T-Mobile US, Inc. (TMUS) Earnings Call Transcript & Summary

Great. Good afternoon, everyone. Thanks for joining us. Again, I'm John Hodulik, the media and telecom analyst here at UBS. And I'm very pleased to announce that our afternoon keynote speaker is Mike Sievert, President and CEO of T-Mobile U.S. Mike, thanks for joining us today.

Hey John. Thanks for having me.

So we've got about 40, 45 minutes for some Q&A. I've got a list here that I can run through. But anybody that has a question in the audience, please feel free to shoot it over to me using the app. I've got it here in front of me, and I'll read that into the conversation. So Mike, big year for T-Mobile in a number of different areas. As we sit here at the end of '21, can you give us a sense of the priorities for the company as you look out to 2022?

Yes, of course. I mean they won't surprise you. First and foremost, we're focused on integration and on building this leadership network. It's the basis of so much in the aspirations that we have over the next few years. And my focus is to get it done in '22, really put the bulk of this integration behind us in '22. And if we're able to do that, it's just going to be fantastic and certainly well ahead of some of the earliest aspirations that we had. And all signs point to that being the plan. So we're really feeling great about the rate and pace of us building this leadership product by integrating the 2 companies, plowing all of those resources into a better network than Americans have ever seen and realizing the synergies from the bulk of that effort being completed in '22. So that's the biggest thing. Secondly, we've got to manage great relationships with our customers. And so part of this integration, it's all about putting an arm around that and making sure that especially those Sprint customers happily find their way to Magenta because underneath it, we see the Magenta performance, it's just firing on all cylinders. And so putting an arm around those Sprint customers is a big priority that we have. And then finally, and this again won't surprise you because we're very consistent around here, it's really continuing to tackle and address those big growth strategies that we have, like smaller markets in rural areas and enterprise and home broadband. And you're going to see us very focused again in '22, as we were this year, on those big underpenetrated segments that we have that differentiates our model because the '22 year, John, is really going to be about us, again, demonstrating that, unlike our competition, we have a synergy-backed model, we have a superior scale and spectrum portfolio and we have massive underpenetrated segments where we have room to run and generate profitable growth. So those are the focus areas.

Why don't we start with competition, and then we'll move to some of these other areas like integration and the churn and new markets? So first of all, how would you characterize the state of competition in the wireless market and maybe any comments on how Black Friday, which is obviously a big selling day, as far as the year, each year?

Yes. Well, if you've got your pen, I'll just take you through our performance week by week for all of Q4 to date. I talked a little bit about competition in our earnings call, which was also during this quarter, because I know a lot of people have big questions. They're looking out there going how can this -- is everybody winning? Is this sustainable? How is this going to work out? And I wish I had answers for you. I don't have answers for you on how it's going to work out for them. But I am clear-eyed about how it's going to work out for us because, unlike everybody else in this milieu, we have the advantages that I just rattled off: a synergy-backed model, the spectrum resources already deployed for a differentiated network and big underpenetrated growth segments that we can chase that are worth billions and billions of dollars. And so that's our focus. And so who knows? Will there be room for everybody to post crazy town growth over the next few years? I don't know. But I'm very clear on what we will deliver. And if you look across the management teams in this industry, we're the ones that have consistently, over and over and over again, posted what we said we would or better.

Right. We've been in a period where the switching pool has been sort of under pressure. I mean does that appear to continue to be the case? And do you expect that to be the case going forward as we look out into '22?

No. You can see, obviously, churn has been suppressed and so that certainly, I guess, can contribute to the switching pool. But look, I don't think it's -- I think this short-term pandemic-related switching pool diminishment that we were talking about last year is behind us. This was a vibrant shopping season. We're really pleased with what we've seen, the Black Friday offers, et cetera. And again, what we do is just a consistent execution. We don't have a new idea every week. I think that helps our overall performance. And so things are unfolding this quarter about like we expected. And that means it's a good news story.

Right. And then lastly on the competition front, just we've seen some increased activity in the cable companies, not so much on the handset side but on the services side. We also have some early-stage build-out going on at DISH. I mean anything you could do -- can you characterize what you see as competition today from those cable companies? And as you look out, does DISH start to become a factor by the end of '22 in your mind? Or when do you think you'll see those guys impact the market?

Well, DISH is a factor now. They have a fantastic MVNO arrangement. And they say that the one they struck with my competitor is even better, so this enables them. They're a factor now with millions of customers. And we're told at some point, they'll start releasing their own network. And we take them at face value on that. We've always assumed that in our forecast. Cable has been consistent. There's been no real change in the cable competitiveness. I've said before, we didn't actually anticipate that they would be able to come and grab this level of SOGA. It's about 10%, and it has been about 10% every single quarter for the last 2 years. So it's just a consistent performance you see from them. I guess, last quarter, it was down a little, 9% or something, but generally flat, has been for a long time. And that run rate is fully embedded in our forecast and anticipated by us.

Maybe now we'll turn to the integration. Take us through sort of what you have in front of you. I guess the decommissioning of the Sprint towers is really the sort of big thing that is going to happen here in '22. What's been the progress so far? And how should we think of sort of exiting the year in '22 in that respect?

Well, I just couldn't be more proud of this. What's happening is, yes, we have tackled this integration on the network front, even in advance of even our internal expectations. As you know, we hit 300 million people with 5G earlier than expected, almost 6 months earlier than our year-end target. We hit our 200 million people with Ultra Capacity 5G earlier than expected, many weeks earlier than expected. And we've begun decommissioning and, therefore, you see run rate synergies. And we're now saying we want to actually wrap that project up in '22. And that's been an aspiration we've had for a couple of years, but that's fully reflected in our plans now. And that's just something to be really proud of. What we are creating here is great for the country. The highest capacity network this country has ever seen which will bring people from the wrong side to the right side of the digital divide. But it's also great for our company because competitively, we aspire to be the leader in this space. And that means that we don't just start out the 5G era with a 2-year head start but that we have a 2-year lead throughout the entirety of the 5G era. And that's what's going to happen. Right now, as we sit, we have 200 million people covered with Ultra Capacity 5G, 200 million now. And by the end of this year, there will be an average of 100 megahertz of spectrum backing that up. And so there's been a lot of talk about C-band coming and eventually, it will get lit up. But that first tranche of C-band is 100 megahertz between AT&T and Verizon and doesn't even have 200 million POPs of licensing. And so for the next couple of years, we are going to have more 5G mid-band lit than AT&T and Verizon combined. And then you might say, "Oh, okay, but sure, they're catching up. Where will it be in a couple of years when they're fully deployed on that and get the second tranche of C-band?" Well, where it will be is we will be at 300 million people because what we're doing through that period is building. 200 million is in the rearview mirror. 300 million is 5x more land mass. It's hard to do. It takes years. And by the time the end of '23 comes around, we'll have 300 million people covered with, get this, a 5G layer that's 200 megahertz deep. And that just means a higher capacity footprint that's ever been built before. And it's one of the reasons why, having gotten that done, we think this is going to be a much more capital-efficient model than our competitors have because the capacity will be built and in place. And you see our capital aspirations for the out-years are, as a percentage of revenues, lower than today's during the height of our integration. And for that reason as well as the ongoing share gains, it's just an incredible cash flow opportunity in front of us and our shareholders.

Talking about the network. So 200 million POPs at 2.5, 300 million just 5G with the base layer of 600. Can you tell us -- I mean, first of all, what percentage of -- remind us what percentage of your customers have a 5G phone that can utilize these capabilities. And then again, it's fairly early, but is there anything -- and I think you've given us 5G usage, but anything you could tell us about ARPU trends or churn within that base of customers that are able to utilize those networks that you deploy.

Yes, thanks. First of all, yes, it's really neat to see how people are responding to this. About 30% of our customers have 5G handsets. And the ones with Magenta MAX, as I mentioned at our recent earnings release, are using 35 gigs a month. They're doing 8x more gaming than people on LTE, for example, many more times of video consumption. And it's just great to see how they're responding to this capacity that we put in their hands. And there are a lot of applications for 5G, and I know you're going to ask me about some of them. But I want to underscore that the first killer app of 5G is the smartphone that you already know. And when you give people a capacity like we have and a rate plan that doesn't slow you down and the freedom to do what you want, you use many times more data and appreciate many times more speed than you could get on LTE. And that's exciting because that conveys into how people feel about the business. In fact, right now, and you mentioned how it could translate into ARPU trends, we now anticipate we'll finish '21 with flat ARPU to last year, not down 1%. And that's really exciting when you consider that more than half of the run rate of our new customers are choosing Magenta MAX. By the way, in my base, I've got well under 20% penetration on that product and at similar price points from prior products. And so there's just a huge opportunity for people to select, to take better and more advantage of the highest capacity network that's ever been built. And that's not at all in our run rate, so that's just potential upside. We're going to see, to your ARPU trends, over the next year, 1.5 years, we're still going to see some pressure from continuing to move the Sprint customers over. And we needed to land them on a great plan, and there will be some pressure there. So we're still out looking. It could be plus or minus 1%. But it definitely will flatten out after that, and there's obviously some potential upside versus our outlooks.

What about churning that base? I mean, again, you've never really addressed churn within the 5G base, but you've got to think -- and we can talk about just Sprint churn versus Magenta churn. But just within the 5G, I mean, you've got to believe that given the difference in the offering versus other carriers that that's going to be lower over time too. So the conversion to 5G is going to be an overall good guide for churn for the company.

Well, here's a stat for you. Our Ultra Capacity 5G network now reaches 80% of our postpaid customers, 8-0, percent. Now I wasn't watching, but I'm not sure, did Hans or John mentioned what percent of their network, of their customers are reached by their highest capacity form of 5G? Because for us, it is 80% of our customers. And so that just kind of speaks to the fact that -- there's this weird sentiment out there right now. So it's 5G, who cares, it's not any faster. That's because you have AT&T or Verizon. I mean I'm not trying to disparage anybody, but when you rely on dynamic spectrum sharing, even your broadly available 5G is roughly the same speed as LTE. Ours is twice as fast because it's dedicated spectrum on a dedicated 5G stand-alone core. We're the only ones doing that on extended range, low-band. And then we have 80% of our customers covered by a network that rips at 300, 400 megabits per second and allows you to do the kinds of things that Magenta MAX customers are enjoying. That is so awesome.

So maybe also getting back to the integration. We talked about decommissioning the Sprint towers being done by the end of '22. Talk a little bit about the billing integration. That always seems like -- that's always a sticking point with transactions in the telecom space. I think 50-plus percent of Sprint subs are on the T-Mo systems now. When do you anticipate sort of completing this process? And anything you can tell us that we could sort of suggest that the risks from that conversion will be sort of minimal or less than we've seen in some other transactions?

Middle of '23. But actually, I expect it to be pretty opaque to the customer. I don't want it to be transparent when they've shifted from one bill to another. And we designed this integration in a way that no company has ever done where each aspect of the integration experience can be handled at the convenience of the customer or, in certain cases, to make sure we have an efficient business model to serve them. So being able to move their data onto T-Mobile, traffic-wise, that happened early on for some. Being able to move their rate plan to a destination rate plan, being able to migrate them to our exclusive team of experts customer care approach, being able to turn on Un-carrier benefits like global roaming and T-Mobile Tuesdays, et cetera, and then, ultimately, having the underlying billing and therefore, care and activation systems be all one. But hopefully, by the time that last thing happens, because we've harmonized the presentation layers and all those predicate moves, it's a nonevent for the customer. It's simply a synergy capture for us.

So it's all done, so the bill that these Sprint -- and I'm going to get to Sprint churn in a second -- that won't change. That has already been changed. And those customers, all of those customers, are getting a T-Mobile bill in a format that would not change after when you...

For some, it will. It will happen stepwise. As you change to being told you're on the T-Mobile brand and the T-Mobile network, we will present you with a T-Mobile bill and hopefully get you on a T-Mobile rate plan such that the fact that you're still on legacy Sprint billing systems underneath it all is not something you should have to worry about. And that allows us to move through and do that conversion of the billing system more at the convenience of our business model. And there's a lot of work to be done. It will be the last piece, it's always the last piece. But I do expect to wrap it up in the middle of '23.

Got you. So I thought one of the most interesting things coming out of the quarter was the commentary on what net adds on the postpaid side would have been if churn at Sprint would have been more in line with what you see on the Magenta side. And it really suggests gross adds are really not an issue. Frankly, you could see gross adds come down and still have 1 million-plus subs a quarter, at least in the third quarter, what we saw in the quarter. When can we expect Sprint churn just to -- has it started to come down? Are there any worries going forward? Do we start taking down the Sprint network or finish up with the billing conversion that suggests that that's going to remain stubborn? Or what has to happen for that to -- maybe it's just getting 5G phones into all these guys' hands. But when can we start to see those 2 churn figures start to really converge?

Yes. It's really a function of phones and plans as much as it is network, and so it's all 3 pieces have to be in place. What we've spoken to so far is mostly network, that about half the customers are end to end now domiciled on the T-Mobile network. Most of their data traffic, even if they're not, is carried by our network, but about half of them are domiciled now on the T-Mobile network. That's a really important predicate. But you also need to make sure that you have the rate plans and phone plans that customers love. And a lot of our Sprint customers still have these leases that aren't, for many, a satisfaction driver. And we've got to get them migrated to installment plans, hopefully on a new 5G phone with 600 megahertz low-band, that there wasn't a lot of on the Sprint side, because that's the basis for the breadth. People are very impressed by 400 megabits per second. What they really want is some signal. They want coverage where they go. And for us, that's 600 megahertz. It reaches further from towers and penetrates better into buildings than any signal. But you need a phone that's compatible. And so all those things are things that have to happen through the period. And what we've decided to do, we think, will lead to less overall churn in our business plan over a 2-year period. But it is more compressed. So what we're doing is tackling this whole project in '22, starting in the last quarter or 2 but then wrapping it up by the end of '22. And we're going to move all those customers to be domiciled on the T-Mobile network. And we think that's going to lead to a faster time frame to when they love T-Mobile just as much as Magenta customers do. We think it's going to result in less overall defections. But because it's compressed into a shorter time frame, you shouldn't probably expect the Sprint churn rate to fall right away. On the other hand, the underlying performance of the Magenta business, as we build this network, continues to improve. And so there's potential offsetting effects here. I mean we're seeing Magenta performance on par with anything ever in our history throughout our entire Un-carrier journey. It's incredible. And so those things will be both blended into what overall looks like it's going to be a good year, but there's puts and takes on all of it. We'll guide you on it all, of course, early in the year.

Okay, some moving parts there. But so just so I have that correct, the Magenta side of things have been improving -- as you get the network moving from a churn standpoint, likely to continue to improve as you get the network where you want it to be. And obviously, you're making tons of progress behind that. But on the Sprint side, it might not be as sort of linear given you got to move these other 50% of these customers over to the end-to-end T-Mobile platform. And that will happen at the end of the year.

Yes. And we're going to be migrating customers. When you touch a customer and deal with them, on the back end of it, they're going to be delighted. But just by having the conversation in some way, it's just the nature of sort of you're talking to them now. And so integrations create integration-driven churn. And that's something we intend to compress into a shorter time frame, get it behind us, love these customers, get them on to the best network that's ever been built and then watch the run rate performance start to take off. And so we think that will lead to less overall defections. But what defections do happen as a function of the integration are happening in a shorter time frame. And so while we're going through it, it will slow down the rate of improvement.

That makes sense. And then in terms of getting that last 50% over to the end-to-end T-Mobile systems, does that happen ratably through the year? Or is that more front-end loaded, more back-end loaded? Is it something that you just block and tackle through the year and you're done by year-end '22?

Yes, it's hard to say. I mean I know you'd love some shaping on it, but it's going to depend. In each geography, we take big tranches of customers and move them when we think that the Magenta network is a better overall experience for them because it's different and different is usually better, but we like to make sure it's almost universally better before we switch you. Otherwise, you'll be like, "Hey, what about this one road where I had coverage before," and you won't be happy about that. And so we try to do it area by area and bring people across. And that's when you do those big final pushes. Before those final pushes, it does happen somewhat ratably. But yes, it's hard to give you kind of shaping on it. But I'll think about whether or not that's something that would be helpful for our next release.

Okay. You started off talking about the new areas of growth that you can penetrate. And I think that in the past, you talked about the business market and sort of more rural and suburban markets. First of all, before we go into each one of those, I think what we've seen is some investment that goes towards capturing that opportunity. And those investments have been somewhat larger than the synergies that we've been realizing so that, as a result, the margins have been relatively flattish, EBITDA growth has been relatively subdued. When do we get to the point where the synergies, maybe because of what we're seeing on the network side, start to really overcome or surpass the investment and we start to see EBITDA start to grow?

It's a great question. I mean, at a certain level, it's kind of cool that we're outgrowing everybody on EBITDA even in the height of this [ coup ] we're talking about, right? Because it's not just investments in new things, it's all of this work that we're doing. And so it's really great that we're able to post these kinds of results and demonstrate progress to shareholders as we go along instead of just asking for blank checks for some wonderful period down the road. That's not our philosophy. So I'm delighted with how things look. And we'll guide '22 for you in the spring. But '21, some of these underlying trends have just been really, really encouraging. Now to your question about when it all starts to pop, obviously, it's when you get to run rate synergies. And the big milestone is to complete the network turndown because that's where the cost is. Those are where the majority of the synergies were. So you start to see those in our run rate in '23, which I think is a big transition time for us financially. But the other thing I want to point out is that for our company, the geography is just different, right? So we have a much less CapEx-intensive business model than our competitors. And so it's important not to focus just on the margins. Where we're really focusing as a management team is on the cash production per service revenue dollar. And on that measure, I mean, cash is king. We see our way to being the leader in cash production per service revenue, call it, cash margins, if you will, by '23, and better than AT&T or Verizon by several points and a lot better by '24. I see us in the 20s percent range in '23 and the mid-20s in '24. And so that's really exciting cash production on a service revenue basis. And again, you have this massive capacity network and underpenetrated segments you can efficiently chase, which means you don't have to -- and our plan is burdened with -- we lease most of our fiber. And so that's over on the margin side. But on our CapEx side, it's much more efficient because you've got this massive network that by then reaches 300 million people with 200 megahertz of 5G. Now you spend a few years filling that up, and you start to see our CapEx per revenue dollar behind this integration really fall to much more efficient levels. Not unprecedented levels, our competitors have been there at points in their past. It's not a crazy CapEx plan, it's just a more efficient one.

Got you. Last thing on those sort of new growth opportunities, business market versus the rural and suburban, I mean when -- first of all, which one are you more excited about? Which one's a bigger opportunity or maybe a more profitable opportunity? And we might be seeing the benefits to that now, but when do you start to see that hit your stride in both of those markets?

Yes. One of the things you'll note is that the things we talk about are in our core business. And it's one of the things that's always made us successful at T-Mobile is we focus on our core. We're interested in other things, we'll come to that. But we're really good when we really focus our energies throughout the entire Un-carrier journey tackling opportunities. And the 3 in the core that we're most excited about, we've been very consistent about, have been small markets and rural areas. I'm not talking about Helena, Montana. Helena is included. I'm talking about 40% of the country, 40%, where we have a share in the mid- to low teens, right, not 35%, 40% like in the big markets. Remember, the only reason we're #2 nationwide is because of SMRA, small markets, rural areas. We're #1 in the top 50 cities. And we have the wherewithal to defend that with a rapidly improving network. We got to be #1 in those top 50 without having the best network. Now we have the best network so we'll be able to defend, probably even extend that capacity. But small markets and rural areas, we've never even been relevant. And now not only are we going to be relevant and catch up, but remember, of the 3 major capital-intensive networks out there, we're the only ones committing to massive 5G in smaller markets and rural areas. We're going to 300 million people. That means we're reaching you with not just a competitive product but a differentiated one in small markets and rural areas. AT&T and Verizon didn't express any ambition to go there with mid-band 5G. And so that's really exciting. Second, enterprise. We've been historically the guy you'd bring in maybe so you can get a better deal on your AT&T service. And then maybe you throw us 5% for our troubles and send us packing. That's been our journey through the Un-carrier. And it's resulted in about a 10% share. And now we have strategic relationships all across the board. This morning, we announced that we've become the preferred provider to Alaska Airlines. We now serve substantially all of the airlines. I mean think about that, that's the torture test customer. They're going into every town and city in this country, and they're picking T-Mobile after extensive testing on who's best, not going on reputation. Lots of progress with oil and gas, with big government, with state government, with financials and so on, strategic relationships growing in importance. So that's a really important area for us. And it doesn't all come in on voice, by the way. Some of the non-voice business we're doing in T-Mobile for Business is as important, strategic and profitable as a voice business. So that becomes interesting as we keep on talking to you about accounts and ARPA, not phone lines and ARPU for reasons that include that. And then finally, home broadband. By the way, we've passed our year-end goal, so we've now got more than 0.5 million customers, which was our aspiration by year-end. The team is just on all cylinders on this product. And it's a phenomenal product with NPS scores that take big jumps from people's prior service, higher Net Promoter Scores than any cable company out there. People, boy, when you give them an option, a lot of our customers on home broadband are coming in suburban and even urban areas from cable, which is fascinating. It's not all just greenfield stuff where nobody's ever had an option before. It's both. And that's really interesting. So we see our way to 7 million to 8 million, which is only single-digit penetration, and that will be a multibillion-dollar highly accretive business for us when we get there. And it's all based on the capacity that's already totally funded by the mobile business, so not a CapEx burden to model, which means we can get there and profitably offer a great value to customers.

So that's great news on the 500,000. So at what point do you sort of hit your run rate in terms of net adds on the fixed wireless side? Sort of what needs to happen? You've got the -- you seemingly have the network in place, right, with the 200 million POPs at 2.5. Is it the CPE? Is it the processes? How quickly can you get to, say, 200,000 net adds a quarter kind of thing? Or when does it sort of peak? And you don't have to tell me where it peaks, but when do you hit your sort of run rate?

I think '22 will be a lot bigger year than '21 was, and '21 was fantastic. I mean we launched it in '21. We were beta before that. We launched it this year, have already surpassed 500,000 customers and we're on our way. Next year will be bigger than this year. So we've already seen some of those trajectory changes that are important, one of the reasons I beat the year-end goal in the middle of the quarter. So that's exciting to see. And it will flow at the capacity. I mean we have to get the -- this network is still being constructed and we're still moving the mobile customers across and still lighting up those Sprint Keep sites, and we've got 10,000 new sites that we're building. And we've also got megahertz to deploy. Remember, we're on our way from what was last quarter, probably 70, 80 megahertz of spectrum lit up on 5G to about 100 by year-end, on its way to 200 by the end of '23. And all that depth of spectrum opens up the potential for more homes passed and/or more opportunity even within the homes passed areas. And so all that colludes to allow us to keep hopefully going at a more rapid pace as we take off.

Got it. So it ramps largely with the network deployment. Anything you can tell us about ARPU? Obviously, probably too early for churn but maybe usage or anything you can tell us about sort of what users are experiencing or what kind of sort of the financial profile that you expect to see from the fixed wireless sub.

Well, the usage is almost exactly like we thought it would be, at several hundred gigs a month. We have a single-digit percent using more than a terabyte per month, like you would expect. It's roughly 10x a mobile user, a good mobile user, on 5G that's consumptive. And so it's about like we predicted, which is great news because at those payloads, even assuming they continue to grow through the period the way data consumption tends to do, when we approve you, we approved you for a reason, which is we predict your household can handle that kind of payload. And that's why Net Promoter Scores are turning out to be quite positive. I'm not saying we've got it all figured out. We're new in this industry. We're going to make mistakes. We're going to learn how to care for customers and how to put them in the right neighborhoods with the right service and so on. But for a new business that we're learning, so far, the first year out of betas looking really good.

Got it. That was interesting, what you said, "When we approve you." So can you explain that process? So I guess, somebody applies for a fixed wireless service, either through a store or online, and you guys go through a process where you see whether or not you've got the facilities or infrastructure in the market to serve you without having any sort of impact on the network. Or maybe could you explain that process? Is it ironed out? And again, probably too early, but is there any numbers you can give us in terms of what percentage of the people that apply are approved?

Sure. Yes. So basically, you take the whole country, what, there are 140 million households, something like that, and you analyze the network based on its capacity and whether or not we can comfortably say that starting now and for any period in the indefinite future, no normal amount of mobile usage can soak up all the capacity on that sector. And we assume that mobile usage goes from where it is to an average of 80 gigs. So we assume that it will take off. And we predict into the future, and it looks like our plan for that sector will exceed the rate of growth of mobile consumption, including market share gains and usage gains. And once you've concluded all that, then if somebody enters that particular address in the system, they get told, "Yes, your address is approved. Would you like us to call you? Would you like to stop in? Or you can engage with us digitally." And we're getting better at all those activation processes. They're not as smooth as they should be yet. But that's how that works. And in any given sector, we can only support so many households. And so it's dynamic, meaning once the nth person applies, we have to start telling people after that, "You're on a waiting list," because we need to make sure that at all times, the mobile experience is unaffected. It's fantastic. And so if your whole neighborhood group gets together and says, "We're all getting the broadband," you better be one of the first people...

"We're all switching to T-Mo"?

Yes. Right now, at least, you better be one of the first people to apply. But as we speak, we have more than 30 million households that if you type in the name right now and said, "I'm in," we would tell you yes. And so that's how it works. Like I said, we have a lot to do to optimize it but I'm really proud of the whole approach. And it basically says, so far, and in our initial planning period for this business, we don't intend to allocate capital cost to it because it's using, literally by definition, as I just explained, excess capital, excess capacity generated by the mobile capital.

Got it. And then you said you can only support so many households per sector. You guys haven't said how many households per sector you could support. And I guess it would certainly depend on, again, the assets in that, so how big of a sell site, how many radios, how much spectrum.

Exactly. And that's a function of time, right? Because right as we spoke last quarter, we had a nationwide average of less than 80 megahertz of 5G lit up in mid-band. It will end the year around 100. It's on its way to 200.

Right. Is it number of households per sector?

Yes, how that capacity just starts to ramp over time. And that means many places where we used to tell you no or you have to wait, we'll then tell you yes. And we keep your name. If you ask, we tell you, "Hey, we're not in your community or on your street, but if you like, we'll get back to you when we are." Now we're gathering database.

Makes sense. And then lastly on fixed wireless, and I'll move on to some more of the financial stuff. You said 30 million households right now, if you type in your address, you can get it. How does that scale? Again, obviously, like you said, it depends on the infrastructure, it depends on the subscriber or the spectrum deployment. I mean I'll pick a number. When does that number hit 100 million? Or how high does it get?

I'm actually focused a lot more on if -- obviously, it grows, but it's a choice we have as to whether to focus on making that top line number grow or whether we're much more focused on the conversion to get the bottom line number. And for me, this business looks like the 7 million to 8 million aspiration is really the North Star. And ultimately, that will be a single-digit percentage of the overall potential, right, because this capacity is going to reach 300 million people with 200 megahertz. No one else is doing anything like that. And so that means, ultimately, you're going to have a lot of flexibility as to who you can approve for home broadband. And what broadband business model have you ever seen before whose big financial success case is a single-digit penetration of the opportunity, right? They always -- their breakevens are 35%. They hope to get to 45%, all these fiber companies going -- this is just phenomenal. And we might do capital-dedicated stuff later. I mean we happen to own a ton of millimeter wave spectrum, for example, which we don't talk about a lot. We owe more than AT&T. But this is a mid-band-centric strategy based on the highest capacity mid-band country portfolio that this country has ever seen or from what we're told by our competitors, we'll see.

Right. Maybe one more follow-up question on ARPU then we'll go to margins and the financial profiles as we exit '22. You said it may be a little bit of sort of flattish, you got some good tailwinds on the ARPU side, one of them being the Magenta MAX plan. First of all, you said 50% of gross adds coming in at Magenta. Is that what we should think about as your sort of long-term opportunity for customers on that Magenta plan? Or do you think, over time, you could surpass that? And then could you also talk about the tailwind that is international roaming. I mean I guess it's relatively small for you guys. I mean it seems to be one of your selling points you have for free. But is that also helpful as international travel comes back or relatively small?

Well, for the second question first, the point of distinction that we have about international, it has been less operative for our customers over the last couple of years, and that's now coming back into the lens. I think that's great. So many people I know tell me they switched to T-Mobile because of international. That's how they found their way to us. And we were the ones and still the only ones that say, "Look, you don't have to sign up for anything. There's literally nothing you have to do. Just go there, turn off airplane mode. And while you're plane's taxiing to the gate, in fly all your messages and e-mails." And it's an incredible delight factor that, even though our competitors are doing way more than they were a few years ago, is still a differentiator. And so obviously, that will come back into the lens. But to your first point, no, look, we're going to get through the Sprint integration. And I want to put an arm around those customers because Sprint churn is more important than trying to see some surprise here. So we've said, we see it next year in the multiyear period. I think in Analyst Day, we said we saw plus or minus 1%. Of course, we beat that this year. It's going to be flat, it's fantastic. So who knows? But our real focus is on ARPA, average revenue per account. And we really feel like this is the more important metric because, first of all, we're the only ones that we know of growing accounts and we're growing them like crazy; and, second of all, ARPA is on the rise. I mean it's up $2 this year. And to us, that's really important because it just shows that customers are willing to deepen their relationship with us. And if that means taking out a data device that's 5G-inspired or it means establishing a new line for a great deal or it means other things, I think that those are all ways for us to welcome more total accounts and then deepen our relationships with them. And so more and more, you'll be hearing our aspirations in the out-periods expressed that way. And I do think there's an opportunity to grow. I mean 5G is opening up lots of use cases and more and more people are going to want to take advantage of that.

Let's shift to margins. This year, we estimate you guys will have core service EBITDA margins of about 40%, with AT&T at 55% and Verizon at more than 60%. And earlier, we talked about having the highest sort of cash conversion ratio among the 3 carriers. Obviously, you're going to have to see some real convergence from a margin standpoint to get there. So I guess, in your mind, how should we think of your margins relative to AT&T? I mean, again, you've got about -- it's almost 1,500 basis points of catch-up there. Is all the leverage in the capital side? Or do you think you can approach what you're seeing at AT&T at this point, even with the leasing, I would say, somewhat lower ARPU and faster growth?

Yes. I mean you know that -- here are the opportunities, right? One is that we continue to grow accounts and ARPA. That's really important. And we love the accretive nature of our growth, particularly on the Magenta side. That's not as good as we've ever seen it through the entirety of the Un-carrier journey, really great. So we have to get through that. And then secondly, we get the synergy run rate into our margins. And there's so much goodness there. And that's one of the reasons we're so focused on executing this integration better and faster than promised because that unlocks that aspect of margins. And then third, it's our more efficient business model. We have the scale because of our resources, particularly on spectrum, that once you have built 300 million people with 200 megahertz, you can be more capital-efficient. And that's important also. So already by 2023, just less than 2 years from now, we believe we have the most cash productive business model in this industry. Just look at our plans versus consensus estimates at AT&T and Verizon, maybe 5 points better, we'll be north of 20%, we think. And then by '24, we see 8 to 10 points better cash production per service revenue than the other guys. And so it really shows you that while we have a different model, they build fiber, we lease it, we all do things a little differently. And frankly, neither of them provides the kind of transparency that we provide on core wireless. We don't even know what Verizon's margins are. We don't, we haven't seen them in 2 years. So we have to rely on you to guess with them and stuff. We provide this transparent picture. And based on everything we know, we have the most cash productive business model in this industry starting in '23.

So final question for me. Given the EBITDA growth that's sort of implied by that catch-up in margins and/or your commentary around CapEx with the deployment of 5G, you're going to delever very quickly over the next 18 months. Based on sort of latest data, when do you expect to reach your leverage target? And when does the conversation shift to buybacks? Frankly, from a lot of investors that we talked to, it already has. But when does that sort of start to enter the conversation?

Well, we wanted to be really transparent about our aspirations earlier this year when we laid out our multiyear plan. And we were very clear that we see a path to investment-grade inside the planning period, which back then was '21 to '25. We see a path to up to $60 billion in buybacks during that time, $65 billion or more in cash flows, et cetera. And what I said in the most recent couple of earnings calls is all of those aspirations are still intact. And they're so exciting. I mean it's huge to think about the potential there in August. And we're making milestones, right? I mean cash flow this year is growing rapidly enough. We gave you a 3-year CAGR of 45%. We were upgraded by all 3 agencies in August. Team just completed a small capital raise last week at phenomenal rates in the investment-grade market. We do see our path to in Corporate Family IG. And it's just going to be choices that we make along the way. We haven't made them yet so I know you'd love to -- but do you focus on delevering? Do you focus on maybe more of a stepwise approach to share buybacks instead of waiting as long? Are there other organic things we need to think about that would be upside to the plan? All those are things that are within our grasp as we're now at the jumping-off point. With the integration so clear in our eyes and focused on being wrapped up by this time next year, now we can start to think about all that stuff because we can see it. We know what we have to do, and we feel very confident in this integration and in this network build, which has been the predicate for so much of the value capture. And the difference versus 1.5 years ago or even 6 months ago when we did the Analyst Day is we can see it so clearly now. And it's just because of great execution by this team.

Got you. And just lastly, does the share price influence that decision? I mean do you also sort of look at all the execution and the progress you've made and the sort of the cash flow that's coming and think to yourself, "Well, because of what we've seen in the dislocation in the market, we have an opportunity here with the stock at these levels." Does that change the equation for you guys at all? Or are you guys largely immune to what's going on in the stock world?

Well, I can't take you into the calculus and any internal deliberations, but I can tell you that our view is that we think we are being painted with a sector brush. We do think that at this point, the T-Mobile stock is significantly undervalued. And normally, it's not a CEO's job to weigh in on that and make views. But look, we're really confident in this business plan. We know exactly what we have to do. This is a team that has always performed and done the things we said we would do. And it's focusing on the core. There's all kinds of cool upsides we can go chase and will. But the things we know how to do that are within our grasp, including some potential scenarios in case we misfire on anything and how we recover, just makes it so clear that we have confidence in this plan. And because of that, we look at the valuations and it looks like we're undervalued. And I think that's because people are worried about the sector. People are like, "Well, they don't understand where all these net adds are coming from and it doesn't look sustainable. Some shoe is going to drop and let's watch that movie and then see what happens." And my view is, yes, I'm watching too, but I'm very clear about what will happen at T-Mobile because we're the only ones with a synergy-backed model, the superior capability on the network backed by the spectrum portfolio that's already built out and massive underpenetrated growth trajectories that we can profitably go chase. The other guys, I don't know what will happen, but it's not my job to predict their business for you.

Makes sense. Well, Mike, I really appreciate the time. You're always very generous each year, so we appreciate you being here. It was very informative.

Thanks a lot, John. We'll see you next time.

Okay, take care. Thank you, all, for joining us.