Adding data center business as well with CoreSite. What's the impact you want to make on the company during your tenure, and what's the most important change in the company you want to catalyze?
Thanks. Well, as you know, I've been a part of the leadership team for a while now, so a lot of the strategy that we've been working on for the past several years, I've been part of. And I think our next decade is really about leveraging this global platform we've built, to have a differentiated value proposition for all of our stakeholders, our customers, our investors, the communities we serve. And what that looks like for us, in the near term, our priorities in the near term, is first, trying to maximize organic growth. You know, we've spent about $25 billion in the past four years, predominantly in the U.S., with CoreSite, InSite, and a couple of small deals in Telxius in Europe. And so maximizing the growth in those assets and the whole portfolio.
You couple that growth organically with some selective investments that we're gonna make, you know, again, predominantly in developed markets, some in the emerging markets, where we partner with the Tier 1 and Tier 2 carriers, but at a lower rate than we have in the past, and then you take the growth from the organic side and from the inorganic side, and you complement that with cost controls and continuing to focus on expanding margins across the enterprise, so that's kind of some of our key focuses there. The fourth thing is really working on our balance sheet, making sure we have a fortress balance sheet, so that if there are other opportunities going forward, either inorganically or share buybacks or whatever we think is gonna create the most value, we can take advantage of those.
I think number five, we've talked a little bit about, is we're working on some board refreshment and governance, but that's a little bit of the kind of behind-the-scenes stuff. Really, the things that are gonna drive value for the next decade for our shareholders is to maximize organic growth, minimize expenses, a little bit of selective investment in the right markets, M&A, if it comes available, and then focusing on our balance sheet.
Great. And then maybe on the last conference call, I think you pretty explicitly said that, you wanted to really increase your level of focus on developed markets. And so if you look over the next five to 10 years, are there specific regions or specific subsegments that you wanna sort of target a little bit more aggressively, where you want to expand your presence effectively?
Sure. And let me just clarify, this isn't a new thing. We were just a little bit more explicit on our call about what we've been doing. If you look at our investments over the past four years, again, we've invested about $25 billion in developed markets. We haven't done an M&A transaction in the emerging markets in that period of time. And if you look at our internal CapEx program, in 2021, the emerging markets made up about 70% of our internal CapEx program. At the midpoint of our guidance this year, it's about 40%. So you've seen us continually emphasizing developed markets more, and on the last call, we just wanted to be a little bit more explicit about what we're doing and why on that.
And so if you look at the entire portfolio across the globe, our first priority is always to add more scale in the markets we're in. That's where we get the best operating leverage. That's where we can continue to drive more value as to those assets being part of the American Tower umbrella. Having said that, are there select other markets that we would consider? You know, absolutely. And we do have a couple of developed markets where we're smaller scale, both France and Canada, and if there are opportunities there, those could prove interesting for us. And if there are other opportunities in the developed world that have the right, you know, market dynamics, and by that, I mean, is it a country that has a healthy ecosystem of carriers, where we can see multiple tenants?
Are the terms and conditions we can get from an anchor tenant, the right terms and conditions? Are the valuations right? We would absolutely look at that. But for us, we think that we've built an amazing portfolio across the globe. We're at scale in all the major markets where we need to be of scale. So there's nothing that's forcing us into a new market. We can just purely be opportunistic at whatever's gonna derive the best long-term shareholder value.
Mm-hmm. I think once you sell your India asset, emerging markets will be about 25% of your overall base or so.
On an AFFO per share basis, yes.
Yeah, and so if you think about kind of some of the smaller markets where you operate, can you maybe help investors understand what your desire is to continue operating in some of those areas when maybe you have a little bit less scale than the other ones that are priorities you mentioned?
Sure. So as we've looked at our portfolio, we've achieved meaningful scale everywhere we think we need to. On the countries where we are subscale, we've worked really hard to put in place an operating model where we can support those from regional hubs. So for those markets, it's just purely opportunistic for us. If someone came along at the right price and made an offer, we could consider that. But in the meantime, we've turned all of those markets cash flow positive. We're able to repatriate those cash flows and invest that cash in our priorities. So there's nothing I would point to you today to say that there's a particular need to divest anything. Having said that, we will be opportunistic.
There was a press release out in the past couple of days where we sold a small land portfolio we had in Australia and New Zealand, and again, that was just being opportunistic. It was worth more in the hands of somebody else. We don't have a presence there. We're not operating towers there, and we just took advantage of the chance to recycle some capital there. So we're open to doing that. The board and the management team is always looking at all of our portfolio and saying: What optimizes that long-term shareholder return? And, you know, does it make sense to recycle capital or harvest the cash flows? And we'll continue to make that evaluation over time.
Mm-hmm. Are the countries and regions where you don't have a big presence today, where you want to be more meaningful in terms of scale? I think Europe kind of comes to mind as one example of that, where your presence is limited to just a few countries. You know, do you see Europe specifically as an attractive market for you relative to others?
Yeah. So Europe, for us, is really predominantly Germany, Spain. We have a small presence in France. We're very excited about the market and what we're seeing there, especially in Germany. We're seeing a lot of demand. We're seeing a lot of new leasing in Germany. Spain's performing very well for us, and we are a little bit smaller in France, but we see, you know, kind of great dynamics in that market as well. There are other countries in Europe that could be interesting, but if you think about when we entered Europe and why it took us so long to build scale there, it's really about being patient. And what we saw is portfolios trading in Europe that just didn't have terms and conditions that lent themselves to long-term growth.
And so we saw portfolios trading that and I won't tell you what they had, but I'll tell you what we look for. You know, we're looking for portfolios where you have uncapped CPI escalators, 'cause even in Europe, you have FX headwinds. We look for a counterparty agreement where we can monetize amendments with the anchor tenant, 'cause that gives us some guaranteed long-term growth in it. We want a portfolio where we can lease up every site. Some of those portfolios had set-asides where you can't lease some sites. Some of those portfolios might have had terms and conditions where the carrier can buy them back, well, that's not for us. We're a strategic long-term operator.
So when we think about going to a different country in Europe, it's not just valuation, and it's not just market dynamics, it's also what are the terms and conditions of the underlying initial agreements with the anchor tenant? If we can find an opportunity that kind of meets all those dynamics for us, then we're open to it. But if we can't, we're happy to take that incredible portfolio we've already built and continue to monetize that and grow that wherever we can.
Mm-hmm. That may be an interesting segue into the next question, which is just, there's been a rumor of a few tower portfolios for sale in the U.S. Maybe just give us an update on how you're thinking about your overall kind of U.S. portfolio trade-offs of organic versus inorganic investment. Start there.
Look, we love the U.S. market, and, you know, as we do with any M&A transaction that's out there, if we're invited to a process, our teams will evaluate it. I think what we've been very public about is there's nothing on mine or Rod's desk today that would take us away from our priority of delevering this year. So we're saying there's nothing in the M&A pipeline today that we think is attractive enough to take us off of that. Having said that, we'll always evaluate those opportunities. We think the U.S. is a great market and has a lot of continued upside in it. We certainly see the carriers continue to invest in our network, so we think that's gonna continue for a very long time. So if we had the right opportunity to invest in the U.S., we would.
When it comes to kind of internal CapEx program in the U.S., most of our investments in the U.S. are in our indoor DAS business, our land portfolio, and enhancing our backup power program in the U.S. We would love to be doing more build-to-suits in the U.S., but we'll only do them on the right economic conditions, and we've seen other people in the market willing to give terms and conditions that we just don't think are conducive to creating long-term value, and that's why you haven't seen us build much in the U.S. If that dynamic changes, we'd certainly be interested in doing that, but when it comes to M&A, you know, like I said, we look at it all. If we find the right deal, we'll act on it. If we don't, we'll sit it out.
Mm-hmm. And I guess maybe just give us your broad view on the disconnect today between public and private valuations. Are there other factors besides valuation that kind of change your view on expanding or how you expand the US footprint? And maybe T&Cs or T's and C's are part of it, but curious about your view on that.
T's and C's are definitely a part of that, and it's also, you know, we tend to kind of lump towers together, but every tower has its own dynamics. So when my team's looking at whether to buy a portfolio of towers, we actually put a business case together for every single tower. Even the one we did a few years ago that had over 11,000 assets, we have a business case for every 11,000 towers, and so you're looking at the growth rate of the towers and what you think that's gonna be going forward. You're looking at the T's and C's with the anchor tenant, and you're just looking at the market dynamic, you know, where is it? Where is it located, and all those kind of have to come together to make that make sense.
Yeah. On the data center side of things, from our discussions with other providers, it kind of seems like some hyperscalers may not be more willing than they have been in the past to actually do large outsourcing deals with third-party data center providers, as opposed to doing more insourcing, work. Do you envision for yourselves the potential for any scale M&A in this sector? Or are you sort of dedicated to a kind of a core, organic-only relationship with CoreSite?
When you look at the reasons why we bought CoreSite, so first of all, we had a strategic alignment with towers because we needed an interconnection hub for what we think will ultimately be a long-term convergence. Setting that aside, when you look at the business by itself, we were very deliberate to buy an interconnection ecosystem-anchored data center platform. And the reason we did that is it's a different business than what you have with the companies who are leasing out single-tenant facilities to hyperscalers, or even undifferentiated colo facility, where people go there to save money. People don't come to CoreSite to save money, they come to CoreSite to make money by interconnecting to other players there. And so when you think about how that business has been built, there's a kind of competitive mode around it.
The vast majority of revenue in CoreSite is with customers who are interconnected to five or more other customers. Now, they may have dozens or hundreds or thousands of interconnections, but once they get into that ecosystem, they're interconnecting that way, it makes it very sticky, so when you do have ups and downs in the business cycle, we're not subject to some of the things that happen in the marketplace, like churn, that other people are subject to. We're also able to underwrite much higher yields because that ecosystem is more valuable to our customers, so if you look at CoreSite, before we bought them, they had the highest returns on invested capital in the industry at kind of mid-teens.
We're still underwriting all of our new development under the same same standards they did, so we're underwriting mid-teens yields in the U.S. on that portfolio, and you just don't see those same dynamics playing out in that wholesale space. So when you think about those hyperscalers outsourcing their builds, that's just not us. That's, that's really more appropriate for someone else. Now, having said that, what I have said publicly is, we do have a private capital partner, and we would be open to the idea of an xScale-type arrangement like you might have seen other people doing
But the only reason we would want to do that is if it gave us the nucleus of a new campus. Our campuses are what give us the ability in a neutral host fashion to drive outsized returns over a period of time. And so if something came along and our private capital partner wanted to fund it, you know, we might be a minority share owner or something like that. But outside of that, that's just not our business, and we'll let other folks focus on that component of the ecosystem.
Mm-hmm. You know, continuing on the data center theme for a second, you know, we've seen a lot of AI training applications in some of the large data center, scale operations we've talked about, but at some point, seems like we should be seeing more inferencing, more inferencing kind of work. Ostensibly, CoreSite should be a prime beneficiary of that trend, when it happens. So you, when do you see that playing out, in terms of the move to inference, and how do you expect to see your data center business benefit directly to that, from that? And do you actually see any synergies between inferencing on the edge or in data centers in general and your core tower business?
Let me start with the first part of that, and I'll get to the second. So inferencing is already starting to happen, and we do have some inferencing business coming into CoreSite, and as you point out, we're a perfect environment for that. You're creating all this information in the large language model somewhere else, but you have to talk to the end users of that data, and that's where you need a distribution network, so you need an interconnection ecosystem to do that. We're the perfect place for that. Having said that, we're not taking counterparty risk. There's a lot of AI startups out there, and we have the luxury in our business of being able to be selective on that, so we are selectively taking some inferencing business.
The other place that that's actually benefiting CoreSite, the kind of overall AI phenomenon, it comes in a couple different flavors. First, because the entire ecosystem is seeing a supply-demand imbalance, we're seeing higher prices. That benefits us. The other thing that we're seeing is that our core customer, and the number one driver customer that we have that drives our business in CoreSite, are enterprises that are doing hybrid cloud deployments. So that's people who either still have servers back in their office and want to go take advantage of cloud tools and want to connect directly into an on-ramp, or people who are cloud native or all cloud, and it was too expensive, so they're pulling their data out and hosting it.
The thing that we're seeing with AI is we have customers who want to take advantage of the large language models and the tools there, but they want to protect their data. So they want to create their own instance of GPUs to house their data connected to a large language model. And so we're starting to see interest from that core enterprise customer expanding their business with us as well, to incorporate that kind of, you know, small language model as part of a large language model. So we're already seeing some of the benefit to that, and we think that's just going to continue to drive demand and continue to drive record sales we've had for the last two years.
They still won't let me say we're going to have a third year, 'cause we set a pretty high comp last year, but I'm hoping to be able to say that. But demand's very strong in the data center business, and AI is one component of that.
Mm-hmm.
So the second component is, h ow does that translate to towers? And when you think about the original reason we bought CoreSite, it's because we see this convergence over time of wireless and wireline, and we believe that there will be demand for edge compute in a way where the wireless networks are appearing at the edge. And when we started looking at that, you know, we deployed a few containers at tower sites, ran a fiber back to a data center, and we realized very quickly that you have to be able to connect that back to this ecosystem, this interconnection ecosystem. And when we were trying to partner with CoreSite and other people, we realized that whoever controls the interconnection ecosystem controls a lot of the pricing at the edge and a lot of the value exchange at the edge.
That's where we said, "We need to own this to be able to play in this space." Now, that's a longer-term aspiration. We do have some proofs of concepts going on that, and we're working with different partners to evolve that. But when you think about how AI may play into that, I think it's still a question mark. We do think there's gonna be potentially a use case for that, and a lot of that has to do with how we're interacting with it. So today, when you're interacting with an AI, you know, bot over your phone, it's generally text or still photos today. That's not as bandwidth-intensive as video is.
So I think as we see these things evolving, and you're looking at, you know, real-time video editing, or you're thinking about, you know, facial recognition software, where it's gonna, you know, I'm gonna look at you with a pair of glasses, and it tells me who you are and all the stats on you, that requires very low latency, very high bandwidth, and that's kind of a perfect test case for edge compute. So I think that you will see that evolve over time. We're just not there yet.
Mm-hmm. I mean, what can you talk about some of the partners that you've been trialing these concepts with, and maybe how far they are away from being, like, kind of a real material revenue driver for you?
I've got one I can talk about publicly because our partner, IBM, blogged about it, so that's, I'm a little bit freer on that one That's kind of a niche use case where we have an automotive client who, you know, they produce automobiles. As soon as they roll them off the lot, off the factory floor, the software's out of date in them, and the way you update them is through a wireless chip in the car. And they store these things in this huge parking lot that has, you know, thousands of cars in it, and there's not enough bandwidth to update them. So that, we have a proof of concept to have an edge data center.
The carriers are doing local breakout at the edge to create more bandwidth to be able to service all those vehicles. So that is live. We're still testing the market. Does it work? What are the economics, et cetera? But, you know, the early results are promising. It works. If it does, it's a niche use case, but there's probably a few thousand factories that could do that. And again, I can't give you economics or anything on that yet. That's part of the proof of concept, is to figure that out. Another use case that's a little bit further behind is an AI use case, and this is not what we're working on, so I'll give you an example that's similar. But say you own a pharmacy, and you have security cameras on every aisle in your pharmacy already.
There's an AI tool that will do inventory management, so it'll tell you, you have to fill in the Advil on aisle three, look in your back stock, and order it for you. Today, the sales proposition is, I'm gonna mail you a server, your pharmacist is gonna unbox it. He has to put it in a place where the cleaning crew is not gonna unplug it, no one's gonna spill Coke on it, whatever. It's, you know, call it $50,000 for that server for each pharmacy.
It starts getting complicated operationally. So the edge use case is that we can deploy a centrally located, at a tower site, you know, centralized node, tether ten pharmacies together, reduce the the operational complexity and the cost because they can share. Instead of $50,000 per, it's maybe $20,000 per, sharing the infrastructure. Now, that one's further behind, but that's, again, another kind of use case of where taking things to an edge compute versus, you know, either on-premises or centralized may make sense.
Okay, maybe let's just pivot back to the bread-and-butter tower business for a moment, if we could. You know, the U.S. carriers have upgraded, I think, around 50%, maybe a little over 50% of their overall kind of footprint with 5G at this point. Kind of given the size of your domestic business and you know, the carrier CapEx trends we've been seeing, what level of comfort do you think investors should have in your ability to sort of drive you know, sort of mid-single digit or 5% organic growth in that business today?
Sure. So we put out a multi-year guide, and just to reiterate what that was, from 2023 through 2027, we expect at least 5% OTBG on average over that period, 6% if you normalize for spread, and when we put that guide together, we did that based on a few factors. We took a look at what it would take to meet consumer demand over that time period, so if you think about what we know, we know that consumer demand is growing 20%-30% a year in the US. We know what frequencies and what spectrum the carriers have to deploy. We know what equipment is out there, and we can look back at 3G and 4G to kind of see where that demand's gonna occur.
That's what we did to put out our multi-year guide, and that involves some things that are contractually committed and some things that are just based on the densification needed to meet that. So as we look at the carrier deployment of 5G, I think what we've said at the end of Q2 is that we had one carrier that had mid-band 5G deployed on about 80% of their sites. One was a little over half, and one was a little under half. So that implies a long runway of deploying 5G just to get to ubiquitous 5G in the U.S. We feel very confident the carriers are gonna do that. They need to produce gigabytes of data cheaper than they did in 4G, and 5G gives them a dramatically lower cost per gigabyte to produce that.
We believe that all of our carrier customers will approach 100%. They won't quite get there 'cause not every site has fiber, and you need that for 5G. But they should get close to a 100% in terms of 5G overlay. You also have consumer demand that's really starting to go up now. When you first deploy a network, you not everyone has devices that work on it. And so that, when that 5G network first went live, there weren't a lot of users using it. Now, our mid-band 5G handset penetration, as of a few months ago, was over 50%. I haven't seen a recent number, but it's probably higher than that now, and we have a few data points from that. The average 5G user is using two to three times as much data as 4G users did.
I think we're using roughly 19 GB of data today per month. That's projected to go up to 59 GB of data by 2029. So if you think about what the carriers need to do to meet that consumer demand, they've got to deploy those networks, and they have to densify those networks, either through adding more spectrum or new sites over the next several years to meet that demand. And that's what gives us the confidence to say that the trajectory that we saw when we put that guidance out holds true today. The carriers will continue to invest in 5G because it gives them cheaper gigabytes of data, consumer demand is there, and it supports their networks.
Mm-hmm. And I think one of your largest U.S. customers, its MLA, has ended with you or between you. In the past, your strategy has been basically locking these agreements to drive consistent ratable revenue growth over time. Given the need for all these large customers, as you just said, to spend, to continue upgrading out the their networks, should we expect you to sign new MLA with that customer coming up? And are there any reasons why you wouldn't do so?
So when you think about our comprehensive agreements that we enter into, the reason we do those is because there are a lot of operational benefits that both parties get. It's easier to deploy, there's less back and forth between the parties. It makes their on-air time faster, saves them money in deployment costs, and it does give us more predictable revenue over time, and it gives them a predictable cost over time. But at the end of the day, we're not discounting to get those deals done, and the carriers don't want to overpay to get those deals done.
So when we're negotiating one of those agreements, you know, we have a set of activity we think they're gonna do over a period of time, and we know what we're gonna charge for that, and the carriers may disagree about what they think they're gonna do in that period of time. And if there's a mismatch in our expectations, you end up not entering into an agreement like that, and that's okay. We've actually only had comprehensive agreements in place with all three of our large carriers for two years, ever in our history. There's always been somebody aside from those two years who is outside of a comprehensive agreement. And we've got a long track record of proving that we can be successful, either in a comprehensive agreement or on an à la carte basis. So we're very comfortable doing whatever.
It's whatever works for us and the customer, and we'll be there to help support their network needs, and they're gonna do what they need to do on the network, whether we're in a comprehensive or an à la carte.
Mm-hmm. Okay, and I guess, you know, as you well know, the U.S. government doesn't seem to have any spectrum auctions lined up. Doesn't seem like there's anything on the horizon, as far as I know. So basically in conversations with customers, how do you think that impacts their, their overall investment plans? Do you think they sort of just spend ratably, and then change that once we get their hands on more, more spectrum? Or do you think they're actually gonna be more active in cell site splitting or, or densifying the networks, because there's no kind of clear visibility of that coming online?
Sure. I think they're gonna do whatever they need to do to meet the consumer demand that they see on their networks. And so if you think about what they've said they need in terms of spectrum, there's on the CTIA website, the Carrier Association, there are a couple of white papers out there, and what they've said is they need 400 MHz of spectrum by 2027, 1,400 MHz by, I think it's 2032, to meet consumer demand. And so the need is real, and I do hope that we get more spectrum coming to the market. But at the end of the day, the carriers are gonna do whatever it takes to meet that 20%-30% demand by the consumer.
So I think that will come in a, you know, whether there's spectrum or not, will come in a combination of adding additional spectrum to existing sites and densifying the network over time. And if you think about the propagation characteristics of mid-band, it doesn't propagate quite as well as some of the lower bands. Now, there are some technologies that improve that, like beamforming and putting higher power through it, but at the end of the day, it, it'll benefit the networks to densify as well, not just from the, the capacity standpoint, but also a quality standpoint. So I think you're gonna see them do a combination. And again, that's very similar to what we saw in 4G and 3G. You know, each of those generations came with a little bit higher bandwidth of technology.
You saw some densification, and ended up being, you know, good for the overall network quality.
Mm-hmm. And then maybe just on a kind of regional basis, maybe just do a quick round the world. Like last quarter, you sort of slightly ticked up your European and African growth guidance, you slightly ticked down LatAm growth guidance. So maybe just spend thirty seconds unpacking, like, kind of what the underlying dynamics are, what's going better, what's going worse than you thought before?
Sure. So just to be clear, we took our guidance in Europe up to the top end of our range that we gave, and that's really being underpinned by higher growth and new leases and amendments than what we expected. So that's a positive demand trend that we're seeing there, and that's also underpinned by a little bit better performance by us. In Germany, like all the other operators there, we encountered some headwinds from regulatory processes and power outcomes, and I'm happy to say that we're able to parachute in our best folks from around the world and help solve some of that. It's part of the benefit of being part of American Towers. We've got some great talent there, and we've really overcome that.
You've seen us kind of really step up in our delivery of stuff in Germany. That's a story of just increased demand, and we see that continue to accelerate going forward as well, so we're very happy with that. In Africa, a large part of the uptick in guidance is really from the escalator going up. CPI was a little bit higher than we originally had guided there, but we also have some really robust sales going on there. Africa continues to have healthy demand for new leases and amendments there. The story there is very positive in terms of the operating leverage that we're getting there, the demand from the customers. They continue to invest in their networks despite some of the headwinds that are out there.
And really, the challenges in Africa have been more around currency translation with FX. Latin America, we did take our guidance down a little bit, and that's really related to some softness in sales there. And that's really related to the carriers, particularly in Brazil, continuing to digest the Oi assets they bought. We do think that turns around once they get through kind of the network integration and once they have kind of completely solidified their plans on the network. But that is a little bit of softness that we saw there, and that's what caused us to take that down just a tad to the lower end of the guidance there.
Mm-hmm. And just, you know, relative to Brazil specifically, I think América Móvil kind of talked earlier this year about investing $8 billion in Claro Brazil through almost the end of the decade, I think 2029. Can you give us a sense on, you know, how optimistic you are in that market in particular, and kind of how you see, you know, kind of the longer term growth trajectory, as you just said?
You know, longer term, that market has all the demand drivers that should see us return to some healthy growth there. And very similar to what happened in the U.S., when we had carrier consolidation, you end up with fewer carriers, but they're financially healthier, and they're able to invest in their networks more. And I wanna stay away from individual customer comments, but I think the announcement you just referenced kind of makes that case, that you know the consumer demand in Brazil is growing 20%-30% a year, just like in the U.S. They have done some 5G rollouts. They've been a little bit more modest. They have some coverage requirements that came with their spectrum.
But we think that once they get through this period of consolidation and they've completely integrated those networks, that you have healthier customers with a lot of healthy consumer demand, and that means they're gonna invest more there. So we, we really believe that you're gonna see that demand curve tick back up there. Now, we have been very clear that we continue to have elevated churn there as a result of Oi, and that means that for the next several years, our growth in Latin America is gonna be a little bit more muted. So we're cutting to low single digits, as we work through the remaining wireless churn for the next two years and have some more wireline churn in 2027 . But that's really on the churn side. The demand side, we think, should be ticking up.
Very good. And I think we've only got a couple more minutes till we're at time. So I wanted to just kind of, like, wrap up with, like, a question I get a lot from investors and just the overall kind of algorithm for the company, because I think, you know, the recent Sprint-T-Mobile merger has sort of, like, made it hard to have, like, a clear, comparable view across companies because of the way you've all accounted for churn a little bit differently in that transition. So maybe just kind of help us understand longer term, once we get this integration and churn behind us, how should investors think about the growth algorithm for American Tower in terms of both, you know, what level of revenue growth, what level of EBITDA and AFFO per share growth you're trying to achieve?
Sure. So if you just think about kind of the long term of our business, the way we view it is our developed markets should be growing at mid-single digits, and we think that's a sustainable trajectory for the long term. In our emerging markets, once you get through some of this carrier consolidation churn that we're seeing in Latin America, we think those emerging markets should be performing a couple hundred basis points higher. And that, that's even in a capital-light environment, where we're seeing less contribution from new assets, but growing those existing assets, that's the beauty of the tower model, is you don't have to put a lot of capital in to grow them organically. So you think about developed markets, kind of mid-single digits, couple hundred basis points higher in emerging markets, once we get through the churn phase.
You have CoreSite, and CoreSite's performing very well. That should be upper single digit or double-digit growth rates there as well. You couple that with some strategic investments that we make in terms of our internal CapEx program, and that should factor into that as well. Then you look at that in terms of also the margin expansion that we're focused on. With our cost controls in place, all that new revenue that we're generating on the tower side organically should be dropping down at kind of an 85%+ margin. When you look at the overall company margin, that should expand the margins over time.
Now, you couple that with there could be some financing headwinds out there, so we've put out our debt stack with all the rates there, so you guys can take your own view on what that's gonna be going forward. The forward-looking curves do indicate there will still be a degree of FX headwinds that are coming our way. And so when you kind of stack all that up, we see a business that's capable of delivering mid to upper single-digit growth, despite some of the headwinds that we're seeing in financing over time. And that's kinda how we think about the algorithm.
Okay, perfect. I think, sadly, we're at the end of our time, but, thank you very much, Steve, for being with us. We really appreciate it.
Thanks.
Thank you.