a shortcut for a new entrant to enter the market? If you buy what's up there already, you're already on second base, if you will.
Yep. There are some constraints in the sense of you need to have antennas, radios that align with whatever the frequencies are that the newer entrant would be using. The gear may or may not work in that sense. But when you look at long lead time items for network deployment, you have local permitting, you have power procurement. There are some parts of the country where the power companies are not able to quickly get meters put out there or get you connected into an existing meter. You may have issues with the ground lessor, where you have to get approvals to get another customer in there or acquire some additional ground space.
But a change of control, the ground lessor will have to approve of that?
Sometimes they have consent just to add a new customer.
Yeah.
You go to add someone to your tower, and the ground lease that was negotiated 25 years ago says, "Hey, anytime you go to add a customer, I need to know, and I need to sign off on it. Some of those things can take a long time. If someone were to come in and look to take equipment that's already up on the site, you've got a poured concrete pad, you have cabinets, you have tower mounts. There is infrastructure that can be reused irrespective of the spectrum. Then those other items, things like building permits, you might be able to transfer the permit, or it may be quicker for the local jurisdiction to approve the permit because they've already approved a like-type construction project for that prior customer who's exiting. There are definitely some synergies that exist.
We've seen it with Sprint and others, where people went in and acquired shelters or used mounts up on their RAD centers.
Got it. Has the way you looked at counterparty risk changed the way you structure contracts? Any lessons learned there with recent events?
Yeah. We've always been extremely disciplined when it comes to how we underwrite contracts. What I will say is the scenario we're talking about isn't about an inability to pay, right? It's an unprecedented situation.
Willingness, not ability.
Yeah, exactly right. As an industry, we've had to work hard. Patrick's sitting out there, right? There's been a lot of work done to try to rectify that. Even though we have three customers that drive a large part of our business, we have several thousand other customers out there that we do business with every month. There is a discipline when it comes to how we underwrite, how we collect, and we have all the landlord-tenant interactions you'd expect with a portfolio of our size. Our ability to keep our churn at a very low rate, I think, speaks to the discipline we've had—
Sure
—around that.
Sure. I want to bring it back to SpaceX. I know we started with the whole femtocell thoughts because I just had to get it out there. SpaceX did acquire 65 MHz of nationwide spectrum from EchoStar, flexible use. I appreciate it's early days and a lot of more questions than answers by far.
Yeah.
Do you see SpaceX ultimately coming in, maybe as a fourth carrier? How do you see this all play out?
Yeah, we'd have to defer to them in terms of what the size of their ambitions are. I think from where we sit, we look at it and say that anyone who is going to be a scale entrant in the U.S. is going to look to portfolios like ours.
Like I said with the MLA question, we talk to customers all the time, and sometimes that's prospective customers, too. We think we're well plugged in to where the potential opportunities are, and we feel strongly that our assets would be very key to somebody rolling out a network of any real size—
Right.
—in the U.S. So, as you look at the mobile data consumption trends and the fact you're going to need denser networks, you're going to need more portfolios like ours where you can get that shared cost model and find ways to leverage existing infrastructure. Yeah, another entrant would be exciting, but we know it's a competitive landscape out there already.
And it seems like in satellite, rural takes care of itself, but you're in a lot of rooftops, too. So that—
We do.
—would obviously help for them augment their network should they need to.
Yep.
Yep. How about direct-to-cell? Is there any change in the network architecture over time? Because today it just seems like you are using the cell, and it finds the satellite and vice versa. Is there anything in terms of changing the network topology that you can think of in the direct-to-cell play?
I think you still have to see what happens, particularly with spectrum, like how that is going to change with the amount of spectrum that is going to come to market over the next decade. See how that goes. Obviously with the satellite networks today, you have the issue of building penetration. You have some latency issues. You have some limitations on handsets in terms of what can be done there. I think a lot of details still have to unfold before we can know exactly what changes would be required. Still a lot to be seen there.
Maybe talk about the spectrum opportunities, past, present, and future. We had Auction 113. Verizon took a lion's share. Is there an opportunity there, or is that a lot of just augmenting the ecosystem that already exists versus what is coming up is the Upper C-band and Auction 115, and then there is chatter out there at 2.7 GHz? Maybe we will just stop right there and talk about—
Yeah.
—the spectrum opportunities.
Yeah. Look, we're always excited about spectrum coming to market, and a lot of credit to the FCC and NTIA for creating that roadmap to identify spectrum, to give the FCC the auction authority back so they can start to plan and think long term. When it comes to spectrum being deployed, it generally involves more gear being put up on towers, more trucks being rolled, which is good for the vendor community, good for our services business. There's just a lot happening there. So investment in capital typically converts to investment on infrastructure like ours.
So contract to contract, spectrum band to spectrum band, there may be different outcomes in terms of what the result is, but when you take the amount of spectrum that's going to be deployed and combine that with the data demand and what AI may bring coming down the pipe as well, then we think you're going to see a lot of deployment—
Yeah.
—across the board.
To help with that deployment with the Upper C-band auction slated for maybe hopefully next spring or summer, knock on wood. Can you help us with that opportunity? What is the typical lag between an auction to actually putting it up on the towers? In this case, it might be a little more delayed, right? Because there is some, I guess, FAA and other issues. Maybe talk to that timeline of the Upper C-band opportunity. Is it an opportunity, because if they have C-band equipment, is that additional equipment they would need on the towers for Upper C-band?
To start with your question on the timing, there is always a lengthy process overall from beginning to end, from identification of spectrum, auctioning, clearing, and then deploying. So, it is natural that there is a lag between when the spectrum is auctioned versus when you actually see gear hung on towers, and that could be a year, two years, three years, depending on the spectrum. In some cases, you have seen the carriers find creative ways to accelerate it, as they did with the 600 MHz on the broadcast repack.
Oh, right. Yeah.
Right. They got together with the broadcasters and figured out a quicker path. With Lower C-band, obviously, there was a lot of coordination with the FAA and different aviation stakeholders.
Right. The A category and then the B, C category. Yeah.
Right. There was a lot going on there. I know that there's been a lot of collaboration on Upper C-band over the last couple of years, learning from the Lower C-band experience. Yeah, I think the conversations are ongoing. It seems like they're trying to find a way to do this collaboratively. If it auctions in 2027, you're still probably looking, we've heard from the carriers, maybe it's 2029 or early 2030—
Sure.
—when you'd start to see that stuff show up on towers. But we may see some activity a little earlier than that, but for the most part—
Right, because you are saying the carriers will collaborate with the owners of that spectrum to try to accelerate it, obviously.
Correct, and with the FAA and others in the aviation space to figure out what is left of concerns around the altimeters and things like that.
Right. Because I recall in the C-band, one of the big resolutions was we will just keep the towers off near the runways, I guess.
Right. If you recall, that was some of those sites had been constructed and ready to go, and then that slowed down till they could figure out what was a viable solution for everyone.
Right.
I think that they've been leveraging a lot of those lessons over the last couple of years as if they've figured out how much of that Upper C-band can actually be brought to market. We anticipate there will be a good resolution there.
Okay. I wanted to switch gears to M&A. As you consider the U.S. market, how would you characterize the tower M&A environment? Is there a case for further consolidation among players at this point in the cycle?
Yeah. We are always on the lookout for what's out there for M&A. We do anything from single towers, rolling up a mom-and-pop tower here and there, to looking at the bigger portfolios that are out there. I think when you talk about things that are more sizable, you need a couple of stars to align. You need a willing counterparty, somebody who wants to come to the dance with you. You need economics that makes sense, and you need a regulatory framework where you can get the deal done. Thus far, we just haven't seen those three things align on something that's very large. But we're always on the lookout. We want to invest in our developed markets, whether it's towers or data centers in the U.S. or towers in Europe. We're active lookers in this space.
Can you talk about the private market tower valuations? There's been this stubborn private to public—
Sure.
—multiple gap. How has that moved over the last year from your perspective? How would you describe that bid-ask spread between buyers and sellers today?
Yeah, there's still a spread there, right? It's still a measurable difference in terms of the multiples on private versus what us publics are trading at. Just one person's opinion, but I think we've seen a little bit of a rollback, just slightly, on some of the private deals that are changing hands. There's also not a ton of data points out there.
Is that a tightening of the spread?
A very small tightening of the spread.
Okay.
But again, not a lot of data points out there. The sample size is small and not big portfolios. So, that can probably fluctuate. I think on the private side, you have investors who sometimes are looking over a longer time horizon. Obviously, with the publics, people are looking quarter to quarter, year to year. I think some of the private investors, they like those secular tailwinds that they see with spectrum intensification and AI and things like that. So they're willing to put down a bigger price on some of these things because they're looking at more of a long-term—
Okay.
—opportunity.
Got it. Wanted to switch gears to fixed wireless. In the cable and broadband space, it's fixed wireless and fiber to the home are all the rage and taking up a lot of, not just the net adds in the broadband space, but the spectrum usage. Can you help describe what you're seeing? Are you seeing fixed wireless specific demand on your towers at all?
I think it generally blends more in with the overall demand that we see.
Yeah.
Is it possible that you see an application that's a capacity colocation that also represents an extension of fixed wireless? Sure.
Could that also explain some of the densification?
It could.
Yeah.
Any use that's driving more consumption of the network is a potential driver of densification. It could be one of multiple factors that could help drive those type of decisions.
It's hard for you to answer this on behalf of your customers. Do you think they have enough spectrum then with their fixed wireless ambitions?
I'd leave that for the carriers to answer. I think the carriers have worked really hard to try to help clear the way for more spectrum to be identified and figuring out the best way it can be put to market. I think the carriers have been open about their desire to get spectrum, and it's a scarce resource. That could just be one of many contributing factors.
Got it. Want to talk about spectral efficiencies and AI-RAN. Nokia's CEO on its recent earnings call said the AI-RAN, that platform could quote unquote, "Deliver more than 100% spectral efficiency gains by 2028." I always thought of spectral efficiencies as a mid-teens percentage efficiency. When you hear 100% over the next couple of years, it's a pretty big deal on their existing spectrum.
Yep.
Is that a software item that can double that network capacity or they need hardware, which is obviously beneficial for you? Otherwise, this is a headwind on these level of spectrum efficiencies that they're talking about.
Well, we don't see it as a headwind. I think that software optimization, software-defined networks, that's been around for a long time. We have carriers today and over the past decade who have done upgrades to our sites where they've just gone on and done a software push to make a modification to the site. That's not going to be a new thing. We know that where the demand is going on mobile data, it's going to take more than just spectrum. It's going to take densification. It's going to be the combination of those things that do it. If there's some level of efficiency that they can create to help create more space for data to be consumed, then that's a good thing.
We don't look at that as being something that's going to be detrimental to our opportunity because we know that you are going to need more sites, you're going to need more equipment, and then also, your AI-RAN also works for distributed real estate portfolios like ours.
Yeah.
Creates opportunities for things like the edge, helps to feed an AI use case ecosystem where you are going to need the networks to have lower latency. We think a lot of those things actually create opportunity for us.
Yeah, part of that huge number or efficiency number, they would say, I imagine there's beamforming technologies. The radios themselves would maybe even move around with motors—
Yeah.
—and things like that's equipment. Equipment for you would be a good tailwind.
That's right.
Yeah, it's a good way of—
Yep.
—thinking about it. Okay. My last sort of questions or set of questions is on the edge. Mike and I were at Connect (X) back in May—
Yeah.
—the edge sounded like 5G 2018 all over again, right? Meaning there was so much talk about it again. Maybe it was lackluster between here and now. Of course, we had COVID, et cetera, but here we are again talking about it, and should we come in with some skepticism like we did last time? Or is AI really finally that use case that will proliferate the edge? Second question would be, could some of that edge reside even at the extension of the base of the tower? I know you've got the CoreSite assets, of course, but just like to hear your views on where you see the edge playing a role in your business.
Yeah. Look, we're bullish on edge, and one of the reasons why we're so excited about acquiring CoreSite in 2021 was the opportunity to marry up the large regional data center model that CoreSite had with the distributed real estate portfolio that we have in tower. We still believe that that is on its way, and it has taken more time. To your point, it's taken longer than we thought it would. You haven't seen the use cases emerge yet where that close proximity, low latency has been really key. You still hear the autonomous vehicles, wearables.
Right.
The same kind of litany of things that we are talking about.
I was going to put you on the spot. What are the use cases you think—
Yeah.
—that might actually drive it?
It sounds a lot like 5G, to your point, and then you add in AI, which is obviously a big disruptor, right?
Smarter autonomous vehicles. Smarter in terms.
Sure.
Yeah.
Sure. I think with things like wearables, people wearing Meta Glasses, stuff like that, it is putting more pressure on the uplink on these networks than you traditionally have seen in the past. I think there is going to need to be some reconfiguration of networks, downlink versus uplink, and it is not because downlink is going to be reduced, it is because uplink is going to have to be more symmetrical. I think you need more overall capacity—
Yeah.
—a little more balanced. On the edge, being able to distribute closer to where we think the action is going to happen, 40,000 U.S. sites puts us in a good position to have a lot of great locations. We have CoreSite, we have their Open Cloud Exchange, so we can have the interconnection between our edge facilities and the big regional facilities that CoreSite has. Distributing power, right? Being able to procure power in some of these areas is very challenging, so being able to take down smaller blocks of power and put them—
Sure.
—in a distributed ecosystem versus having to aggregate and potentially wait years to get the power is a positive thing, too. The really great thing for us is we look at CoreSite continues to have a ton of activity, AI and otherwise, and we think the stuff that you're going to see at the edge that has to be processed more locally is all additive to that. It's not a shift of taking from here to there. It's the how do you handle the more traffic coming through, and we think you're going to handle that stuff locally, to some extent, with these smaller facilities.
Yeah, a lot more distributed.
Yep.
Yep. The data gravity sort of goes towards the outside as well.
Yep.
The PowerPoint is actually interesting, right? If you've got—
Yeah.
—a megawatt here or there, and we are seeing some of that with central offices being redesigned for these little mini couple megawatt spaces in the closer to the edge. It is interesting.
Yeah. That is the type of stuff that we are following on with. We feel really good that that is going to emerge and emerge soon. It is just as you said, it just may not be tomorrow.
Right. Great. With that, we are all out of time, so thank you very much.
All right. Awesome.
It's great. Yeah.
Thanks so much.
All right.
Appreciate it.