American Tower Corporation ($AMT)

Good morning, everyone. My name is Benjamin Goy. I'm an equity analyst here at Deutsche Bank, and I'm very pleased to be joined this morning by Rod Smith, American Tower's CFO. Thanks for being here Rod.

Welcome, Ben. Thanks for having me, and welcome. Good morning, everyone.

You reported 4Q earnings a couple of weeks ago. Looking back to 2025, what were some of the highlights? And what are some of the key areas of focus for American Tower in 2026?

Yes. We had a great 2025. I would say at the outset, we are making a lot of progress on our strategic priorities that we've laid out in the past. So I'll highlight a few of the economic accomplishments for the year. Most notably, on an as adjusted basis, we grew AFFO per share by about 8%. So upper single digits there, very happy with that. We also expanded margins, partly supported by our focus on efficiency and cost controls around the globe, certainly -- we are -- we invested almost a little less than $2 billion up in the $1.8 billion range, and we are putting that capital increasingly into the developed markets. That includes towers in the U.S., towers in Europe and our data center platform in the U.S. Our data center platform continues to perform exceptionally well, growing in the double-digit range. Our tower business globally as well as in the U.S. and Europe on a organic tenant billings growth basis is in the mid-single digits, has been kind of solidly there. So things are performing very well from that perspective. Yes, so we couldn't be more pleased with the way 2025 rolled out and our focus going forward is continue to maximize the organic growth across the assets we have globally. And we focus on that every day of the week, driving efficiency through the business and continuing to expand margins priority for us, certainly. And then being very smart and disciplined when it comes to allocating capital, allocating capital in a risk-adjusted way that gives our investors the highest returns over time. And in the environment that we're in today, we certainly look at that and say putting more capital into the U.S. through towers and data centers is mission-critical, maybe priority #1 from our capital perspective and continues to divest and develop the business we have in Europe.

That's a great overview, and we're definitely going to hit on a few of those topics. But starting in the U.S., the big 3 carriers seem to be reaching their 5G coverage targets after several years of robust activity. What are you seeing the carriers invest in today? And how would you characterize the demand environment in the U.S. as we think about 2026?

Yes, the demand environment is healthy. You saw us come off at '25 with a very robust record-setting level of services that really underpins the amount of activity that we've seen across our towers, certainly. Yes, the carriers have made great progress in deploying mid-band spectrum across their networks and transitioning to 5G. There's a little bit more work to do there. So we're seeing that kind of play out. Growth in mobile data continues to be in the 30%, 35% range. We expect that to continue. That means carriers will be focused on beyond getting their networks 5G equipment deployed they will be focusing on the quality of the coverage and eventually the density of the network. So we do see a path forward where the carriers will continue to roll out 5G. And then they will come back through and increase the quality of the networks by continuing to invest in it. They'll be investing in it through amendment in order to keep up with growing demand across the networks that they have. And then eventually, they'll shift most likely to densifying the network. We're seeing an uptick in co-locations, the number of new installations on new towers that they're not already on. That's a form of kind of expanding coverage but also increasing the density. We think that, that effort has to continue over the longer term as more and more of their network relies on higher-band spectrum. We do -- certainly, the C-band spectrum is a higher band spectrum than the traditional spectrum for years back. That means it doesn't propagate quite as far. It can handle more capacity, but it doesn't propagate as far. So they need to fill in the network in order to make it denser. The industry will need more spectrum. And as that spectrum comes, it will be even higher band spectrum over time that the wireless carriers rely on, which means there's more infrastructure needed to make the networks denser in order to operate properly on that higher band spectrum.

So one of the things you mentioned was that really the driver of leasing activity over the long run is mobile data traffic growth. In your view, what are some of the key drivers of data traffic and by extension tower leasing in the U.S?

Yes. It really is just more and more handsets being out in the marketplace, higher level handsets, 5G capable handsets are really getting to a point where lots of people have them. And we've seen that transition over a couple of years, right? When you have a new 5G capable phone, not everyone has it, it takes a while for people to get it. When they get it the adoption of applications that are more data intent that require more capacity, use up more capacity. Those applications grow on the handsets, and people use them more and more. As more and more people have the handsets, more and more people develop applications that will run on the phones and other wireless devices that will drive traffic across the board. That's the cycle that we are certainly seeing today. So 5G is out there well and good, handsets are ubiquitous across the market applications are going up. That's driving the growth in mobile data consumption, which kind of is the backdrop for the continued capital investments that our customers make in the networks. Going forward, we expect that you will see AI-type applications eventually hitting the wireless networks, right? These large language models that people are building, everyone's developing applications that will be AI generated. The idea of having AI applications require uplink and downlink speeds and capabilities that are similar is a different type of a structure for the wireless network. So we think investments will be required in order to keep up with that. And then 6 Gs are around the corner, and then that will come. That will be another wave of network components that will be added to the networks. And it all is activity that underpins kind of that growth in mobile data consumption, which we expect to just continue in that 30%, 35% range for a little bit here.

What about fixed wireless, where your customers are increasingly leaning into that as a driver of growth? Obviously, those users consume a lot more data than mobile users. So how are you thinking about that as a driver?

Yes. It's a driver of mobile data consumption. Even though it's fixed wireless today is running on the mobile network, so it's coming, let's say, coming off of landline networks and now being provided through wireless networks. That means it's part of the growth in the data consumption going across mobile networks, it's using up capacity on the wireless networks. From everything I hear, the carriers like where it's headed. They like their success, their positioning on it. They like the economics of it. So if that continues to grow, it just takes up more capacity on the wireless network, which means more capacity needs to be built into the wireless networks. And we're here to help customers do that. So I think it's a very exciting part of the wireless network. It's one of the few instances where you really see an incremental revenue opportunity, bringing revenue that was satisfied in some other way, a service that was satisfied in a totally different way on a different network now transitioning into the wireless network. So I think it's an interesting development and I think it's good for the wireless carriers, and it could be good for the infrastructure providers, too.

You mentioned spectrum before. One of the priorities in this administration has been trying to find new bands of spectrum to put into use for wireless. How do you view the spectrum pipeline in the U.S.? And what could that mean for leasing activity in the future?

Yes. I think -- I mean it is a priority for the administration to have new spectrum release, the big beautiful bill or the one big beautiful bill has some requirements in there for the FCC to release more spectrum. So we do expect some more spectrum to come in. It's clear the industry needs more spectrum to satisfy that growth in mobile data consumption. But maybe I'll take a minute and just explain the 2 different paths. If you -- if the carriers bring in more spectrum, they're able to put more spectrum into the network and then they can handle more capacity within the network. In order for that spectrum to work, they need to pair it with antennas and lines and radio. So that's all equipment that they would put on the tower sites, and we would benefit from that. If the spectrum doesn't come, which we think some will and certainly over the next 3 to 5 years and then 5 to 10 years, there should be more spectrum being allocated, higher-band spectrum, I would remind you. If the spectrum doesn't come and the growth in mobile data kind of gets to the point where the networks run into capacity issues in the carriers, we'll deploy more equipment and reuse the spectrum they have more frequently, which is kind of accelerates that density. And again, it's then putting more equipment, towers, cables, antennas on the towers in order to reuse that spectrum. So either path, new spectrum reusing existing spectrum requires more equipment on the towers.

Filing earlier this year, you noted that DISH had stopped meeting its obligations under your agreement and you're now pursuing legal action. Can you remind us what you expect the impact from that could be? And how should investors think about the potential next steps in that process?

Yes. What I -- that pretty much summarizes kind of where we're at. I won't say too much more than that, other than highlighting the fact that we took DISH out of our outlook completely for 2026. So our 2026 outlook is completely derisked from a DISH perspective. Any future collections or settlement with DISH could be a tailwind to the P&L outlook and/or the balance sheet if there's some sort of a settlement. But within our outlook, there is zero revenue from DISH, zero economic benefit from any kind of a settlement. So there's only upside. And with that said, I would say it is a litigation. Some of that will be made public and everyone here that's interested can certainly follow that but we won't be commenting along the way in terms of what stage we're in and what's happening in the litigation.

Okay. Makes sense. When we put all of these factors together, how are you thinking about organic growth in the U.S. tower business this year and longer term?

Yes. So this year, the DISH -- removing DISH and having all the DISH discern the '26 affects the outlook, I would say, when you ignore the DISH or before the DISH churn, we're up in the mid-single digits in the 4.5% range. One of the key elements of that 4.5% organic tenant billings growth in the U.S. excluding DISH is that we're seeing about 250 basis points of growth there from new leasing activity, new co-locations and amendments. That is very similar to almost the same number we saw in the prior year if you exclude DISH altogether, right? No revenue contribution and no churn. And so we are seeing kind of a consistent 2.5% growth on our key net new business, colocations and amendments, not including any churn, just colocations and amendments in there. So that's good. And that -- and with all the things that I've said earlier, and that feels like a pretty good range for us to be in, given that 35% mobile data consumption and the growth that's coming -- we haven't yet seen AI applications run through wireless networks. That could happen in the future, and that could be an inflection point one way or another. I wouldn't talk too much long term, but I would say where we sit that 2.5% looking forward, we don't see any reason why that level of activity shouldn't be somewhat consistent.

I wanted to pivot to the international business. You've been seeing strong growth in your European business recently. What are some of the main factors driving that performance?

Yes. We -- I mean, we have a great business in Europe underpinned by our acquisition of the Telefonica assets and there's really just a couple of factors. One is what we're seeing organic tenant billings growth over the last several years that's exceeded the U.S. growth. Now it's coming more back in line, but I think it's still slightly above the U.S. growth, which is good. The underpinnings there is we're seeing healthy new business activity across the carriers that we have. That's amendments, it's colocation on towers. It's new rooftop development. It's some existing customers just increasing their network components. And it's also somewhat supported by Drillisch 101, which is a new carrier in Germany kind of building out a greenfield network, we're getting contributions from them as well. So the new business activity is solid. And the way that our contracts work is our escalation is basically local CPI, uncapped with a 0% floor. So that's a good place to be in an international contract. We've seen higher levels of inflation in the last few years. That's moderated down a little below where the U.S. is. That's what's making its way into our organic tenant billings number that is still slightly higher than the U.S. But we have that organic tenant billings growth number, that CPI escalation piece of it tied to local inflation. Again, it's uncapped and it has a zero pool for the most part. In France, there's a little different since we have a 2% escalator in France. It's a smaller business compared to Germany and Spain. And then we're in a good position with churn in Europe where the vast majority of the revenues come from Telefonica on long-term contracts as part of the leaseback transaction we did when we bought the asset. So churn is very much controllable. It's running in a little higher than 1% now or so. But certainly staying within that 1% to 2% even at the lower end of that, it's clear that we're in a good position in Europe. So to summarize, it's solid new business activity kind of across the market. An escalator that's tied to CPI that's uncapped, which is very good and kind of a modest lower level of churn expected based on the contracts that we have.

It seems like you're doing more new builds in Europe than in the U.S. What are the factors that make these markets attractive to build in?

Yes. So again, I would just highlight the fact that the organic revenue growth has been higher than the U.S. for a couple of years. The markets certainly are high-quality markets, some of the highest economic quality that you would find we like that. We like the developed markets as we've talked about for many years. We do think that's the best place to drive consistent long-term economic growth and returns. So that is all good. We have a skill set that transfers to Europe pretty well in developing towers and rooftops. The carriers that are in those markets need new assets. And we -- I would say we performed better than other infrastructure providers. That's been our experience and therefore, they come to us and we're able to drive terms and conditions that we are happy with. And therefore, we put capital to work and we build up more components and very high-quality markets for the right customers. So we're doing that. We'll continue to do that. Some of that is with Telefonica. Some of those build-to-suits came with the leaseback arrangement we had. I think we signed up 3,000 new tower bills over a 10-year period as part of the acquisition contract. We're building sites for Orange, and we're also building sites for Drillisch 101.

In Africa and in APAC, you're guiding to an acceleration in new leasing activity this year. On the other hand, your Latin America business is experiencing some headwinds related to carrier consolidation. How should investors view the fundamental outlook across these different regions?

Yes. I mean there are different regions. And within the regions, the countries are all different. So there is some uniquenesses there that really should be explored if you want to get into a lot of detail there. I would say the differences between Africa and Latin America. I'll start with Africa, that the new business growth there has been solid, upper single digits, kind of consistently over time. So we're seeing 6%, 7%, 8%, 9% growth from new co-locations and amendments. We expect that activity to continue. So in local currency, that business performs really well. Because the demand for new infrastructure and the carriers' willingness to pay kind of matches up in it and it works pretty good. There is also an element of escalations that are tied to CPI. The inflation was much higher a few years ago, double digits, even 20%, 30%, that's come way down. So now that the inflation is generally in the upper single digit or even mid-single-digit kind of range. And then we're seeing 4% or 5% churn across the market there. And then, of course, FX is a concern when you're in Africa and in Nigeria and Ghana and other places, Uganda where we are at. But if you look at on a local currency basis, those businesses perform very well. Our contracts are solid. Our teams perform exceptionally well. The infrastructure is important. And it's really the monetary policy, the FX issues that create challenges in those markets. And for that reason, we have pulled back our investments. That does not mean that the invested capital we have in the towers, we have can't perform exceptionally well in some years when FX behaves itself. In some years, there may be some FX headwinds. We are looking to minimize the impact that those FX headwinds can have on our overall business. That's why you've heard us talking about shrinking the exposure to Africa relative to our other places we can do that by growing and developed markets or we can shrink Africa from time to time here and there when it makes sense for us and for the investors. So that is Africa. It could do very well over the long term. The economic risk around FX and some other things have caused us to not want to put more development CapEx there, but we will ride that bet that we have and I think it will be constructive. Many years. In some years, we'll have some FX headwinds, certainly. Latin America is a little bit different. It's a more mature market, some of the backdrop in terms of the number of carriers and the competition is different. We've seen more carriers kind of in the market there, and now we're working through consolidation. And that's kind of a major event in Brazil and in a few other places. So we're seeing significant consolidation churn. And in some cases, as the market works through the consolidation churn, they've also slowed down on their new business activity, their amendments. They're really focused on absorbing the customers that are being -- the carriers that are being consolidated and integrating things in. And so the new business has slowed. The churn has gone up, and we generally have inflation-based escalators in there as well. You put all that together, and we've been projecting low single-digit OTBG for a number of years. We've kind of been in that cycle. We delayed some churn from '25 to '26. We've also accelerated some expected churn from '27 into '26. That puts us in a negative position relative to organic tenant billings growth. But that will pull us out of this quicker, too. So we'll accelerate the recovery and we say accelerating growth. We expect that growth to go from negative to begin to accelerate to a more normal organic tenant billings growth across the region over the next couple of years. And we are really very excited about the backdrop in Brazil. They've worked their way through to 3 primary carriers pretty good distribution of market share, pretty healthy across the board. It's a big market, and it's one that we think we could do very well in over time.

I wanted to ask next about your data center business. CoreSite, as you mentioned earlier, has posted consistent double-digit growth the past few years, and you're now guiding to low teens revenue growth again in 2026. What's driving that performance? And are you starting to see a greater contribution from AI related to that?

Yes. So our data center business, CoreSite has been and continues to perform exceptionally well. That great performance has been accelerating, going up from upper single digits to higher to getting it to double digit to now staying in the double-digit economic growth. That's what we were projecting. And and that's where it is. I would remind people that, that's well above the range we underwrote when we did the acquisition. Just as a reminder, we were in the 6% to 8% range when we underwrote it. We've been above that almost the whole time, and now we're in the double-digit growth. The -- I mean, what's driving the double-digit growth really is the new business, the demand side of the equation. Not just for our assets, but kind of across the board. It is a time when more and more of the world's greatest companies in networks, they all want to get into cloud on ramps, they all want to interconnect each other. They all want space where they can build out their business, to grow their own business. They require that interconnection and the ecosystem that CoreSite offers to grow their own business. So we've seen the demand strengthened. We've had 3 or 4 years of record sales repeatedly that not only does that drive in our case, double-digit revenue growth that can translate down into overall economic growth, but it also predicts it out over a couple of years because you're delivering capacity that you've already sold. So we feel really good about the next few years in terms of where that business is going to go and keeping revenue up. We're investing more capital because we need to do that in order to satisfy the growing demand. In the increment -- the returns on the incremental capital has been really good. We're looking at mid-single, mid teens to better than mid-teens on a stabilized basis for the incremental capital that we're putting in place. A couple of years ago, we would talk about how before we even started building a building, it would be 55%, 60%, 65% preleased. That's come back just a little bit because we're accelerating some builds. And much of the building that we're doing is connected the campuses we already own. It's not out speculative building. It is just expanding the campuses because our existing customers within those campuses want more space. They want more compute power. So the business is performing exceptionally well. We are seeing AI use cases show up in our pipeline. We're leasing space to folks that will use it for AI inferencing. But in 2025, it was very early, not a meaningful impact at all. In terms of the P&L, we think that will grow a little bit in '26. But over the next couple of years, we do see AI inferencing tipping into these interconnection data centers with cloud on ramps, our centers are -- have the ability to be very dense from a GPU basis. So you can get a lot of compute power in the small spaces. We think that AI inferencing is going to go up. It could be a wave of demand on top of already very strong demand.

One of the major themes recently in the data center space has been the imbalance between demand and supply of capacity in part because it's so hard to get access to power. Are you seeing that play out? And if so, what does that mean for pricing power across your business?

Yes. I think -- I mean, a lot of it comes back down to the chipsets that are out, faster, more powerful chipsets, companies can do more with that and to help grow their own business. If they're going to do that, they need a place to do that. They need power to do that. They have a place where that power and the heat generated can also be cooled. And that's where you get into these great ecosystems that we have with liquid cooling in most of our facilities across the U.S. We can put the GPU density is going up in our facilities, and we are able to do that for our customers, and they can compute more and do more. So we -- again, we've had great performance in CoreSite. It is even before AI really kind hits in there. That is causing a lot of demand for our site specifically, and I would highlight the fact that we believe our sites are differentiated from others. The fact that our sites have multiple cloud on-ramps in each of our facilities, each of our campuses, that is key. We've got over 400 networks terminating within our campuses across the U.S. Our customers rely on interconnection. They interconnect with each other. We're seeing high single-digit, even transitioning to double-digit growth in the interconnection world. So companies that are interconnected with one another. That works so well for them in terms of growing their own business. They want more of it. And the interconnection grows. That is kind of the demand profile. And that's why you've seen us increase our CapEx to increase capacity to stay ahead of that. We are building in order to provide 2 years' worth of absorption. Typically, we've pre-leased a lot of that 2 years, so it's a low-risk capital, and that's why we're able to drive these high mid-teens or better returns on a stabilized basis.

You recently unveiled a new cost-saving initiative with the goal to generate 200 to 300 basis points of margin expansion over the next 5 years. Tell us a bit more about this initiative what areas of the cost structure are you planning to focus on? And why does this make sense to roll out now?

Yes. I would say, first off, the idea of being efficient in managing costs, we've been focused on that for quite a while. Of course, I think anyone that knows us know we grew materially over a number of years through M&A. That M&A activity post CoreSite has slowed down. Our last 3 acquisitions was the CoreSite acquisition in the U.S. We bought Telefonica towers in Europe, and then we bought another good-sized tower company in the U.S. So our last 3 acquisitions were all centered around the U.S. and Europe in high-quality economies and assets. Since then, the M&A activity has slowed a little bit. We've been focused on integrating everything fully and making our operations very efficient. You've seen us reduce SG&A year-over-year even in light of and working against 7%, 8%, 9% inflation. That's not an easy thing to do. And we did that even though our SG&A as a percent of revenue was already kind of industry-leading and our margins have been industry-leading as well. Now we're transitioning with the appointment of Bud Knoll as the Chief Operating Officer globally to bring some global standards lessons learned and efficiency that comes out of systems and approaches around the globe to really make sure that we're very efficient. He will be focused on things like managing our land expense and land renewal. So a lot of things there that the U.S. has learned over a long period of time that can be more fully integrated outside of the U.S. Unifying supply chains around the globe and really driving contracts where we maximize the benefit that we can see in our procurement areas across IT and systems and really trying to use the best practices from a systems perspective and get them in the right places and make the service better for the customers and reduce the cost. So we think we've got a nice outlook there. And the way we describe it is that those efforts over the next 3 years, 4 years will contribute to margin expansion along with our general business and growth in revenue. So we're projecting that our margins will expand about 300 basis points out to 2030.

It sounds like you're starting to do some efficiency work using AI that's separate from these initiatives. Can you discuss some of the things you're looking at with that?

Yes. AI is rapidly developing. We're on it, looking at it. We've been for a little while for a year or so. We certainly think there will be applications that will make our business more efficient. And as and when we completely analyze the opportunities and build things in, we'll let people know. But I would say that even above the 300 basis point expected margin expansion. AI could have another wave of kind of dramatic results there. In areas like lease processing, lease abstraction and some financial reporting and some of the accounting areas, predicting where sites may be needed and other things, doing analysis on towers through drone footage, having AI review that to make it much quicker and instantaneous in terms of figuring out what's on the towers. Is it actually the equipment that's listed out in the agreements. Is there any equipment differences. So it's an exciting time. I think it could be important to the way our business operates, and it could be in addition to the 300 basis point expansion that we expect.

You've been delevering your balance sheet over the past few years, and you're now back within your target range of 3 to 5x. How is American Tower thinking about capital allocation in general? And what does it mean for the company to now be within your leverage target?

Yes. So we are back within our target. We're below 5x, which is a nice place for us to be. We've enjoyed 2 credit rating expansions or upgrades. We're at BBB+ now across 2 of the 3 firms, which, again, is a place that we like to be because we're focused on quality of earnings. We're focused on quality of balance sheet. The way we allocate capital really has not changed in terms of the way we approach it, maybe our focus areas certainly are different. So that's absolutely the case. But first and foremost, we provide a dividend over $3 billion this year, about $3 billion last year, over $3 billion this year, grew 5% last year, probably in the same similar way this year. That dividend and dividend growth is paramount to us. It will continue. We will protect it and really be focused on it. After that, we go back and we invest capital. We think that's the highest return possible. We're investing more of it, 80% of the growth capital now is going into the developed markets, which again is the highest quality markets in the world with the highest quality customers in the world, we think that is a really good thing. So we typically invest anywhere between $1.5 billion and $2 billion a year in capital programs. Then beyond that, we look for M&A opportunities where we can leverage scale, increase assets in places we already are, get an outsized benefit that someone else might be able to value in. And we'll continue to be very selective. I just highlighting again, our last few acquisitions were in tower -- a big tower portfolio in Europe, a medium -- a good-sized tower portfolio in the U.S. and then the CoreSite business in the U.S. all of which are working out very, very well for us. And so we'll continue on that disciplined, developed markets kind of approach, really trying to figure out the infrastructure that the networks of the future are going to need in the highest quality markets. That's really -- that's where we're focused. And then beyond that, we will and always do put up M&A against share buybacks. We'll do the math and make the right decision there when we can [indiscernible] invest over $365 million at the end of last year in a relatively short time buying back shares. We announced we continue to do that as we turn the corner into 2026. So we bought back more shares in 2026. Not making any prediction going forward, but that is in our toolbox. We'll be looking at that and putting it up against any M&A. We'll be balancing that with the benefits of just paying down debt and preserving capital for future deployment. So our approach there hasn't changed materially at all.

One of the questions investors have been asking recently is whether satellite broadband could end up competing with terrestrial wireless infrastructure or it will mainly be a complement to towers. I know you have a relationship with AST SpaceMobile, who's building out satellite broadband. So I'm curious to hear your perspective on that.

Yes. We view it as a complementary. An important complementary technology to terrestrial networks. It is built and designed for certain purposes like extending coverage into rural areas, hard-to-reach areas. Those hard-to-reach areas can be in places like the U.S. and also over in Africa, LatAm throughout Asia. So it really is a very important service, but it's complementary to terrestrial land-based networks. It also provides immediate coverage support and disaster recovery and extreme weather events and those sorts of things. So that's clear. There are fundamental challenges with trying to have a satellite wireless service compete with a terrestrial network. There are capacity issues in terms of the spectrum that is available for satellites and how much capacity you can actually run through that, and it is nowhere near sufficient to compete with terrestrial networks in places where people live and work and travel. And I think that's kind of broadly and widely known. So we do view it as an important complement to the terrestrial networks, not a competitor.

Maybe just to wrap up, it's still fairly early, but the industry is starting to discuss 6G. Do you have any initial thoughts on what spectrum could you use for 6G? And when you'd expect that activity to start picking up more meaningfully?

Yes. People are talking about 6G. We're just getting 5G deployed in the discussion already leapfrog to the next technology. New technology is always good for tower companies and infrastructure players. Those new technologies allow new applications, which often are higher bandwidth. And in an AI setting, again, it's going to be an uplink and downlink. It's going to fundamentally change the way the networks need to operate in it's going to require new infrastructure. I would expect that there will be additional spectrum coming along for 6G and won't be for a few years yet. So it's too early to talk too much about that. But it will be higher band spectrum, maybe significantly higher band spectrum, which means the capacity will be there, but the propagation won't be there. And that, again, is another kind of fundamental driver for network densification.

Well, that seems like a pretty good place to wrap it up. Thanks, Rod.

Excellent. Thanks, Ben. Thanks, everyone, for joining.