All right, well, hello everyone and welcome to our 2024 Analyst and Investor Day. We are coming to you live from AT&T Stadium just outside of our hometown of Dallas, Texas. So thanks to all of you for making the trip down here to be with us. Before we begin, I need to call your attention to our safe harbor language. It says some of our comments today may be forward-looking. As such, they are subject to risks and uncertainties described in AT&T's SEC filings. Results may differ materially. Our Analyst and Investor Day materials will be posted to our IR website following today's prepared remarks, but please pay attention. We have an awesome agenda for you today. You're going to hear from several members of our senior leadership team.
After we get through our presentations, we are going to have a Q&A session, so we ask you to hold your questions until then. The presentations and the Q&A are going to run a little over three hours, and we are going to have a short break about halfway through. With that, I'm going to get off the stage and I'm going to turn it over to our CEO, John Stankey. John.
Thanks, Brett. And thank you. I've never been in a room with all of you where you've lifted your heads up from your laptop and applauded. That's amazing. So it's never happened to me before. It's good to have you all with us today, those of you online, especially those of you in the room who made the journey here to Dallas to join us today, and we hope to make this worthwhile for you. When I think back of why we're doing it in this particular location, obviously we've got our name on the building, but equally as important, it's convenient to the airport, but it's also a manifestation of what we've been working on to really try to drive the right degree of perfection. And this is a business in a location we spend a lot of time working on connections. We do a lot with the Cowboys.
We've had a great partnership with them and all their telecommunications needs. But I was thinking back earlier in my career in about 2009 when this place opened, and we were working to engineer it to actually handle smartphones at the time and wanted it to be a unique experience, and we worked really, really hard to make sure we could get one to three megs to each handset in the location, and I think about all the work we had to go through and about $100 million later where we are today. This has really been ground zero of where we've driven perfection of Wideband CDMA and large dense buildings and facilities, and it's indicative of what we're trying to do and getting back to our roots in this business, and pleased that you could join us today, but we don't want to just talk about the stadium.
We want to talk about where we're going with the business, and I want to move into that and give you a sense of our direction and our longer-term strategic plan to, in fact, be the best connectivity provider in America, and I think the leadership's going to do a good job with you today explaining how we're going to actually execute around that. As I started with, for about 150 years, AT&T has been connecting folks, and we've taken it as part and parcel and critical to everything that we do, but in 2020, we kind of took a timeout and realized we needed to rededicate ourselves to this mission and this direction, and we felt like that was a better path for our shareholders and our owners to think about how to invest in our business.
We articulated three primary focus areas around how we would allocate capital during this period of time. The first, we wanted to invest at the top of the industry in our core capabilities of our product, our infrastructures, and our network. And we did that: $140 billion over a five-year period in spectrum and infrastructure in this business. The second priority we set out is that we said we wanted to make sure that our balance sheet was in a better place, and we were going to have to take a time and a period where we took excess capital and made sure we got ourselves a little bit more financial flexibility. Standing here today, $25 billion less net debt.
As a result of that, you know where we'll arrive in the middle of next year at 2.5x adjusted net debt to EBITDA and in a position where we feel like we've got the financial flexibility we need to run the business, and the third thing we wanted to do is we knew we were going to take our investors and shareholders on a bit of a ride here. We didn't want to lose sight of that. We wanted to ensure that we could protect the dividend and return about 45% of our free cash flow to our investors in the form of dividends as we were going through this transition, and we did all those things, and I stand here today with a business that I think is financially stronger. It's more streamlined, and it's much more focused on connectivity in our direction.
And that gives us an opportunity to think about this inflection point or this pivot. And we did a lot of things to operate the business in a better way during that period that achieved it. And I think it's worth spending just a couple of minutes on it before I get into what we're doing in the future. In our wireless business, our largest business, our most important business, we focused on prioritizing customers, simplifying our operations, simplifying our offers, and being consistent in what we did in treating all of our customers, whether it's those that we already had or those that were coming to us with the same level of respect. We did that under the context of Best Deals for Everyone and how we engineered our service experiences. We also felt it was very important for us to invest in our networks and our product.
We did that. I think we did it in a very smart way in our 5G investments. We timed them right. We did them in the right places where our customers were. We've left opportunities open for us to take advantage of the next generation of technology and price points as we move forward. The net of all that work was we grew 10 million net adds in the postpaid voice business since the second quarter of 2020. It didn't stop there. In order for us to grow at that level, we needed to treat our base effectively. I think we did that as well. We were able to go and grow our POPs in that base. We were able to drive more revenues. We grew over $9 billion of service revenues since 2020.
And if you look at that within the context of the industry and adjust out equipment revenues and look at the large players, that was about 40% of the service revenue share during that period of time. The same plays that we ran to grow customers were important about keeping our current customers happy. We did it on the backs of respect. We did it on the back of better service. We did it on the back of simpler constructs. And in 13 of the last 15 quarters, we were able to drive the industry's leading postpaid phone churn, showing that we had a formula of network performance, value, service that resonated with the customers and kept them with AT&T. And it wasn't just enough to do that.
We wanted to make sure we were growing the right way, that we were investing the right way, and we were able to do all those things while at the same time improving our margins. And since 2022, there's been a 420 basis points increase in our wireless operating margin. And I think you know what this all resulted in: better performance, better transition into EBITDA from service revenues. And that was a good, good journey for the business. But the best news about this is we still have more that we can do. And we're confident about that. And we're going to talk a little bit about it. The story was very similar in our consumer wireline business.
If you think about 2020 and the decision we had to make, it was a decision that we felt like fiber and high-performance networking was going to be critical in this space, and we needed to refocus the business on connectivity. Since 2020, we've passed 11 million more fiber locations. We will end this year passing 29 million locations. If you think about it, since the second quarter of 2020, we more than doubled our AT&T Fiber subscriber base to over 9 million subscribers. Since 2021, we've nearly doubled consumer fiber revenues. We've done this again in a way where we've been profitable in how we've approached the market. We turned wireline EBITDA positive in 2022, and then, through disciplined cost management and the scaling benefits as we started to receive them by growing our footprint in various areas, you saw our profitability and margin performance increase.
That trend is going to continue as you go forward. You'll hear a lot about that in today's material and presentation. Maybe summarizing the last four years, I would point out that it's been marked by a lot of hard decisions. We're midstream on two of those hard decisions where I don't think we've achieved the level of performance that we think we can ultimately get to. One of them is where we are in redirecting our focus in business markets. I think many of you in this room have heard me say this is a place that I don't lose sleep over our right and our ability to be relevant in this space. I think our brand, our capabilities, how we go to market, what we know how to do that's core to our business are well matched to this market segment.
I'm going to make a few more comments a little bit later as to why I think it's economically imperative that we are successful in this segment. We're not quite where we need to be. Similarly, in a second area, we started charting our path out of the last century's technology. We said that we needed to get out of kind of our core and legacy business. And I'm proud to say that I think we have done more operationally and from a regulatory perspective than any other company in our industry on this front, but we're not done yet. Today, we're going to spend a little bit of time talking to you about what do we need to do to turn the crank to get that over the line and how we finish that work.
When we complete these additional two steps of the work that we've already done to operate other parts of our business more effectively, we're going to have an even stronger AT&T. We stand here in a strong position today, an opportunity to make a decision and transition to more offense, but we will be even more agile, more focused, and more profitable than we are at this time after we complete this work because our focus is now on building the largest, highest capacity converged broadband network with the lowest marginal cost in the United States. And that's what we're focused on as we move forward in the future. There is more opportunity ahead of us for growth in this business. And we're going to talk a little bit about what that means to us. We have no intention of slowing our momentum.
We're continuing to work on our cost structure and re-engineer it. We're continuing to work to improve our networks, and we're continuing to improve our product portfolio. As we think about capital allocations against those actions, there are four fundamental beliefs that we keep in mind that have guided us to this point and will guide us in the future. They're rooted on the basis of connectivity, and they're rooted on the basis of consumer demand and behavior. And I want to spend just a couple of minutes talking about those four beliefs with you and how I view them and how the management team looks through this lens as we're making decisions. The first key belief, no surprise, it's probably not earth-shattering, demand for high-performance connectivity is going to increase. It's not a newsflash. It's been happening for decades.
We've seen since 2020, mobile bandwidth up 70%, fixed bandwidth up 60%. If we look at data consumption over the next five years, we expect that's going to increase about 80%. That trend is probably not going to shift. We still think there's tailwinds behind this, just fundamental structural changes that are occurring. How consumers entertain themselves is only midstream in the transition. We're moving from broadcast and multicast technologies to discrete streams to devices in the home. We're thinking about upgrades from HD to 4K and the bandwidth that that consumes, the continued rise of gaming in the household. All those things are tailwinds before we see any new innovation, any new applications that we know are going to drive demand higher.
But we also know that there's dynamics going on in new products and services that are driving more upstream demand that are going to grow, that are relatively new technologies: AI, user-generated content, Augmented Reality, Virtual Reality. As all those things start to work our way into our work lives, into our homes, into our businesses, there'll be further demand increases. IoT is often cited, sometimes dismissed for being a bunch of little small things with small amounts of bandwidth. But as I think about the number of cameras that are manifesting themselves in our daily lives, in our vehicles, around our homes, in shopping malls, in retail stores, and we start to think about all that video that can be moved up for analysis into AI engines, I think there's additional demand, and a lot of that is going to be upstream driven.
And that demand moves to our second belief that ultimately fiber is destined to be nearly ubiquitous by the end of this decade. And you know what? This isn't new either. This trend's been going on for about 40 years. We started deploying fiber back in the mid-1980s, and it was about connecting major switching centers in metropolitan areas. And then the economics started to prove out to say you could move beyond major switching centers, and you could go to places where large amounts of voice traffic and trunks were aggregated. And the 1990s were characterized by that. And then we hit the 2000s where data became meaningful. And suddenly, we started to see fiber distributed to nodes and businesses to pick up all that data. And then the 2010s rolled around, and we're at a point where we needed these dense aggregation networks for all that data.
It was aggregating traffic at cell sites with fiber. It was making sure that you had the right collector layers for middle-mile infrastructure. And then about 10 years ago, we kind of crossed that final threshold. The economics and the business model of getting fiber all the way to the end user started to make sense, and we began the final stages of this race. So our belief is that by about 2030, we'll be done with that. It'll have been about 45 years from the start to the finish. That may seem like a long time, but it's actually about half the amount of time it took copper to move the same level in distribution and pick up the same number of customers. Our belief is that fiber ultimately wins. It's destined to be nearly ubiquitous because the physics say it will win. RF technologies great.
We've done remarkable things with it. Satellites are going to play a role. They're going to be very important for picking up coverage in areas. We're going to continue to push the limits of what wireless technology can do with Wideband CDMA and other technologies. But at the end of the day, fiber has better performance upstream, downstream, latency, resiliency, scalability, and marginal cost, and it will win. In the next decade, networks aren't fixed and mobile. Networks are fiber with different access technologies hanging on the outside of it. And that's the fundamental direction of converging our networks and our approach to the markets that we believe is taking shape as we move through the balance of this decade. We believe that strategically placed fiber collecting workloads is the best way to earn returns today and will be a determinant of returns in the future.
And that's why that second belief is so important. Our third key belief is that customers actually don't relish complexity in their lives, and they would like to do business with one company in a more simple fashion. Again, probably not an earth-shattering belief, but an important one. If we think about where we started in our industry, constructs of what we did in the 1920s, the 1930s, and then in some way reshaped in the 1990s were the laws under which licenses were granted, industries were structured, and what we did. When you think about what happened during that period of time, regulators set up the construct about how customers were to deal with their particular service provider. What we're seeing now is technology has moved beyond that. What we're seeing now is that customers can buy from one place, and they actually want to do that.
When we go and we evaluate and we survey the market, 77% of our customers tell us they prefer to get on the internet with one company, and if you don't believe that they're telling you the truth, we look at those customers that buy converged services, and what do we know about them? We know that they churn less frequently. We know that they have much higher NPS scores, and we know that they drive better value back to the company that they're doing business with, so we see this as an important dynamic that it's not only the technology that's helping us down this path, but the customer is going to drive us down this path as well, and it's time that the industry make this final step, and we believe this move to convergence drives a unique opportunity for AT&T.
We're positioned to make the shift over the next decade. We lead with quality and differentiation. Fiber drives the best product. We own and operate and can build the best wireless networks that are out there. We can put it all together and build better features and capability, and we drive this on value, performance, and not price. We move away from handsets and routers and cars that are connected to households that are connected. Same thing for a business, and we believe this is the direction the industry is going, and it's how we lean in. In fact, I would submit to you that the technology foundation that we've been building networks on right now is really in its adolescence.
As the mindset and the intellectual property of talented companies and innovation gets driven to bring these together, we're going to see some very unique capabilities start to emerge. Now, let me just touch on the fourth and final key belief that we hold. Ultimately, we believe that those who lead in this industry are those that are going to have the most dense and distributed access points. Anybody who drives scale and distributed access points is ultimately going to be in a preferred position. The inherent value of a network increases when you can connect more people to more people. That's the very definition of networking. If you can do that with better performance at a lower marginal cost, you're going to drive better returns.
We believe ultimately the structure of the industry and playing out how we invest should be driven by that thought process as we think about this. It becomes even more important now as we start to look at behaviors where the line between your business and your personal life starts to change. Working at home or working at the office changes. Fixed and mobile changes. Our belief is that if you're going to get to the right cost structure and the right density, you're going to have to be serving workloads, whether at home or at work, in order to get the right kind of scale and the right kind of structure. This is one reason why I remain as committed as we are to the business markets and ultimately finding our path to success there.
Aggregating the most traffic on the largest, lowest marginal cost, most converged network is going to be critical to our success, and I tell you, just think about where we're going to end up starting in 2020 and finishing this decade in 2029. We started in 2020 with about 18 million fiber passed locations. You're going to hear us today talk about over 50 million passed fiber locations. Just in business alone, sitting where we stand today, we'll increase businesses passed by another 40% before the end of this decade. That's up to 6.5 million business locations that we're going to be able to market and sell to, and we'll talk a little bit about that.
We'll move from about 300 million POPs of 4G LTE in 2020 to over 300 million POPs on 5G, plus most of the rest of the nation will be covered by satellite. By the time we get to 2029, we'll have moved away from proprietary purpose-built network elements to a network that's largely driven by flexible and open software that gives us flexibility to innovate and drive software into the customer experience on a much more dynamic basis. We started this decade in 2020 with a pervasive copper footprint. We will end 2029 exiting copper entirely, and we'll talk about that today. And we'll move from a very fragmented IT structure that was built in the history and the legacy of separate companies to one that is a much more simplified, modernized converged tech stack that gives us the kind of cost structure and flexibility we need.
This will indeed have been a significant decade in changing the industry. So our belief is, at this point in time, we're uniquely positioned to be the operator that has an organic path to do all this, to wind up with a scaled, modernized fiber-based converged network infrastructure that we've got the right kind of owner economics on. Ultimately, what do we think this does? It drives an improved customer experience. It drives better product innovation. It lowers cost structure. And all those things drive better returns for shareholders. And as I've said before, we are absolutely focused that this is a definitive time, and it's a scale time, and this management team will not fall behind in that race. So that gets us to what do we need to do to make that happen?
I hope what you've seen from this management team over the last four years is that when we invest, we can drive the right kind of growth. We're going to invest about $22 billion a year in this business for the next three years. And in return for that investment, we're going to talk to you today about what we can do. You're going to see consistent growth in service revenues. You're going to see consistent growth in adjusted EBITDA. You're going to see consistent growth in free cash flow. You're going to see us accelerate our adjusted EPS growth. And you're going to see us deliver a business as we exit this decade that is incredibly sustainable and durable and built to win.
Because we're arriving at our financial targets for how our balance sheet is structured, as we shared with you in our release this morning, we think we have about $50 billion of financial capacity to decide what we want to do. And because we're at this transition moment, we have an opportunity to make some decisions that we haven't been able to make for the last four years. That strengthened balance sheet is going to allow us to now shift to driving better shareholder returns. We're going to do that by dedicating $40 billion over the next three years, 20 in dividends, 20 in share buyback, keeping about $10 billion on the balance sheet to give us the flexibility to execute that organic plan that I'm talking about.
And the board, with their confidence in this, has already authorized the first tranche of $10 billion that we'll begin to execute in the middle of next year. And Pascal will give you a little bit more color and texture on that when he makes his remarks. We believe this is the right plan right now. We believe it's the right plan competitively. We believe it's the right plan for our shareholders, and we believe it is the plan that we can, in fact, execute. I hope you have the confidence that this management team has demonstrated over the last four years that we can, in fact, execute on an organically driven investment-led approach in this business, and we intend to do that. We're going to position AT&T to be the best connectivity provider in America.
We're absolutely excited about the future of what this means for us and our customers. And I think as you go through the material today, you'll come out in the same place. And with that, I'd like to turn it over to Jeff to get into some of the operations and how we're going to make this happen. Jeff.
Thank you, John. All right, good afternoon. So I'm Jeff McElfresh. I am the Chief Operating Officer that's dropping the top of the water bottle of AT&T. And I've had the privilege of leading the operating teams that over the last several years have delivered very consistent and durable results in both our wireless and our 5G and our fiber businesses. John shared the vision and where we're headed, that we will be the best converged connectivity provider in America, period.
So, I'd like to set the stage for you, what you're going to hear today about how we're going to go achieve that. You're going to hear specifically the five things that we intend to accomplish over this planning period to deliver that vision. It starts off with, first, we expect to extend our converged leadership as we expand our fiber footprint and we grow our mobility businesses nationwide. Jenifer Robertson, who leads our Mass Markets business, will walk us through exactly how we'll do that. Second, and John touched on this, we will position our Business Solutions organization for growth. Business Solutions is the combination of our business wireline segment plus our business mobility product set. And it's a very important part of our long-term plan. We've got an enviable customer base here.
We've got the only dedicated public safety platform and a really strong mobility business that enables us to serve these really important customers with the converged products and services that they need, and we've got a plan to reposition this business for long-term growth as the financial pressures from secular declines and the legacy services ease. We've chosen the right leader to make this transition, Melissa Arnoldi, our newly appointed executive that's leading this segment. It's going to outline the steps that were taken to better serve these important customers. Third, we have confidence that we will continue to lead this industry in defining the path to transition customers from legacy copper services to 5G and fiber services, as well as exiting legacy copper network operations.
And to cover that, Susan Johnson will discuss our roadmap for achieving this across the large majority of our footprint, along with the significant cost savings that result from that. And underpinning these first three is our network, and it's foundational to absolutely everything that we do. Jeremy Legg, our Chief Technology Officer, is going to discuss how we're transforming and modernizing our core networks, as well as streamlining our technology stack so that it can unlock future growth opportunities for us, as well as drive cost efficiencies. And number five, extending our converged broadband network. It's where I plan to spend my time with you this afternoon. I'm going to talk about the fiber and the wireless edge networks. It's where we spend the majority of our capital, and it's what our customers touch each and every day.
And so to be clear, over the next three to five years, AT&T will complete the expansion of our fiber footprint and the modernization of our wireless network, solidifying our position as the nation's largest converged network operator with owner's economics at scale. And there's three really important components to this. The first is our fiber expansion plans and how they're informed by better than expected performance and really attractive economics. Second, our wireless RAN modernization and how it will improve our wireless network performance, but also it unlocks future network monetization opportunities and how it's going to improve our operating costs. And then third, how we're going to execute all these projects, all within the $22 billion capital guidance range that we provided here today over the course of the next three years. So let's get started off with fiber.
As John shared, we intend to expand the reach of our AT&T Fiber to over 50 million plus locations. We're going to exit 2024 with 29 million locations passed with our owned and operated fiber network. Before the end of 2029, we will expand our reach by 15 million locations. Kind of doing the math, that's an average of an incremental 3 million AT&T owned and operated fiber passings that we will deliver annually to achieve a total of 45 million passings by the end of 2029. To get to the additional 5 million plus locations, we intend to expand our Gigapower and commercial open access network relationships to reach more than 5 million locations. That gets us to a total of 50 million plus.
What's really important in this plan is the attractive economics and why we at AT&T have confidence in our ability to continue to generate attractive returns for our share owners with this plan. It's important to think about fiber investment in three buckets. The first is the cost to build, which drives our cost per passing. The second is the cost to connect a customer's home or a business when they actually order service. Then the third is the cost to maintain and repair that network. Starting with the cost to build. Over the last two years, during a heightened inflationary environment, our cost per passing has increased by less than 2% annually, all while improving our build quality. A key reason that we've been successful at bending that cost curve is the way we have chosen to architect and engineer our fiber network.
Most notably, we connectorize the endpoints in this network. This results in higher materials costs, but it yields meaningful gains and efficiencies because we don't have to splice in the field. We just plug in prefabricated connectors. And this plug and play approach, it lowers our labor costs for building fiber. It also results in really a higher quality network architecture, which drives a superior performance and lowers our installation and maintenance costs. Also, given that no one in the United States is building fiber nearly at the scale and pace of AT&T, vendors want to work with us. And as a result of that, we generally are able to get better unit cost economics on our programs. Recent example, you probably noticed, was a deal that we just announced with Corning in October.
That deal enables us to accelerate our fiber network expansion and enhance network performance while minimizing deployment costs. That's the cost to build. Let's talk about the cost to connect. It's the cost to penetrate our build, bring revenues onto the network, and how do we manage those costs? Over the last few years, our frontline employees have upskilled through a rigorous training process that we call Expert Path, and it improves both their customer interaction skills as well as their technical skills in a consistent and a repeatable way, and one of the key benefits of this investment in training is that we have improved our install quality. For example, our same-day completion rates and post-install network reliability have both reached all-time highs.
But more importantly, when our frontline experts connect a home for the very first time, they ensure that that connection is durable and that location is enabled for customer self-install in the future. And why that's important is because any subsequent resident has the ability to perform the fiber installation themselves, which only requires us to send them a gateway. Doesn't require a technician, doesn't require truck roll. Why is this important? This capability is important, number one, well, because customers demand and expect immediate service. As well, number two, it is significantly more efficient. A net add performed via self-install is one -quarter the cost of one without. These install and reliability improvements, they result in fewer overall truck rolls, fewer calls to care, and in turn, drive cost efficiencies while at the same time actually improving the customer experience.
Today, approximately one in four of our fiber net activations are performed by the customers themselves through a self-install, no tech, no truck roll, and as we scale this network and we continue to penetrate this network, the self-install rate will improve to one in three, and as a higher percentage of our net growth comes from self-install, the more capital efficient our program becomes at penetrating fiber, which enables us to reinvest those savings into building more fiber, and lastly, the cost to maintain, while this one isn't necessarily CapEx, it sure is cash. Through these improvements that we've made in the quality of our build, as well as our installations, we've also bent the cost curve on maintenance and repair.
To put the benefits into context, the cost for us to maintain a fiber subscriber is 35% less than copper subscribers due to significantly fewer calls to care and the fiber network being superior in reliability and requiring fewer repairs and truck rolls. So this matters for what reason? Because bending the cost curves within our own fiber program unlocks financial capacity for us to invest more in the fiber footprint expansion and the customer acquisition.
And given our track record, as we expand our footprint, we have confidence that we will continue to drive meaningful returns. In fact, since we accelerated our fiber build back in 2021, the returns on our fiber investment have been better than we expected. 12-month consumer penetration rates are higher, 850 basis points higher. Our ARPUs are higher as customers see the value in our service, and higher speeds are more popular than we expected.
Churn's better, and we're seeing more convergence with mobility attach rates exceeding our expectations by more than 500 basis points. The bottom line is that the results are even better than we had anticipated. Time to revenue is faster. The quality of the revenue is better, and customers are staying with us longer. All in, we have accelerated our organic fiber return business case by two years. So I'm sure you would agree that these attractive fiber economics are a key reason why we are expanding our fiber footprint to roughly 45 million locations and why we are confident that we can continue to sustain our performance and deliver attractive returns. For the remaining 5 million plus, we're going to approach that with the relationships with Gigapower and our commercial open access providers.
Today, in that footprint, we're actually achieving equal to or better coverage performance than we are in our own in-footprint territory. And given that performance, we have confidence we can continue that as we expand the reach of these relationships and address this footprint with the right economics and in a capital-light manner. So that is our fiber expansion plan. Now let me transition to where we are with modernizing our wireless network. We first announced these plans to modernize our wireless RAN or our radio access network roughly a year ago. And this effort is important for three reasons. Kind of the first reason is through this transformation, we're going to modernize nearly every one of the AT&T cell sites. It allows us to boost capacity, to increase coverage and speed, but also it helps us optimize our cost structure.
As part of this program, by the end of 2026, we will efficiently deploy our spectrum assets to cover more than 300 million POPs with our midband spectrum, up from nearly 270 million by the end of this year, and beyond the performance improvements, we also expect to drive significant efficiencies in recurring network operating costs, including energy savings, fewer tower configurations, and turf vendor consolidations. The second reason that this is important is that this transformation of modernizing our wireless RAN is actually to transform it to a more open architecture that will enable us to serve the growing traffic demands with a better customer experience and at a better cost curve. By the end of 2026, we expect that 70% of our 5G traffic will flow across open hardware.
And this transition to a more open radio architecture will enable and invite new disruptive technology partners to engineer wireless access endpoints for our network at a much better cost curve. Why is this important? Well, as John mentioned, based on our core beliefs that demand will continue to rise, the evolution of that customer bandwidth, as well as our core belief that all networks over time will be fiber networks with different access technologies at the edge, when you couple the nation's largest fiber network and one that is expanding with our open-enabled wireless network, you can see how AT&T is uniquely positioned to serve the growing demand across our infrastructure at favorable economics. This will become increasingly important as the industry becomes more spectrum constrained and begins its transition to the next generation of wireless technology.
Bottom line, our wireless modernization coupled with our fiber expansion will allow us to deploy our spectrum and densify our network in a much more cost-efficient way. And the third reason that this program is so important is that with our modern network, we will unlock a consistent network API architecture. This will enable more rapid product deployment and open up the network for future network monetization opportunities. I'd like to pull all of this together for you so that you can see the puts and the takes of our capital plan. First, we expect to continue bending the cost curves for fiber build, installation, and maintenance, just as we have been doing over the past several years.
Second, the capital investment related to our wireless modernization project is expected to peak next year, and it will start to decline in 2026, and there'll be a further decrease in 2027 as we near the completion of our modernization efforts. Then third, as Pascal shared during our third quarter earnings call, the high-end of our $21 billion-$22 billion capital investment guidance this year includes payments for vendor financing. We've significantly reduced our vendor financing balance, and we expect to remain within this lower, more stable range going forward. With these three efficiencies coming into our capital plan, this gives us the financial capacity beginning in 2025 to increase our annual fiber CapEx over the next three years to expand our footprint while staying within our $22 billion annual capital investment target.
This will further position us to continue to generate attractive returns on our fiber investments while also seeing an improvement in our overall capital intensity over time. So with our results over the last four years and our plan going forward, we at AT&T have confidence that our customer-centric and efficient operating model, along with our significant network investments, will position AT&T to drive further growth in 5G and fiber customers, and convergence will only deepen and increase the value of those relationships. Thanks for letting me set the stage for you here this morning. Let me turn this over now to Jenifer Robertson, our Mass Markets leader, to talk about how we're going to extend our converged customer leadership. Jenifer?
Thank you. Thank you, Jeff. All right, I am excited to be with all of you today to share our Mass Markets plan for the next three years, and I'll give you a little bit of a spoiler alert. The spoiler is this: it's going to sound a lot like the plan we've had for the last four years. In that, specifically, it is centered around our customers. Over the last four years, we've won by listening to our customers and responding to their needs, and what you'll see us do going forward is focus on three things.
First, you've heard it so far, we will absolutely grow with fiber. We have the best broadband product in the market with AT&T Fiber. We know that. Our customers tell us that, and I'll share a little bit more detail with you here in just a minute. We are going to win with convergence. John shared some perspective with you just a minute ago. Customers prefer converged. We are able to deliver that simple and seamless, and we know we can continue to evolve that to make it better, and then our third priority here is to sustain our momentum in mobility nationwide. We have a good set of results over the last four years.
We'll keep that momentum. We know we have opportunity to continue to grow, so we like where we're positioned. We're going to start with fiber first. Let's dive in. Okay, we've shared a few results. Fiber as a product, it is absolutely the most reliable internet speed in the nation, AT&T Fiber. We have symmetrical download and upload speeds, and we have the capacity to efficiently scale. Jeff just covered that, but our superior performance with the product is showing up in our financial results and in our customers.
I'll breeze through these. You've seen them. About 20% ARPU growth over the last three to four years. We've doubled our fiber subscriber base. And on an annual basis, we've grown our revenues about 25%. Those stats you've seen; they're business outcomes. What we haven't yet shared is why we feel so strongly, why we have conviction about what our customers think on AT&T Fiber. Our customers tell us this: they're willing to recommend AT&T Fiber above other products in the market at a greater rate. In fact, over the last three years, AT&T Fiber has led the industry in Net Promoter Score of major scaled fiber providers out there. And I'll say major scaled broadband providers out there. AT&T Fiber wins, hands down. Above that, AT&T Fiber customers stay longer. They stay with us.
We've seen, excuse me, we've seen about a 20 basis points improvement over this period of time in our churn. So they are staying with AT&T longer than we see with other products. So they're telling other customers that they are willing to recommend it better than any other broadband product out there. They're staying with us longer, and we are winning external awards to boot. We're seeing results. Customers are telling us that we are better for a product. Where do we go from here? We grow. We have more room to grow with AT&T Fiber. In fact, we are outperforming our penetration targets and our newer build cohorts, and we're doing that sooner than expected, so we're already achieving what we had set for ourselves internally as an aspirational goal.
Initially, we had said, "Hey, we'd like to achieve 50% penetration in these cohorts, in these areas that we build fiber. Long term, let's shoot for 50." What I'd tell you is in the newer build, we are already achieving 50% in some of our markets. And we have vision and line of sight to achieving 50% Fiber penetration in the long run in all markets where we have AT&T Fiber. That is our new goal. At that, with the expanded build we have, we have line of sight to doubling our AT&T Fiber base across the 50-plus million Fiber footprint. We've doubled it over the last three to four years, line of sight to doubling it on the 50-plus million Fiber locations. All right. We know that where we have fiber, we win. And our goal, second area of focus, is winning with convergence.
We have 9 million fiber subscribers today. Four out of 10 of those we've shared also have wireless with us, and we've grown that converged penetration rate about 400 basis points over the last three years. In that, we know that our share of postpaid phone subscribers has increased about 500 basis points. It's higher with those cohorts than it is in areas where we don't have our broadband. Okay, so we have share of postpaid phone higher than our national average. Why is that? Why do customers in a converged area have higher wireless footprint? Well, we've already said 77% of customers want converged. But do they come in and are they happier with it? The answer is yes. Again, what are customers telling us? Customers actually that are converged with AT&T Fiber and AT&T Wireless have 2x higher Net Promoter Score.
They're more willing to recommend AT&T Fiber and AT&T Wireless. They also have better churn. They stay with us longer, just like an AT&T Fiber customer, but in this case, it works both ways. So when a converged customer has wireless with us in our fiber footprint, the postpaid phone churn is about 25 basis points lower than in areas where we don't offer broadband. We get double benefit because when that converged customer is on the fiber side, our AT&T Fiber churn actually sees a 45 basis point improvement versus a standalone AT&T Fiber customer. Now, we've had a lot of speculation out there. Is converged just about a discount? Are we out there going and saying, "Let's just grab the customers and anchoring the home," and is there true value? What I'll tell you is the answer is yes.
Not only are customers happier with us recommending it and staying longer, but converged customers at AT&T are spending more. In fact, converged customers are actually more likely to have more lines per account, select higher speed tiers, and that results in a 15% or more increase in lifetime value, so it's a more valuable customer base, and we have room to unlock new benefits as we look at evolving the customer experience and adding on to products. How do we do that? We are looking at evolving experiences and building new products. A couple of very recent examples and soon-to-be-launched-to-market examples. We have streamlining our digital experiences for new wireless and new fiber customers coming to the market in a new way to purchase.
We have our recently launched Internet Backup, where we have the ability for networks to actually hand off to each other. AT&T Business coming up next launched this in 2018. We've recently launched this on the Mass Markets side. And what we're seeing there is that offering this to customers at no additional cost, any customer that has an unlimited plan and AT&T Fiber is able to carry the benefit. And what customers value most in this space is reliability of connectivity. So we're giving that to customers. You'll see that in certain locations today and more to come. So the opportunity to grow higher value customer willing to recommend, happier and staying longer, and the ability to unlock new benefits as we go towards converged gives us a lot of confidence that we have runway to grow in this space. And our question is, how much runway?
Very similar to fiber, we know that we are seeing success as we expand our footprint in our organic areas. We've mentioned Gigapower. We're seeing early success in Gigapower as well. So our penetration rates are ahead of the business case in our non-traditional footprints outside with Gigapower. And in that area, we're seeing higher adoption of speed tiers, right? We are seeing higher adoption of converged. So we have proven to ourselves the investment thesis of, can the AT&T brand go out of its traditional wireline territory and hunt with wireless and fiber? That gives us confidence to keep growing there. As we expand our footprint with fiber, we know our performance is strong and the potential is clear there. So in converged, we see a path to 50% converged penetration over the long term and more than double our converged subscribers.
That's our vision for fiber and for converged growth. It leaves for us the mobility space where our goal there is to sustain momentum nationwide. As we do that, let's talk about a little bit of where we've come from. John mentioned it earlier, 10 million postpaid phone subscriber growth at $9 billion of service revenue growth, and we did it the right way. We started with the customer. We stayed disciplined, and we focused on quality. It's a customer-centric approach that we will not steer away from. In the fall of 2020, as we looked across, we did one thing, and we found the biggest customer pain point out there, and that was that our existing customers didn't feel as valued as new. That launched into our Best Deals for Everyone offer, but our focus on the customer was much broader than just an offer.
It actually spanned offers, distribution, pricing, and product. We've iterated on that and launched activities across all four of those streams ever since. It's what's contributed to the growth that we've covered so far this afternoon. We will repeat those steps as we look forward. Those steps led to 13 of 15 quarters of industry-leading churn in the postpaid phone space. That success has also given us the space to be able to lean in and focus on where we underperform versus the industry in specific audiences, in specific geographies, in spaces like converged. So as we look forward, we'll stay focused on treating existing customers the same as new. We'll invest in modernizing our network, as Jeff just covered, and we know we have opportunity to focus on those under-penetrated areas that we've discussed previously: converged, price-sensitive, Hispanic audiences, and the business segment that Melissa will talk about.
We also know we can innovate. That's a space that we have not leaned into as much, and there are products and features that we know customers value that we've recently started to bring to market. Some of those include ActiveArmor, a security-based feature that resonates with customers, protects data. It offers safe browsing. It protects you with public Wi-Fi. We launched AT&T Turbo earlier this year. Turbo was a first of its kind. It offers real-time data, safe browsing, responsiveness, stability. It allows gameplay as well as certain video aspects to it. Above all, it offers choice, and it offers purchase capabilities in a two-step digital buy flow that provides an opportunity for customers to see the product in context at the moment they need the service most. That's the new type of engagement we'd like to see with customers, and you'll see us lean into more.
All right, as we head into 2025, the bottom line is we'll keep building on what's working. We've proven we can execute here. We have four strong years of growth behind us, but we're not in a leading position just yet. There's no structural reason for that. We absolutely have more miles of coverage than anybody else. We have the spectrum position that's on par with the share leader out there, and we have a winning strategy that focuses on the customer. When we put those three things together with the fact that we will continue to build on the plays we've seen work for us, we have room to grow in mobility nationwide while leaning into winning in fiber and winning in convergence. So we're really excited about where we land. I'm going to wrap up where I started.
AT&T Fiber is the absolute best broadband product in the market, and we are proud to be the ones that have it and proud to be the ones that continue to evolve with it. Customers prefer convergence. They tell us that. They spend more with us, and we will make that product better together. And then lastly, we'll stick with consistency, focusing on customers all the way through and sustaining our momentum and mobility nationwide. I'll say thank you to all the AT&T customers out there in the audience and those watching live. You're the reason why we're here, and we're excited about what we do. And with that, I'll turn it over to Melissa Arnoldi to talk about AT&T Business.
Okay, good afternoon. First and foremost, I am excited to be here, and I'm excited to meet you all. I'm going to discuss business solutions. As Jeff and John said earlier, this is a combination of our business wireline and our wireless relationships that we have with our coveted business customers. As you've consistently heard throughout the day, AT&T is investing in the right assets to solidify our position to be the nation's leading connectivity provider. We lead with quality and asset differentiation, and that could not be more important and more true for the business solutions space.
As John talked about earlier, we understand. I know the importance of repositioning AT&T Business Solutions, and I really look forward to the opportunity. There are three things that I'm going to talk to you about today. The first one, I'll cover the state of AT&T Business Solutions as it is today. The second, I want to talk about our roadmap, specifically how our network leadership is going to fuel our product and distribution strategy to drive growth. And third, I'm going to give you a look under the hood at some numbers that we have not shared with you before. To start, here are some important business facts around AT&T Business Solutions.
You may know this, but we serve nearly 2.5 million of the largest global enterprises, government, small, and medium businesses. We provide services to nearly all of the Fortune 1000 companies. And we do that because we have deep relationships in key verticals. Those verticals are manufacturing, finance, healthcare, retail, and public sector, just to name a few. Plus, we have well over a couple of million relationships with small and medium businesses. And we are the number one fiber share provider in our 21-state footprint, number one.
We are number two in our mobility subscriber. That's nationwide. We serve over 6.4 million connections on our FirstNet network. And we do that by providing first responders with mission-critical services to more than 29,000 public safety agencies. Okay, so the question you're probably asking, that's great. So why aren't we seeing these strengths in our financial results? That's the right question. That large customer base that I just talked to you about that's so coveted and desirable, they also are the ones that rely on us for legacy services. You know how I said we were number one in fiber and our footprint? We are also number one in several legacy and transitional services like virtual private networking, like MPLS, and local voice and data services. So while our customers rely on legacy infrastructure, our customers are also midstream in their transformation from a technology perspective.
And they're going to move to our fiber and our wireless access. We also, though, are in midstream, simplifying our back-end technology stack. Why is that important? That helps us provide agile, innovative products in a timely manner. We've been going through that process, and you're going to hear from Jeremy Legg later, and he'll talk to you more about that. So as I've said, we have experienced headwinds in business wireline, primarily from the industry undergoing a secular shift. But I'm here to tell you that we are positioned in business solutions to succeed that shift and extend our growth through our leadership in fiber and wireless. Our fiber and wireless-owned assets and advanced technology give us the right to win in every segment of the business. And that's what we plan to do. And also think about our brand reputation. What matters to our customers?
I can tell you in talking with them, it's reliability, it's high performance, and it's secure connectivity. We do all of that, so let's talk about how we're going to reestablish ourselves and grow our business in business solutions. Here's our playbook. We're going to start with 5G. As Jeff said, we have a desirable customer base, and it's a strong mobility business. In fact, in business solutions, we have achieved 23 consecutive quarters of wireless service revenue growth, and our business wireless revenue growth rate has exceeded our total wireless service growth rate every quarter since 2019. And it's business grade. What does that mean? That means it allows us to monetize solutions such as network traffic prioritization for important business applications, as well as video and voice. We also deliver the only dedicated public safety platform. It's called FirstNet.
It's a unique network responsive to the public safety needs. Supporting those first responders, that's in our DNA. The strength of our wireless business is a testament to the 5G network assets and the deep relationships that our account teams work hard at each and every day with their customers. And we will leverage those relationships to cross-sell fiber. Which brings me to fiber. As Jen said, where we have fiber, we win. And as you've heard from Jeff earlier, we are exceeding our consumer fiber business case expectations. But guess what? We have room in our business segment. I'll talk to you about that in a minute. Today, our penetration rate, where we have owned and operated fiber at a business location, is below 30% overall. We have room to grow. We can do better.
No reason why we cannot penetrate more, and I'll talk to you more about that. In fact, if you think about it, close to 70% of the business locations where we have fiber, they are also occupied by small businesses. And small businesses, we even have more upside with a lower penetration rate. That is a significant opportunity for us. How are we going to do that? Two things. We're going to retool our distribution as well as our digital capabilities. Let's talk about distribution first. We're great from a direct sales perspective. We've got deep relationships. We always have room to improve, but now what we're going to do is we're going to expand our indirect presence. We're going to take new routes to market with a robust third-party partnership. And through this agent model, we will win in small and medium businesses.
This will help reposition our shift in distribution from a direct one, largely today, one that's equally capable through indirect over time. Digital. We're making investments in our platforms to continue to build momentum through our digital channels. We're confident that the simplification of our platform and the development of a converged experience for our customers will unlock sales growth. And as we develop new product innovations, we're going to develop them with a digital-first mindset. Two examples of proof points I'll talk to you about. The first one is our fixed wireless product, AT&T Internet Air for Business. The other one is a network-embedded security service, Dynamic Defense. Both of those were natively digital-enabled. Okay, next, let's talk about AT&T Internet Air. Our strategy is we are going to lead with fiber.
However, if we do not have fiber, we have a nationwide fixed wireless solution, and it is built for business. So what do I mean by that? It has no speed caps. It can serve as primary access for small businesses and as primary or backup network for our fiber service for medium or larger businesses. That brings me to our fourth play, convergence. Where we sell fiber and fixed wireless connectivity, we must converge our customers. We have to attach wireless. I mentioned a few minutes ago that our direct sales channels are now cross-selling fiber and wireless. As a result, we aim to increase our distribution coverage for direct sales by 20%.
However, the converged customer centricity, it doesn't stop at just cross-selling. It extends into our product innovation, such as with our new integrated gateway. It's a first of a kind in the market, the first. It brings us.
I would take it to say connected firefighters or connected police officers. The future of technology is in things like wearables. It's in things like helmet technology that is going to provide real-time data feeds to incident commanders to make better decisions. That connection, whether it's 5G, 4G, and even as we move into satellite technology, that whole Internet of Things, and what is that going to look like? I believe we're going to be even more reliant on connection than we are today because of the advancements in technology that are coming our way, that are completely a good thing for us, for our firefighter safety, and their health, and our response to the community. So absolutely almost a mandate in terms of criticality that we have those connections.
Fabulous. And one of the things that is so unique, I think, to your branch of service is that firefighters are not just permanent full-time, but represent a lot of volunteer firefighters. And what's the importance of ensuring that every public safety community, even though they may not be badged in full-time, has access to the capabilities of FirstNet?
Well, I'll jokingly say, and we talk about this a lot, fires don't burn differently in rural communities than they do in my own community, which is an urban area. So the challenges that maybe a rural community fire department faces, the only difference is scale and scope, right? I have a lot more employees. They have a lot less. But the need for connection, regardless of community size, is limited. I think there's a lot of opportunities as we continue to build out to better serve rural communities and those volunteer firefighters. And at the same time, we'll be benefiting folks like myself and my department as we continue to build out as well.
Thank you very much.
Thank you, Jeff. And again, thank you to FirstNet. Thank you to AT&T for your ongoing support. It means a lot to us. And I go back to that partnership and that trust, and we'll continue that.
Wonderful.
All right. Good afternoon. Hello. I am Susan Johnson, and I'm here to talk to you today about copper. Or more specifically, I want to talk about how AT&T is leading the transition to upgrade our customers from our legacy services to next-generation services. You've all heard a lot today about our next-generation services, fiber and wireless.
And it probably does not surprise any of you that our legacy services are no longer meeting our customers' needs for speed, reliability, and always-on connectivity. NPS scores for fiber are five times higher than copper broadband. And customers are voting with their feet and moving off of the 70-year-old copper services. In fact, only 5% of our residential customers are still using copper voice technology. Right now, we are operating two parallel wireline networks, our copper network and our modern fiber network. And our wireline footprint spans over 500,000 sq mi and requires lots of people, power, and infrastructure to manage and maintain. And the copper network is very inefficient. We are seeing declining reliability with storms and increased copper theft. Copper simply does not do well with water and flooding. And repairs are very labor-intensive. And unlike fiber, copper is an energy hog. It's called an active technology.
That means every line and every element is powered all of the time, and the IT stack is a morass that has been built around multiple pieces of hardware with custom back-end systems. This is a call to action. This is a descaling business with high fixed costs. We've been successful in removing some of the variable costs as customers have left, but we are now ready to take the next step. The bottom line is we expect to no longer provide copper-based services across the vast majority of our footprint by the end of 2029, so with that backdrop, there are three things that I want to talk to you about today. First, I want to give you an update on our leadership in developing and launching some products that can deliver a better experience for our customers.
Next, I'm going to cover the strategy on how we are orchestrating our copper exit, and then third, I'm going to lay out the impact and what all this means to AT&T, so you've heard today, customers are demanding faster and more reliable services, and we've developed many products that can better meet their needs. AT&T Fiber, you've heard a lot about it today, rated number one in customer satisfaction for residential Internet service. AT&T Internet Air provides a much better experience for businesses and consumers compared to our copper broadband today. In fact, up to 25%, about 25 times faster speeds than legacy ADSL, and then AT&T Phone Advanced. This is a new product that you've heard less about. It is a new innovative product.
It works just like a traditional landline phone, and it works with many of the services that our POTS customers, our landline customers, are used to using. Think fax machines, medical monitoring devices, alarm systems, even elevators. It connects over our wireless network or a broadband connection. And we've gotten great feedback from our customers who have made the switch. Now, as you know, AT&T has the largest wireless network in North America, and our AT&T Phone Advanced product is going to serve the vast majority of our existing copper-based customers. However, there are going to be some areas where we will need to work with our customers to move them to other technologies, including satellite. But we've made a pledge that we're going to keep our customers connected through the process and make sure that no customer loses access to voice or 911 services. Last, legacy emulation.
Some large enterprise customers may not be ready to upgrade their legacy equipment. This product enables them to maintain all of their existing equipment on their prem while we upgrade the network. We've had some really good migration success already as we've been migrating customers to these products. This means we've got the right message, the right price point, and the right channel to reach these customers. And I'm happy to say it's working really well. Let me give you some data points. Where we have fiber today, we've had strong migration success rates, about 85% for consumer broadband customers and about 60% for our business broadband customers. Now, before I move into how we are going to orchestrate this exit, I want to ground you in some numbers. We have roughly 88 million customer locations in a wireline footprint today. This is a number you may not recognize.
21 million of these locations are voice only. This leaves about 67 million locations that are broadband eligible. Of these, we have 29 million locations that will have fiber this year, growing to 45 by the end of 2029, as you heard Jeff lay out. The remaining locations are going to be served with wireless solutions, and I will tell you more about that in just a minute. Now, our copper network also includes 4,600 wire centers. For those of you that are less familiar with legacy architecture, we define a wire center as the legacy equipment that's in that central office, think voice switch, plus all of the outside plant, remote terminals, pedestals, and copper lines that all connect back into that central office. Now that I've grounded you in some of the numbers, you might remember that a couple of years ago, we made a commitment.
We made a commitment that we would reduce the square mileage where we offer regulated legacy services, think POTS and special access, by 50% by 2025. I am proud to tell you that we have already achieved this milestone ahead of plan. This represents about 250,000 sq mi where we've met the regulatory requirements to no longer offer regulated services because our customers have moved on to other services. This is really good progress against our goals and has enabled some good customer-related cost reduction. However, without the full discontinuance of services across an entire wire center geography, we're unable to stop the maintenance, repair, and attack the more fixed infrastructure costs. So with that in mind, our exit strategy is going to be geographically focused at a wire center level. As John explained earlier, our wireline footprint is going to look very different in 2029.
Let me tell you how we're going to do this. Our path to migrate customers and exit copper is going to be carefully orchestrated based on two things: the profitability of the wire center and the plan for our fiber build. Wireless First is the name for our wire center areas where we have not built and do not plan to build residential fiber. There's not an economic path to do so. These wire centers may still have fiber supporting businesses or cell sites, but no consumer fiber. This is about 50% of our land area, but it's only 10% of the population. Pretty low density in these areas with four remaining copper customers per sq mi. In these areas, we will be offering customers a better experience with our mobility services, think fixed wireless, including fixed wireless and AT&T Phone Advanced.
We'll rely on competitive alternatives and satellite connectivity to ensure that all customers stay connected. The plan is to have no customers using copper services in these wire center areas by the end of 2027. Now moving on to Fiber First, which is the other 50% of our land area, but it's 90% of the population. These are the areas where we have today or plan to build fiber. In these fiber areas are much more dense. We've got density about 10 times more dense than the wireless first areas I just talked about. In these areas, we will be working to upgrade our customers to fiber. These migrations will be ongoing over the next five years. I want to be clear, not every customer location will be reached with fiber in these areas.
We will still serve some of the customers in these areas with wireless alternatives. Our plan is to have no customer left using copper services in these wire center areas by the end of 2029. That is when we expect to have our fiber build to 45 million locations completed. Now, I want to make two important notes. California is not included in the plans I just laid out for you. We are continuing to work with policymakers to define our path in that state. We will work very closely with Melissa's team to ensure that we honor the long-term contracts we have with our enterprise customers, particularly in the public segment. Now, as many of you know, the services that we deliver over this copper network are heavily regulated. We've been working at the state and the federal level to update the regulatory frameworks.
For example, at the state level, we've worked to modernize carrier of last resort regulations or COLR regulations. I am happy to say that 20 out of the 21 states where we operate wireline networks, we've achieved it to modernize these regulations. While California is the last state to modernize, we've started a process there, and we will continue to work towards this objective. Now, while the work has been going on with the states, we've also been working very closely with the FCC. We've achieved two great milestones. First, earlier this year, we gained FCC approval to stop all new copper-based inward orders in 60 wire centers across 13 states. Now, this is a small geography, but an absolutely critical precedent for us to be able to stop all copper inward orders. Second, as I talked earlier, we have launched a product called AT&T Phone Advanced.
This product was specifically designed to meet the FCC's criteria as an adequate replacement product for our traditional landline phone service, and we have successfully completed the testing with the FCC, and we are continuing to move through their preview process. We have an application pending with the FCC for a small number of wire centers, which, if approved, would allow us to replace traditional landline phone service, think POTS, our most regulated product with AT&T Phone Advanced. This is a critical step, and we are leading the industry to define our path to exit copper and eliminate legacy costs, so with the path now set and the precedent laid out, we are going to work with the FCC to speed up and scale this process.
And with the new administration, we are optimistic that we can make even more progress in simplifying our networks and migrating our customers over the next several years. Okay, so I've said a lot. What does all this mean financially? Overall, our legacy business is profitable today, but the revenue declines are accelerating. With the migration rates that we are seeing, we are confident that we will retain many of these customers as we migrate them to fiber and wireless. The job of my team is to optimize these rates and to drive discipline into the operations to ensure that we remove the legacy costs. We currently spend about $6 billion annually on direct costs to keep this network running, and there's about two broad cost categories. First, about 40% of the costs are direct customer costs, more variable in nature. Think installation and customer care.
These will fall as the customers migrate off the copper services. The second area, the geographic focus costs, which is 60% of the cost, they're more fixed in nature. Think power, network maintenance, and real estate costs. By targeting the complete customer transition in a wire center with the least profitable wire centers first, we are able to remove these geographic costs and really optimize margins as we move towards exiting copper services. Now, I will note that this $6 billion that you see doesn't include any of the capital expenditures associated with the business or any of the legacy IT costs. This legacy IT stack, as I said earlier, presents a large opportunity. Jeremy is going to talk more about that next. Our plan defines cost savings to be achieved over the next five years as we make progress.
This is all included in the guidance that Pascal will review. We are also looking for ways to monetize the copper infrastructure as we exit. First, the reselling of copper as we decommission it, and we've been very successful in that. And second, repurposing and leasing out some of our central offices, many of which are in very desirable real estate locations. So modernizing our network and exiting copper substantially simplifies our wireline footprint. We're going to need less people, less power, less infrastructure, and we'll be able to target the $6 billion of costs tied to legacy services. And with our proven success to upgrade customers and the regulatory precedents that we have achieved, we are very confident that we can provide the right products for our customers to migrate them, maintain them, and exit the copper network.
All of this is letting us control our own destiny to geographically exit copper and preserve our margins. We are leading the industry and very confident in our ability to exit the legacy footprint by 2029. Thank you for your time. I'm going to now turn things over to Mr. Jeremy Legg.
Give me a bump. Thank you. How's everybody doing? All right, I'm all that stands between you and a break. They put the tech guy there. Don't often speak to a large room of financial analysts, but it's a pleasure to be here. My name is Jeremy Legg. I have the privilege of being AT&T's Chief Technology Officer and probably more importantly, represent an incredible group of men and women that bring together all of these different products and services that we talked about. And you've heard obviously a lot of information here. You've heard Jen and Melissa talk about new products and services and how we're going to be acquiring new customers as well as maintaining those customer relationships. You heard from John and Jeff about the expansion of endpoints and establishing connection points for our customers as close to where they are as possible.
I'm going to take you through a bit of information around how we're actually enabling all of those capabilities across the technology stack, specifically looking in three areas: network simplification, application consolidation, and generative AI. And within each of the three of these areas, really talking about both the benefits to the company as well as the competency that we're trying to develop here. This isn't technology for technologies sake. This is technology that we're using for some very specific outcomes. Let's jump into the network simplification piece. And I'm going to start on the wireline network, some of which you just heard a bit about from Susan. And this also incorporates our broader fiber network. We have actually been in the business of establishing software-defined networking for quite some time. Many of you have probably heard that term. We started that several years ago in our core fiber backbone.
That's the big ring that goes around the United States that delivers bits to the different regions of the country. We actually started that three years ago, and we now have 80% of the core traffic on that backbone riding through software-defined networking. And what that means is it's standard compute, Linux boxes, and open-source systems that are routing that traffic around the country. What we are now in the midst of doing is moving that same set of technology closer to the edge. So as Jeff talked about establishing all of those endpoints and those consumer and customer endpoints, this is just the other side of the last mile. This is what's actually sitting in those central offices that delivers those products and services to those endpoints. Today, inside of many of those central offices, we have 4,600 of them sprinkled around the country.
There's a very heterogeneous stack of technologies. You've got the legacy copper stuff that Susan just spoke to you about. There are also modern technologies in there, such as fiber and wireless. There are enterprise customer technologies that exist inside of those. The key takeaway here is that each of those technology stacks are different, and each of them persists inside of each of these different central offices. What we're moving towards is an intelligent software converged network. So as you heard about convergence from our market-facing business units, convergence in the technology world means this. And so what that means is that we can take those same things we did in our core backbone where 80% of our traffic is moving, put that in the edge, and collapse these technology stacks onto one stack. And you don't have to have all of them sitting in every central office either.
So these technologies historically were proprietary. They were very labor-intensive. And if you wanted to build a new product or a service, you had to send someone to go install things in the network. The competency that we're developing here is the ability to deliver software-based, high-margin, lower capital cost networking down to those endpoints, as well as create the opportunity to develop new products and services that sit on top of this. Melissa spoke of one earlier today in Dynamic Defense. That is actually a software-based product that sits on top of these Linux-based routers through traditional software development capabilities that you've seen traditionally outside of the telecommunications industry. From a financial standpoint, the benefit is obvious. It's right on the slide. Consumption is continuing to increase. You saw that in what John was presenting earlier today. You've got consumption rates going up 20%-30% a year.
By doing this, it actually drives our per-unit cost way down. So at the completion of this journey, we'll have 70% lower per-unit cost associated with the transport of the traffic, even though in the aggregate, consumption is increasing markedly. Similar to the wireline network, we're also making some huge changes to the wireless network. Jeff spoke about what's happening in O-RAN and the touching of every cell tower that exists across the United States. This is further back in the network where our mobility cores are. And a mobility core is effectively the kind of the brains that sit inside of a mobile network that determines how a call is routed. Historically, just as we had in wireline, we had separate mobility cores for lots of different products and services: Internet of Things, enterprise customers, voice, data, 4G, et cetera.
We're in the process of now actually collapsing the majority of those cores onto a single standalone core that is a cloud-native core. This is built for the purpose of being distributed around the edges of the country. And it's also purpose-built so that you can do the same kinds of software development that we're doing in our wireline network on top of our wireless network. So whether you're talking about the ability to ultimately do things for the automobile industry and dedicate certain portions of spectrum, or if you're talking about the ability to deliver network services like AT&T Turbo that Jenifer Robertson talked about earlier, it's this kind of technology that enables that across our mobility footprint. And we already have millions of customers already on this core. So both of these things, they aren't things we plan to do.
They're things we're already well underway in doing today. All right, I think Susan hit me with the IT stack is a morass that hurt me on the inside. Here's how we're going to fix that problem. So two things before I get into the IT stack, let me just sort of ground people in what's going on relative to customer expectations. You all already know this, but I'm going to say it anyway. If you're an enterprise or a business customer, the public cloud was very good at enabling you to pull out a credit card, spin up compute, spin up storage, leverage a service in the cloud, and you could do it from a laptop while you were literally sitting right here. If you think on the consumer side of the world and what the big e-commerce giants have done, it's effectively the same thing.
You can almost instantaneously buy and purchase and instantiate a video service or purchase a physical good that shows up at your house the next day. Those are the sets of consumer expectations that we have to be able to meet as we change the systems that face our customers, whether they sit on the business side or they sit on the consumer side. A few things that I want to point out here. The first is you got to know who your customers are. We are very well down the path of actually pulling CRM capabilities across all of these different touch points so that no matter where a customer enters into the AT&T ecosystem, we actually know who they are and what their preferences are.
Second thing is actually enabling digital buy flows to buy the products and services seamlessly through a shopping cart, just as what Jen and Melissa spoke to you earlier about. Just over the course of the next few days, we'll actually be enabling the convergence that you just talked about by allowing a new customer to purchase fiber as well as wireless through a single shopping cart and instantiate that digitally. All of this has to be presented to a customer on a single bill. Simple thing and seemingly. But as what Susan talked about, many of these products and services historically sat on separate billing platforms. We have already put our Internet Air product onto our new billing platform. And over the course of the next couple of years, we'll be moving our other products and services onto this same platform.
The lower portion of this is something that probably doesn't get talked about nearly as much. This is the technology platform that sits on top of the network. So there are things called APIs, Application Programming Interfaces. They're simple things seemingly, but basically they're things that allow developers or other folks to plug into a system or plug into a network easily. One of the things that we've already done, whether you're an MVNO or you were a branded fiber partnership or Gigapower, is enable people to plug into our back office capabilities as well as into our network so that it takes weeks, not months, to do so. In the scaling that Jeff was talking about, that already exists. Gigapower and those different partners that we have are already leveraging this technology. What it also enables us to do is create reusable sets of capabilities.
AT&T Turbo is a packet acceleration capability, effectively. It's a prioritization capability. You can decide to use that for lots of things. And that capability already sits there. You can rebuild on top of the network because that's already there. That speeds your product development. And as I mentioned before, it enables us to actually inject that software into the network. We don't have to send a truck roll. We don't have to put a new piece of network or a new piece of gear into the network. We can instantiate it digitally. This is a big deal for our company, and we think it's something that is actually a big deal for the broader telecommunications industry. AI, who's not talking about AI, particularly generative AI? I want to point out a few things that some of you may not know.
AT&T has been in the AI business since the AI business was created. I have the privilege of representing AT&T Labs inside of the company. They actually were at the forefront of this way back in the 1950s. This isn't a new thing for our company. If you fast forward to today, we're actually the sixth largest holder of AI patents in the United States. We have over 600 AI and ML models already in production inside of AT&T. There are areas, the IT stack being an example, where I think we aspire to be world-class. This is an area where I believe we already are world-class. What I would point out to you here is the item here called the BIRD GenAI SQL Award. That is actually an award about being a practitioner of AI.
That is about embedding data inside of AI, text, and retrieving that information back in the most accurate ways possible. We are presently number one in the world in that award. And so I'm very happy to say that this is an area where I think our company truly excels. Now, the journey with AI is. It's the latest in a disruptive set of technology trends. By my guesstimation, these happen every 10, 15, 20 years. I'm now at a vintage where I've lived through a couple of them myself. And in the 1970s, it was CPU computing. In the 1990s, it was the emergence of the internet. In the 2000s, it was the iPhone. And I would argue in the 2010s, it was really the emergence of public cloud computing and the impact that that had across the technology landscape.
GenAI is rolling out faster than all of these technologies rolled out before. I do believe it's going to have a more profound effect not only upon our company, but also our industry. What underlies GenAI are large language models. I'm sure a lot of you have heard of them: Gemini, OpenAI, Meta's, Llama. NVIDIA obviously powers the chips that provide the chips that power these large language models. AT&T isn't in the business of creating chips, obviously. We could create large language models, but our preference is to actually leverage the billions of dollars of investment that are going into those large language models and those chips and to become the best practitioner of artificial intelligence in our industry. That we believe is the differentiation point associated with this. Which LLM do you use? What is the right question to ask an LLM?
Because they're not all good at the exact same thing. I look at the, and I think our company looks at generative AI in a somewhat unique way. These models have existed for quite some time, and they were trained on the corpus of information that sits on the internet. Essentially, everything that has been published on the internet has gone into these large language models. That's an incredible thing. It has democratized incredible amounts of information across many, many industries. But what I would also argue that it's done is it has commoditized that information so that everyone has the exact same set of access. Our information is not on the internet. Our network data is not on the internet. Our customer information data is not on the internet. What's happening with our churn rates, what's happening with all of the data sets that we have are not there.
So what we are in the midst of doing is actually embedding that information into these LLMs in a proprietary way, cul-de-sac, so that it does not go out into the broader internet. We believe that all telcos are going to have access to this information. They're going to have access to these chips. They're going to have access to their own sets of data. The differentiation point is how good are you at actually using it? And to my prior point, we have 600 models already implemented. We have many more along the way. We've greenlighted over 90 new GenAI cases for 2025. There's four things that I would say about GenAI that we've learned. The first is every model isn't good at the same thing. You have to know how to score the models. Some models are better at math. Some models are better at image interpretation.
Some models are better at doing text conversion. You have to know how to score the models and know how accurate the models are so that you use the right model for the right task. Second thing, and I mentioned this already, host and control your own data. If you don't want your data commoditized, then don't put it out into the open internet models for these LLMs. And you've read about some companies where that fate befell them earlier this year. Third point, be an expert in how you access your own data. Prompt engineering. The way you ask a question is incredibly important to the way that the response actually comes back. Tuning. All of these models have parameters that have weightings on them. How you weight a parameter will drive the response that you get back. Are you putting the right weighting on it?
These models aren't plug and play. You have to tune them, and you have to understand how to ask them the right questions in order to actually get the information back in an accurate way and then utilize that information. I believe we're actually experts in that area. So let me give you an example. Jen talked about fiber. We're doing thousands and thousands of fiber installs every quarter. If you, as customers, I would argue you want two things when we show up at your house: to be on time and to do the installation flawlessly. We actually today leverage GenAI to enable us to do that.
We enable an operator to go to the greatest number of homes and the fewest miles driven so that we can maximize our ability and our efficiency in terms of a technician and how many homes they can service over the course of a day. They have these same tools when they're doing the installation that helps them improve the way that we do the installation inside of the home. And lastly, and this is a future state piece, we're actually going to be embedding those same sets of capabilities into the network, in those technologies that I described before, so that we can proactively identify and fix problems before consumers ever see them. How do we know it's working? I'm not going to tell you about our model scoring numbers or how often we use Gemini versus Llama versus OpenAI. That's inside baseball stuff. This is what matters.
We have a 96% NPS score on our fiber installs. We're doing something right. If you want to know whether you're doing something right, ask your customers, and they're giving us a 96% NPS associated with that. The things that we talk about internally around GenAI and how we use GenAI, it's around penetration rates. It's around ARPU. It's around lifetime value. How do we leverage these technologies to improve those sets of capabilities and those metrics across our organization? It's not all just tech speak. So with that, I think I'm done. Turning it over to Brett. Get the clicker out of the tech guy's hands before all hell breaks loose. Thank you.
We're going to take a 15-minute break. After that, we'll have one more presentation from our CFO, Pascal Desroches, and then we're going to go right into Q&A, so see you guys in 15 minutes.
Ladies and gentlemen, our session will begin in five minutes. Ladies and gentlemen, please take your seats. Our session is about to begin. The power to connect is the power to change. Connect two dots and change the whole plot. Connect one day and change a long week. Connect to one thought and change what you speak. Scared of the storms, change their thunder. Don't share the dream, dream another. Connect and change. Face a place. A dance. Town. Fight the night. The right. Anything you like. Connect. Connect and change. Connecting changes everything. No one connects more than AT&T. Bienvenida a casa. Bienvenida a casa. Thanks for having me. Rápido. Muy tarde. Él es el feo Caribe. ¿Me entiendes? Recuérdame. Recuérdame. Bienvenida a casa. Hola, abuela. Bienvenida a casa. They send you a too... Miguel. ¿Recámara? Sí. Venga, we'll speak a lot of Spanish. Connecting changes everything. AT&T.
How's everybody doing? All right. Good afternoon. I am Pascal Desroches, the Chief Financial Officer at AT&T. Today, I plan to cover three topics with you. First, why we have an approach to capital allocation that balances continued investments in the business with returns of capital to shareholders in the form of dividends and share buybacks. Two, the financial results we expect to deliver by executing on our strategic plan. And three, the financial capacity generated by our plan and our plans for allocating that financial capacity. Simply put, we believe our balanced approach to capital allocation is the best way to optimize returns to shareholders over the long term. It does so by investing to drive growth in revenues, earnings, and free cash flow while returning significant capital to shareholders in the form of an attractive dividend and share repurchases.
The strategic plan we've put together allows us to accomplish both goals while exiting this three-year period really well-positioned for accelerated growth and returns thereafter. Our confidence in this outlook is in part shaped by everything we've accomplished the last four years. During this period, we took decisive steps to streamline our operations and exit non-core businesses. We also went back to our roots by investing heavily in our 5G and fiber networks and putting customers in the center of everything we do. The results speak for themselves. We've delivered sustained growth in postpaid phone and fiber subscribers. We've nearly doubled our AT&T Fiber revenue since 2021, and we've transformed our consumer wireline business from a business in secular decline to one we expect to grow at a double-digit CAGR over the next three years.
Since the beginning of 2020, we have invested more than $100 billion in our networks and nearly $40 billion in spectrum. We did all this while we were able to reduce our net debt by $25 billion as we unlocked efficiencies and monetized non-core assets. While achieving these results, we've aggressively transformed our cost base, delivering more than $7 billion in run rate cost savings in areas like customer care and field, administrative expenses, real estate, workforce benefits, and vendor rationalization. These strategic moves have enabled us to consistently grow total service revenue and adjusted EBITDA while driving improving free cash flow trends. We have every confidence we'll continue to deliver these positive trends for the foreseeable future. Let's turn next to our expectations for 2024.
While we still have about a month to go before closing out the year, we have a pretty good line of sight in terms of where things will land. The takeaway here is that we are increasing the low end of our 2024 guidance for adjusted EPS and reiterating all of the full-year financial guidance we shared on our third quarter earnings call. A couple of points to keep in mind as you begin to develop your estimates for 2025 and beyond. First, we expect free cash flow to be near the midpoint of the $17 billion-$18 billion range we previously provided. This includes approximately $2.5 billion of after-tax cash payments related to our 70% stake in DIRECTV, which we've agreed to sell TPG. So if you adjust for these payments, we expect full-year free cash flow to be in the $15 billion range when excluding DIRECTV.
Second, based on the strong year-to-date momentum, we're coming in at the higher end of our prior expectations for adjusted EPS. We now expect full-year adjusted EPS in the $2.20-$2.25 range compared to our prior guidance of $2.15-$2.25. We expect DIRECTV to contribute about $0.30 to adjusted EPS in 2024. So when excluding DIRECTV, our adjusted EPS for 2024 would be in the $1.90-$1.95 range. Now, let's spend a few minutes on discussing our financial guidance through 2027. First, our guidance assumes a macroeconomic environment with low single-digit GDP growth and relatively tame inflation that's basically in line with the current environment we're operating in. We expect consolidated service revenue growth in the low single-digit range annually. We're not providing guidance on consolidated revenue because equipment revenues, which are low margins, can fluctuate from period to period.
We also expect to benefit from additional transformation savings over the next three years. Overall, we expect to deliver $3 billion in run rate cost savings between 2025 and 2027, inclusive of the remainder of the previously announced $2 billion run rate cost savings target we set out to achieve by the middle of 2026. We expect these savings to be largely driven by rationalization of legacy network costs, real estate, force reductions. For adjusted EBITDA, we anticipate an annual growth rate of 3% or better, driven by service revenue growth and margin expansion. Specifically now for each business unit. For mobility, we expect service revenues to grow in the 2%-3% range annually, with mobility EBITDA growing in the 3%-4% range annually as we continue to grow our margins and move customers to higher value plans and execute on disciplined cost takeouts.
In 2025, we anticipate mobility service revenue and EBITDA growth to be in the higher end of these ranges. We expect consumer wireline to grow EBITDA at a double-digit CAGR, with consumer fiber revenues growing mid-teens through 2027. Over the long term, as we scale fiber and exit copper, we expect to achieve EBITDA margins reaching peer-like levels, which are in the 40% range. And in business wireline, we expect EBITDA declines to steadily moderate from high teens in 2024 to single digits in 2027. Altogether, business wireline, we expect EBITDA to decline at low double-digit CAGRs through 2027. We also expect to exit 2027 with the vast majority of legacy business revenues behind us. Additionally, we expect business wireless service revenues to grow faster than our overall mobility business. This should result in business solutions EBITDA approaching stabilization by the end of 2027.
I'd also like to share how pleased I am with the performance of our team in Mexico. The team has successfully repositioned AT&T Mexico, and we expect continued EBITDA growth over the course of this three-year planning period. We expect capital investment to be in the $22 billion range for all three years of the plan. This level of investment will support the modernization of our wireless network while ramping the build-out of our in-region fiber network at attractive returns. In fiber, we're generating returns in the mid-teens range, and we expect similar returns from an additional 15 million in-footprint locations we plan to build. Given our expected range of capital investment as well as consolidated revenue growth, we anticipate that our capital intensity will decline over the course of this three-year planning period.
As described in the 8-K, beginning in the first quarter of next year, we plan to report free cash flow and adjusted EPS, excluding DIRECTV, and we continue to expect that we will close on the sale of our 70% stake in DIRECTV by mid-2025. Excluding DIRECTV, we expect to generate free cash flows of $16 billion or more in 2025, with free cash flow growing about $1 billion annually to more than $18 billion, primarily driven by expected EBITDA growth. In 2025, we expect to benefit from materially lower vendor financing payments versus 2024. This allows us to increase our capital expenditure in fiber and 5G while keeping total capital investment levels essentially flat relative to 2024. This growth is expected to be partially offset by higher cash taxes. In 2024, we anticipate cash taxes of about $2 billion, excluding DIRECTV.
In 2025, we expect cash taxes to increase by $1.5 billion year- over- year. We then expect cash taxes to increase by an additional $1 billion in 2026 before leveling off in 2027. Our estimates for cash taxes are based on the current tax law, including the continued phase-out of bonus depreciation. Excluding DIRECTV, we expect adjusted EPS in the $1.97-$2.07 range in 2025, accelerating double-digit % growth in 2027. Over the course of this three-year plan, this would result in adjusted EPS growing at a double-digit CAGR. The drivers of this expected EPS growth primarily come from EBITDA growth and the benefits from share repurchases over the next three years. Turning next to our leverage target, many of you have asked us why we've remained so focused on achieving our target of net debt to adjusted EBITDA of the 2.5 x range.
The answer is this: 2.5 x range provides us with many benefits. These include flexibility to weather variability in the economic environment without disturbing our ability to invest in the business, while also delivering dividends and share repurchases. Additionally, it allows us to be opportunistic for things such as spectrum or small complementary assets should the opportunity arise. Importantly, during this period, the plan does not contemplate any major M&A, and it's a range that is competitive with sector-level leverage. Simply stated, we don't believe moving to a lower leverage target would provide us with meaningful benefit given the constraints that it would put on us. Let's move next to capital returns. We realize many of you have been patient as we've laid the foundation for sustainable growth by investing at historic levels.
As we've shared before, we believe getting to our net leverage target is a key catalyst for returning more value to shareholders, and that's exactly what we're announcing today. We worked with the board for many months to determine how to optimize shareholder returns over the long term, and we're very pleased with the outcome of these efforts. The operating plan and financial outlook that we have outlined today should generate more than $50 billion of financial capacity during the 2025 to 2027 timeframe. This represents our expected free cash flow after distributions to non-controlling interests, plus the cash we expect to receive from the sale of our stake in DIRECTV, as well as incremental borrowing capacity associated with our EBITDA growth over the next three years.
We expect that more than $40 billion of this financial capacity will be returned to AT&T shareholders in the form of share repurchases and dividends. While the board will re-examine the dividend annually, we currently expect to maintain the annualized common dividend at $1.11 per share. This would result in approximately $20 billion being available for share repurchases over the next three years. We expect to begin repurchasing stock once we achieve our net leverage target. This would result in share repurchases beginning in the second half of 2025, with an initial share repurchase authorization of $10 billion that management expects us to execute through the end of 2026. This means we have an additional $10 billion for share repurchases in 2027, subject to board authorization. Our three-year plan has about $10 billion of additional financial capacity for potential strategic investments, further balance sheet deleveraging, or additional capital returns.
Through this robust shareholder return plan, we expect to return nearly 80% of our free cash flows to shareholders over the next three years. This approach to capital allocation should enable us to exit this three-year period well-positioned to drive accelerated growth and attractive returns thereafter. The team has worked diligently to arrive at this pivotal moment, and we believe this is the beginning of a new era of growth and returns for AT&T and its shareholders. With that, I'll turn it back to Brett to start the Q&A.
We're going to need like one minute to change out the stage, so if you can bear with us, and then we are going to jump into the Q&A.
Calm before the storm? No, thank you. This is Jim's time to shine. And mics and jumpers. Helmets check. Drones check. Checking that thing we just checked? Double check. We ain't got time to chill. We're out here driving this thing. This is when we get to work. To keep you connected, no matter what. Connect before the storm. Change how we weather them. Connecting changes everything. AT&T. You can't see the over $400 billion we invested in our network, but you rely on it every day. You can't see the nearly 2.5 million connected businesses, but they fuel growth, and you can't see the interconnectivity of 28,000 public safety teams, but they protect us. You can't see it, but for 148 years, AT&T has proudly connected Americans.
All right. So we've got time to do a Q&A for a little while here. There are microphones on your desk. When you get called on, you pick it up. They're going to turn them on automatically, so you don't have to worry about that. Raise your hand. I can see most people. And just say your name and your firm for the benefit of everyone on the webcast. And I see a hand right there from Jim Schneider. Let's go with you.
Oops. Okay. It's terrible. Thanks, everybody. Jim Schneider, Goldman Sachs.
I'm not surprised.
That's inside baseball.
What firm are you from?
Yeah. Just two questions related to sort of returns and capital allocation, if I could. One is relative to the $10 billion in incremental financial flexibility you expect, how would you expect you might best kind of direct that capital? Should we think about that as being sort of held in reserve for things like spectrum purchases or potential M&A, or is that something that's more likely to be used for additional returns like share repurchases? That's the first question.
Thanks, Jim. I think the answer is it depends, but it's more likely that we're going to find opportunities to invest in the business and further our objectives to grow in the right way. And that could be spectrum. It could be a capability that helps us drive some new product revenues in a way that we can accelerate it. It doesn't mean that we might do so well or overperform in other places operationally that we open up opportunities to do some other things. But I would say at the front end, we expect things are going to come up, and we expect the market is going to be out there to make some opportunistic decisions, similar to what we just did, for example, with UScellular as an example of something that we might choose to pursue.
All right. More questions? Give the walt a try.
Same question, but a different angle on it. You're super bold up on convergence and fiber. Is $10 billion, and there's also possibly, with Trump now, spectrum opportunities, depending on what happens with DISH, is $10 billion really enough over the next three years to take advantage of the exciting opportunities you have in fiber and then maybe some pretty large spectrum opportunities?
That's a good question, Walt. I mean, there's a lot of things we always have to adjust to that move forward. Is it possible we could have different tax policy than what we have today? Sure. Is it possible that that could drive incremental free cash flow? Sure. Is it possible it's about the same? Yeah, that's possible too. I think it's entirely possible that something breaks left or right, and we'll adjust accordingly when it does it. I would hope at some point, as we execute along the lines of the plan that we put out there, if there's something that comes up that looks like a great opportunity and this team is executing well, investor base would look back at us and say, "Well, of course you want to go do that.
It makes perfect sense to go do that," but right now, I think what we have in front of you is what we believe is the likely case, the most likely scenario that we'll run to, and we feel pretty good about it.
Just one follow-up. One quick. Just one quick follow-up. Is the dividend never growing again? Because you said stabilize at $1.11. Is that it? And any incremental free cash flow that we look at going forward is just share repurchase?
I heard Pascal be pretty clear about it. The board's going to relook it every year and kind of make a decision of what it wants to do. And as I said this morning when I was asked a similar question on television, I want to see where the stock price is, want to look at where the yields are, and want to understand where interest rates are. And all those things are going to go into a decision as we move forward in what we do. But right now, we think this balance for what we have in front of us in this chapter, especially given the float that we have on the stock right now, that this is probably the right thing to do in the ratios that Pascal described to you.
I go to the back of the room. Peter?
Hi. Thanks. Nicely done, everybody. Thank you. Appreciate the event. I wanted to ask you about consumer wireline. Two questions. First, on the $6 billion of cost reductions over the long run, how much of that shows up in your 2027 guidance, and how much of it is incremental to that in the subsequent years? And on a related note, I wondered if you could comment on CapEx in consumer wireline. Is any of your run rate CapEx supporting what's left of that business, and does that mean that over five years there's a step-down opportunity in CapEx? Thank you.
Pascal, why don't you go ahead?
Sure thing. Peter, the way to think about the $3 billion of cost savings that I cited in my remarks, part of that is legacy declines. It's not all of it. So there are other areas we are reducing expenses beyond the legacy network rationalization. And as Susan alluded to, we're not going to be completely out until 2029. And so sometimes as the cost comes out, it's not necessarily linear. So it's a meaningful portion of the $3 billion, but it's not the entirety. And your second question on the capital that's being deployed to support the legacy business, I would describe it as fairly modest, but it's still, look, it does displace some other investments that we would otherwise make annually. And so as the network gets decommissioned, that's another opportunity.
I'll stick in the back of the room. Simon, why don't you ask your last question at our panel?
All right. Thank you, Brett. It was great to get the long-term guidance. Thank you for that. One of the things that you have talked about in the past is the normalization of the wireless industry, yet you seem to be pretty constructive about growth continuing and AT&T benefiting from that. Can you just unpack that for a little bit? 2025 and beyond, volume, price, et cetera. What gives you confidence in projecting out growth continuing through 2027?
What are you going to do with that long-term guidance? Contemplate it from a beach chair? Congratulations, Simon. We'll miss having you around. Jeff, why don't you offer some thoughts on the industry and what you see there?
Yeah, certainly. I mean, as we cited in the third quarter earnings, we're really pleased with the performance of our businesses right now, and that includes our wireless business. I'll remind you that we are not number one in market share. Our priority has been discipline. We've been consistent with our offers in the market, and that's yielded really good attractive growth, great churn, and taking the lion's share of wireless service revenue growth. As a result of that, we see the normalization is occurring. Things are slowing down as we had predicted, but we still have opportunities ourselves to continue to grow. We touched on a few of those in the session today. We are not in the position that we believe we have a right to play in many parts of the business segment, particularly down market that Melissa touched on.
Jen touched on several audiences inside the consumer space where we are hardly present today, and we intend to participate in that part of the market. So as we look to the long-term growth of the business and the guide we've offered today in terms of service revenues and EBITDA, it's a steady as she goes growth engine of mobility leading into 2025 and beyond.
All right. Jonathan?
I guess it's a question for Pascal. The guidance that you've given of 2016, 2017, 2018 adds up to $51 billion. And if you're keeping leverage at 2.5 x, that's going to unlock some capacity from the balance sheet of $5 billion-$10 billion as well, given 3% EBITDA growth. So why isn't there more than $50 billion in financial capacity over the next three years?
Actually, we said it is more than $50 billion. So yeah, you've done the math right.
But it's like $60 billion. It's a lot more than $50 billion. Plus.
We like our business. All right. Sebastiano.
Next question.
We're going to go to Sebastiano.
Sebastiano Petti, J.P. Morgan. Just, I guess, Pascal, sticking with the $50 billion, I think you said the $40 billion of capital returns equates to about 80% payout of your free cash flow. Is that maybe the longer-term kind of level to think that the company is comfortable with as we extrapolate forward beyond 2027?
Really, it depends. It depends what we see as the opportunities that are available outside of our business, and that's always a factor in determining. But look, I think reaching our leverage target allows us to give back a much larger percentage of our free cash flow than we've been doing the last several years. So I think it's fair to say that giving back the majority of our free cash flow to shareholders going forward will be part of the equation. We worked to get to this moment, and I think that should be your expectation going forward.
Thank you.
We'll go to Mike.
Thanks. And thanks for hosting us here today. Two questions, if I could. John, first for you. In the past, you've talked about the reordering of this industry around fiber and convergence. I'm curious if you can give us an update on how you're seeing the industry and AT&T's positioning expanding on what you described earlier. And then second, with the 45 million fiber passings by 2029, is that a destination, or is that just another milestone on the road to continuing to fiberize your broader incumbent footprint?
I don't want to give too qualified an answer to the second part. Let me start at the first part. I think what I see happening right now is consistent with the point of view I've been articulating for a period of time, which is we're not going to have wireless and fixed networks moving forward. We're going to have networks that provide access to the internet, and they're going to be more integrated. My guess is in the United States that that's probably three generally fairly close to national providers, maybe four. I don't know. I think we're going to see that reordering start to occur. We're already seeing evidence of it where people are picking up footprint that, in some cases, repicking up footprint that they sold because they think it's apparently a good decision now to do that.
What I've tried to stress, in my point of view, is I think we can run our plays and choose to acquire the kind of customers and the kind of capabilities we want organically and come out better as a result of that. No offense to Susan. I love the work she's doing. She's done a remarkable job in getting the business oriented around it. We're making fantastic progress, but I don't need any more of her headaches right now. She can take the lunch that she has, and we're going to figure that out. We're going to finish that work and we're going to get through it, but I also want to keep the business focused on what we can control and what we can grow, and we can do it on our architecture. We can do it with our approach to the product.
And I hope you got a sense from what people talked about that that formula works very, very well. And if you didn't get it from the discussion that was here in the room, for those of you who have made the trip to Dallas after we break away, go to the exhibits out there, and you can talk to some people who are even a little bit closer to the work than some of us, and they can share with you why we think this formula works as well as it does and where we're taking it. So I think we're started down that path. I think we're going to see assets reassemble to ultimately achieve what I just described.
If I think about kind of the long term and where I think things settle in, it's hard to say it's the end end, but it's going to be pretty darn close to the end unless BEAD entirely implodes. If BEAD entirely imploded, which to Walt's question earlier, maybe that's one of the things that could happen over the next couple of years if somebody got really upset about it and didn't feel it was executed properly and it all went away. I don't think that's going to happen, but if it did, there may be 15%, 10% of the homes that are still sitting out there uncared for. They're definitely going to require some kind of overt subsidy to be built and to have reasonable broadband to them, whether it's fiber or some other technology.
But my guess is by the time we get to 2030, most of the asset infrastructure in this industry is probably settled in, and I think you've probably pretty well seen most of what has to occur take place.
Yeah. Hey, guys. Thanks. David Barden from Bank of America. John, I just wanted to follow up on that a little bit. There was an interesting slide that showed that about.
Just one.
The most interesting for me. And I know you want to keep me happy. Was that there's 500,000 postpaid phone net adds that got added in the last year in just 9 million fiber passed homes, which means that the other million came from the other 130 million homes around the country. And that is interesting because given how impassioned you are about convergence, it suggests that the industry is on the verge of a balkanization where you're going to have home field and a way field. And does that mean that we start competing differently, home versus away, and we have to have all sorts of different playbooks? And how does that evolve with your competition?
Yeah, Jen, I mean, you're living this day in and day out, so why don't you talk about your home games and away games?
Sure.
I kind of like that analogy, Dave. I think we'll kind of stay with it.
We're here at a football stadium, too.
It's appropriate. I think what I would say is similar playbook. The difference is at the home game, we have converged that we add in. Outside of the home game, when we're the visiting team, we have the plays of continuing to grow where we underpenetrate in areas of the Hispanic segment we've talked about. That is an audience that is the largest portion of the population that's growing. We undercompete there relative to our peers. So we know that we can go out and grow share in that space. We also know that in the price-sensitive audience, we have opportunity to grow relative to our peers there as well. So those are two audiences that we have declared we have opportunity, and we continue to test and iterate.
We have several different types of both pricing constructs, different types of offers, also distribution plays in various markets that we have in the market today and will continue to evolve. Those are two. And then the third we addressed is business segment. And so those are the visiting team plays that we have nationwide to grow share. And I don't know, Melissa, if you'd want to build on business.
Yeah, I'd say from a business perspective, right, we lean into, of course, FirstNet, which is strong from a mobility perspective. The other aspect that I'd say is when we sell to business customers our fiber product, we have a great opportunity to sell mobility not only to the business, but also to the employees of those business customers where we can offer the unique mobility offers that we offer on the consumer side as well.
Could I offer one additional point? While we play a certain amount of home games this season, when we're done with this plan, we're going to have a whole lot more home games than we do today. And this company right now is fighting hard and producing great results in our away game as well. So we know what we have to do outside of that home turf territory. And that's why you see us kind of excited. We've worked really hard to get to the spot, and we are building out that home court advantage each and every day as our fiber network expands.
John, in the back.
Great. Thank you. John Hodulik from UBS. We've been talking about copper decommissioning for a while now, and it sounds like you guys are very confident that you can get that done in the next five years. I think, Susan, you talked about some progress at the state PUCs, but what gives you that confidence? Is it what you're seeing at the FCC with Commissioner Carr coming in as chairman? I'd love to just hear what underpins that confidence. And then I think, Pascal, you talked about margins in that business getting into the 40s. I think Comcast, last time they reported, it was like 48% was their cable margins. You guys are in the low 30s. And they've got video.
So why can't we be thinking sort of high 40s, maybe even low 50s for that kind of business, exit the copper sort of once we get past this piece?
Susan, I'll let you take the first crack at it, and if I need to, I'll backstop it.
All right. Great. John, good to see you. Thanks for the question. What's given us the confidence? We've been focused on this now for about two years, and I would say that there's a couple of things that we've achieved that are giving us the confidence to lay out a roadmap and make a commitment here. The first is, and you alluded to this, where we've made progress on the regulatory front. 20 out of the 21 states clear for COLR regulation, which then starts the FCC process. I went through two of the precedents on the FCC side, one stopping all inward orders. The second, which is really a huge step forward for the team, a new product in the market that we will be allowed, with FCC approval, to replace a POTS customer with.
Without those two pieces on the regulatory front, we would not have the confidence that we can move forward on the regulatory side. I also talked about the product and customer front. We've done a lot of work on what does it take to migrate a customer? Where do we get the right price point? Where do we meet them? Where's the messaging? How do we have the right products out there? We're confident now that we can move the customer there, and we can move them in a way where we can maintain them as customers, which, by the way, clearly is the goal. And then I think the third piece I'd point to is there's a lot of instrumentation that's put in place that allowed us to set up the strategy. We've got profitability down the wire center level.
Now, that matters to finance geeks because they know where to go first, and it kind of is helping us orchestrate how we exit, how we manage the process, so I think those three together have really given us the confidence to put this together and make the kind of commitment that we're headed on.
Yeah, I think what Susan just described to you is that the wireline has been built. All the aspects of what has to happen from replacement products to the regulatory funds, that wireline is now built. The wireline now has to have scale. I believe the administration change is going to help on this. I'm going to go out on a limb. I believe the incoming, assuming he's ultimately affirmed in his role, the incoming chair will lean in aggressively from a policy perspective to try to move this along faster than maybe the previous administration that was not against it by any stretch, but maybe a little bit more cautious posture. And so I think that creates a degree of tailwinds.
The other thing that I think is going to help over this time frame, I don't expect that you're going to see a complete implosion of BEAD, partly because, look, there's a lot of rural areas that are in red states, and people want good broadband in those areas as well. And as a result of that, when money starts getting deployed and broadband alternatives start to show up in some of these places, the notion that we're still talking about preserving POTS services is going to start seeming very, very quaint when the farmhouse has a high-quality internet access capability. And so I think there's a lot of momentum moving this way. We've got a power problem in this country. People are going to need a lot of power to deal with AI.
There's no reason to be cooling and running huge energy-consumptive switches that were built 30 years ago to serve four customers. And so there's going to be social reasons why people are going to finally say, "Oh, yeah, this doesn't make a whole lot of sense. Let's do something better." So we have work to do, John, but this is the kind of work that this team, our DC office, our state organizations are built to do, and we will do it.
John, on your question on consumer wireline, we are in total agreement. There is no reason why long term we shouldn't have margins like our cable peers. Absolutely not. Fiber has, I would argue, even a better cost profile.
Thanks for throwing that can of corn to the CFO, John.
You know what? I figure let's get to 40 before I promise 50%.
We are sandwiched.
We're going to go to Tim.
Thanks, Brett. Two questions. There's a big difference between the 45 million fiber lines you're building and the 88 million homes passed. Can you just kind of walk us through where we are longer term? If California wasn't in the equation, I know you said 10% will be wireless, but that's still a big difference between 45 and 78. Do you think we get the subsidies over time that you build out fiber to that entirely to the 90%? And then secondly, I know you had satellite up for some of the broadband connections. Can you talk about who you're using for satellite, and do you think direct-to-phone can be material for the wireless business and the business in general? Thanks.
Yeah, Tim, let me. I'll start, and Susan can jump in. Remember, she's quoting you. Franchise service area numbers don't conclude that there's any customers hanging off on a lot of those locations, right? They're well gone and past at this juncture. And no, I don't expect we're going to be building out to the bulk of those areas. I think, again, unless something changes dramatically in how BEAD is getting awarded, it's pretty clear to me that in many states, it's kind of had, under the current guidelines that are put in place, a bit of an anti-incumbent bias. And as a result of that, we've been disciplined in looking at it and saying, "I'm not going to chase it just for the sake of chasing it." And do I expect we'll get some BEAD funding again under a current construct? Sure.
Do I expect that that might be more than a million homes that we would build? Yeah, but it's not going to get anything close to reconstituting what's on that chart in terms of the number of voice locations that we use to support. And the reason that map came up during my conversation, behind me, that showed where we were reflecting our service territory is that's kind of what it's going to look like. And it's going to be a much more consolidated, focused service territory and a very different business moving forward as we execute through that. The second part of your question was the satellite?
Yeah.
AST SpaceMobile is our partner. If you're familiar with the construct around it, they have partners basically on each continent. We were one of their primary partners in the U.S. that established with them. We worked with them to bring Verizon in as a partner as well. That contribution of spectrum that we both have brought in is making the product more robust for both of us. We expect we're going to continue in that. As you know, they've put their first satellites up. We've got a full year ahead of us of getting more birds in the sky to ultimately get service up and running, but feel like it's going to be a differentiated approach to doing that. Most importantly, it starts with the customer. Most customers want to use the service exactly the way they use it today on the handset that they're accustomed to.
They don't want to have some degradation in the service definition. They don't want to use some kind of a unique device that has to be equipped a particular way, and we think the approach with AST that we're using is going to fit well into that.
Yeah, can I add some color on that? The other thing I'd add on this, and I think the numbers are fascinating, that 88 million customer locations that I talked about, that is what we essentially built the copper network to serve 70 years ago. But if you look at the data underlying it, and you've got some of this in my presentation, 50% of our territory, we are no longer offering copper services. So they've moved to fiber. They've moved to competitors. So those have already been kind of discontinued through the process. And then I also set up just how few customers are left in those wireless-first areas for copper circuits per sq mi.
So just a few customers left, but the magnitude of that 88 million is to let you know the size of the infrastructure that was built and is still running to support those customer locations.
Kannan?
So, Pascal, maybe one for you. In terms of the cash flow guidance that you've provided versus the EBITDA growth guidance, I mean, I think the conversion rate over the next three years of cash flow is pretty high, and especially if you include also the potential tax break if bonus depreciation was to be extended, you're going to have more cash flow growth than EBITDA growth potentially. So what's driving that? What's driving such a healthy free cash flow conversion beyond just what we see in EBITDA?
Sure thing. First, let's go from 2024 to 2025. There are a few things to keep in mind. One, there will be a significant benefit associated with not having the paydown to vendor financing we had in 2024 and even 2023. So the last two years, our organic conversion has been much higher than our reported conversion because we've taken excess and paid down vendor financing. We are not expecting that drag over the next few years. That's one. Two, for next year, remember this year we had a DIRECTV termination payment, $500 million. It's not going to be there next year. And if you look out, there are some modest things with working capital as you look out the next few years, but on balance, between 2024 and 2025, there are some natural benefits to working capital that should help us.
And then from there, we have every confidence we should be able to convert a high percentage of our free cash flow, of our earnings into free cash flow. Now, none of that obviously factors in anything that's done in tax reform, and all that would be upside.
Here's Greg.
Thanks. Greg Williams, TD Cowen. I wanted to talk a little bit about the fiber-to-the-home build costs. I was pleasantly surprised to see on the slide that the costs were only up 2%. Is that the same sort of cadence you're looking for going forward? Because one of the biggest factors we think about at cost is the density of the markets. And as you go to 35, 40, 45 million homes, it's going to be less dense markets. And I'm just concerned that the costs would rise as you think about your long-term CapEx profile. Thanks.
Yeah. Thanks for calling that out. There's been a lot of hard work done by the organization to maintain cost. And don't discount or overlook the fact that scale, when you're building at the pace and the rate at which we are, we get to work with some of the best third-party labor partners. We have the training that we have done to help ensure that we have low error rates and defect rates. And we have assumed in our plan that things will get more expensive. I mean, make no mistake, we know that inflation is going to occur in the cost of goods or the cost of labor. Our plan is not assuming that does not happen. Our plan does assume that we will continue to bend the cost curves, that we will continue to improve.
The returns that we have seen in our fiber business isn't simply the math equation of what it costs to pass a home and what our ARPU we get and at what rate do we penetrate it. It's deeper than that. That's why I walked you through the three cost buckets of fiber, and the second CapEx that we call it, the cost to connect the home, we have been working and bending the cost on that with the self-install and the training that we've done with our front line such that any level of efficiency we get there allows us to spend a little bit more in cost per passing to reach a little less dense neighborhood and still make our returns in our fiber business.
When we do that at the penetration rates that we're doing it, coupled with the pull-through of convergence with mobility, we're still golden on our business case, and that's why we've guided you to the 45 million that we intend to get built over the next five years.
Kutgun?
Thanks. Kutgun Maral with Evercore ISI. I had a longer-term question. When you look at the last four years, it was very important for the company to refocus its efforts around the customer experience, 5G, and fiber. All that foundation has allowed you to lay out today's multi-year targets where you're investing a lot more and investing at an impressive clip, but you're also starting to perhaps harvest some of the or see some of the benefits through higher margins and the buybacks. Can you talk a little bit about your vision for what today's multi-year plan, how it positions you for the next era for AT&T? I'm not trying to front-run the 2027 Analyst Day, but if you execute on all your plans that you outlined today, you'll have a much larger, healthier customer base, converged customer base. Your cost to serve will be much lower.
Presumably, your CapEx profile could be structurally lower after the fiber build is largely completed. Are you just shifting to a period where you're going to be increasingly in harvest mode and perhaps ramp up the buybacks? Or what does AT&T look like after this multi-year plan?
I don't know. Here's what I do know. You're asking the right question. And I kind of joke internally, whoever sits in my chair at that time and the individual should be sending me a very, very nice fruit basket every Christmas because they're going to be sitting on some of the most pristine infrastructure in this industry, and it's going to just print a lot of cash. And I started my career fairly early in my career. We used to run all the numbers to say, "Well, gee, what if we could take 35% of the cost out of managing the outside plant on a per-customer basis? And what kind of architectures could we deploy to do that?" Because if we could do that, man, this business would be one hell of a business.
And we pounded our heads for decades trying to work through it, including efforts in deploying hybrid fiber coax as an architecture to serve telephony-based services and a variety of other things. Folks, it's here. This is it. This is that architecture. Jeff gave you those numbers. And the best part about it is the stuff still works when it's 20 years old, assuming that the gardener doesn't go put a shovel through it. And yes, the characteristics of this business, a software-driven network that is tied together with the operation support infrastructure that's all software-driven, that has this kind of utility in the hands of a customer, can be a very, very attractive business. I hope we're having this discussion as to what we do with it.
Maybe at that point in time, we understand what adjacencies are that we can pursue, that the business and the brand has built some credibility in doing some things that make sense to invest in. But on the other hand, maybe it's just printing a lot of cash. And our job is to return that to shareholders. I don't know. We'll see when that time arrives, but I think you're on to what the characteristics of this business can be when Susan gets rid of my headache and we finish this work on copper and get it done, and we get the infrastructure out there in the way we're talking about it. This could be very, very powerful. It's always hard to beat the lowest marginal cost provider in an industry.
Ivan.
Hi, Ivan Feinseth, Tigress Financial Partners. Thank you for taking my question and putting on this great event. Can you give a little bit more insight to your views on satellite, like as phones start to converge and have satellite capabilities incorporated, and maybe the ability to roll out satellite-based bandwidth would be less capital-intensive and cover greater areas? So can you kind of give some thoughts on that?
It's your hobby. Why don't you go ahead? You can answer it.
It's my hobby. Yeah. Our strategy with satellite is it's going to augment terrestrial services when it comes to telephony. John laid it out in the answer to the prior question. We've been involved in satellite as AT&T. We've been involved in it from the very beginning with the first one in orbit, in fact. And we understand exactly how much capacity each of these birds is limited to. The amount that we can task up for broadband applications can serve a number of customers. It cannot serve enough customers to meet the demand that's on the ground. And so we've chosen to take the path of it being complementary to the products and the services that we offer to our business customers and our consumers.
Having said that, you're touching on something really important, and that is rather than having satellite dedicated spectrum, lifting up spectrum that's already embedded in the handsets that each and every one of you are carrying, the direct-to-device strategy is an important one. But in order to make that strategy highly effective, you need a fleet of satellites that are engineered to have that kind of RF gain. The antenna array of those satellites needs to be strong enough, large enough to ensure the level of service to that handset is equal to, as best as possible, to what the consumer would expect from a cell site on the ground, and that's why we partnered with AST. That's why we invested with them to get their technology lifted up into orbit.
And you'll see us plug in their services on the edge of where we don't cover over the ocean in the Grand Canyon and places where today it doesn't pay for us to put up wireless cell towers to cover that footprint.
It still operates at a higher cost per bit at the end of the day. And so if you have embedded infrastructure where your marginal cost structure continues to be incredibly attractive, that's going to be the best place to service it over time. But it's going to be a great complementary approach to networks moving forward. You probably get to the next natural question, well, should you own a satellite fleet at some point in time? If I owned footprint outside of the United States on a more global basis, maybe. But probably people who manage global satellite fleets and have expertise in doing that are better to partner with, that can then manage the relationships to use that asset on a global basis, not a U.S. basis, because those fleets are going to be global in nature.
Yeah. Can I ask one more question?
Sure.
But I would agree that it's much better to rent the satellite than to own it. Do you think at some point the industry comes together with some kind of consortium to build, let's say, a LEO-based fleet that would also offer much better high-speed connectivity? Because the problem with the geosatellites is the travel time for the signal. But eventually, if we were to cover the Earth with a group of LEO satellites that would handle high-speed connectivity, then companies could rent space and then have their customer base like you do to access it.
I may be missing your point, but that's effectively what we're doing with AST. It's a LEO-based architecture, and it's a group of companies that have come together to contribute spectrum on a global basis, and we're cooperating, in this case, with a landlord who owns it. It might at some point a different organization structure that makes sense. Maybe. I don't know. It's not on the cards right now.
Thank you.
Andrew.
Hi. It's Andrew Beale from Arete Research. A question for Pascal. I mean, you've got about $35 billion of financings outside your net debt definition: preferred interest, preferred stocks, vendor and supply financing, receivables, et cetera. I guess some of these cost more than your debt, and some of them cost more than your dividend. So I'm just wondering why the decision is that the first use of capital post-dividend is to retire equity, first of all. And secondly, just whether you can tell us what you think the plans are for all of those. I mean, obviously, you said you started running down the vendor financing and supply financing. But what are you doing with the other ones?
Here's the way I think about it. Given the size of our balance sheet, it's really important that we have diversified sources of capital. And yes, some of the preferred subsidiary preferreds yield a higher rate than we would get in the traditional debt markets. With that said, I look at where our equity is trading and the cost of our equity. I think that is right now the best use of capital of this company once we've satisfied all the other obligations. So we've got many of them at pretty attractive rates. Many of them are taxed. Most of them are tax deductible. And so when you look at the net effective cost of them, they're pretty reasonable, and they are less than our cost of equity.
All right. We had a question way in the back.
Yeah. Hi there. It's Sam McHugh from BNP Paribas. I had two questions. Just on fiber to begin with, you talked about seeing accelerating penetration on fiber earlier to getting to 50%. I don't know if you can give us a bit of color, excuse me, on how quickly you're getting to 50% and if that 50% rate is the right way to think about the SMB opportunity, or I think you said today you're at 30%. Thank you.
Jen, you want to touch on your rate and pace success, and then maybe Melissa can pick up the SMB discussion.
Sure. We set our early tenure cohorts and didn't offer any specificity on the months there. But we have seen several markets where we've achieved that 50% in our more recent build and have set the long-term goal to have 50% across all of our AT&T Fiber markets. Melissa, SMB on 30%?
Yeah. So a couple of things that offers. 30% is the average across business. We've got higher penetration on some of our larger segments. We are under-penetrated on the small and mid. What I would say is the infrastructure is in the ground. It's up to us to go get it. And when you look at the plays that we're looking at, I discussed two in particular: distribution. More and more of our customers, particularly small and mid, they have coverage and relationships with indirect partners. We have some of those relationships today, but we're going to continue to extend that, make sure we've got the full coverage and the suite of offerings that we need. And the other thing I would just say from a digital perspective, Jeremy talked to some of the complexities with our digital platform.
We're working that through to make sure we've got a seamless experience. So do I think penetration should be higher than 30%? Absolutely. We've got the infrastructure, we've got the means, and we've got the resources to do it. Look, there was an acceleration down market that we did not accelerate at the same degree. We now know that. We've got the plays to go after, and I expect that will continue to grow.
I'd just add one thing on as you think about the business market. And I've talked about this as usually it comes on the tails of being asked the question about fixed wireless. Small businesses are different than four-bedroom households and single-family units. And a nail salon doesn't sit around and watch Netflix all day long. And the construction yard for the general contractor who's building custom homes isn't in the yard all day long working, but he needs an internet connection. So there's a bunch of small businesses out there that, frankly, they're not necessarily going to need the highest-performing networking capability that fiber necessarily affords itself. And they may be highly mobile businesses, and how they operate and the right answer for them is a great fixed wireless solution for the kind of demand that they have at their home office, call it that.
And when they're out in the field, they need 10 handsets to deal with their employees. And that's the right answer in their case. And I think as you think about the portfolio of small businesses, they tend to be a little bit different, and they all have unique characteristics. And the great thing about where we are with our fixed wireless product, and you heard Melissa talk about it, we'll take all comers for our business-grade fixed wireless product. It's a very, very different dynamic on profitability than dropping it into a home that chews up a terabyte a month. And especially if I can drag some handsets along with it, we'll do that business all day long. And incrementally, we'll invest to do that business.
So I don't think you're ever going to see small businesses per se get to 50% penetration on fiber because some of them are going to want different solutions than fiber. And that's perfectly okay with us.
All right. You got a question, Ryan?
Yes. Thank you. I wanted to ask you a question on AI. First one is, how can you monetize your network to benefit from AI externally, not internally, as a corporation? And the second one, as you decommission your copper, your central offices have a lot of power feeding them. How can you convert them to become edge data centers? Is AT&T going to be a major player as a data center company going forward?
Yeah. You want to go ahead and touch on that, and then I'll fill in behind the.
Yeah. So we have taken a look at exploring the use of some of the space we're freeing up in central offices for the purposes of becoming AI compute centers. The issue is actually less around the physical space. It's about the power required and the electricity demand going into those central office locations and how many GPUs you can actually fit there. So there are some that we are considering, but I would say as a broader matter, we have to fix parts of the electrical grid in order to get enough power in there in order to generate, in order to put enough GPUs in there to make it self-sustaining.
Yeah. I just, I think what Jeremy says, you know, look, we built the kind of power that's going to be necessary for core processing on AI is not the kind of grid and characterizations you're going to see where we drop central offices in typically. There are clearly some clients out there who are going to have some very unique high-performance capabilities, going to have to be very carefully isolated at the edge. And I think selectively you will see us play with that, but it's not necessarily going to be, oh, isn't it great? We can drop it in on the power architecture, that's, by the way, largely DC, into these central offices, and that's going to be the solution as much as it's going to be we're building nodes today that make sense to offer some capabilities around it.
But I'm not going to sit here and tell you I think that's our next fantastic growth business that we're going to come to you on and drive some, to this point in time, unseen margins back into the business. And Jeff, you want to quickly hit the issue on where we are on the product side and how we think about it.
Yeah. And so we touched on a few of the products today, like AT&T Turbo. I think Jen touched on it, and so did Jeremy. We're actually actively looking at a consortium of network APIs with other operators around the world right now, coming up with a standard so that software developers and partners can code against one consistent set of APIs. And Jeremy actually leads the team that does this for us. Why do we get excited about that? It's not because we want to open up the door to any and all to come in and play around on the network just for the sake of doing that. We actually want to go drive some product evolution, some innovation with our third parties.
We touched on Gigapower and some of our open access relationships, network relationships, where they are counting on us to provide some core basic fundamental capabilities in our network and our OSS and BSS stacks. So now, as a partner, we can turn on these network APIs, give them access to what we already have without custom code being built. And that speeds up their time to revenue, just like we enjoy a faster time to revenue when we launch a new fiber footprint. We're looking to monetize that and to help others help us monetize that more quickly.
So, safe, reliable, secure that you heard Melissa talk about is what a customer looks at AT&T, says they believe we have credibility in servicing those needs and meeting their expectations. So when you go into Dynamic Defense and you see how we're applying algorithms and looking at scrubbing of traffic and information, there's an application of AI, but it's not sold per se as AI. It's sold as a great add-on security capability to your connectivity. If you go in and you look at what we can do in the home to understand the dynamics of how the home network is performing since the wireless LAN is becoming more important, we can use algorithmically determined capabilities to understand how we can better serve that customer.
That will be manifested in the wireless home network we sell the customer, not per se manifested in the AI selling them per se as AI. That's how I think about the applications that we're working on right now and where we've been able to drive the technology for monetization purposes.
We take one more question. We got one in the back.
Hey, guys. Rob Palmisano with Raymond James. Thank you for organizing such a great event for us today. You took us through the benefits of wireless RAN modernization earlier. When do you expect to see those benefits out in your network? Will they start to be apparent to customers as you're building it out, or will the realization of those benefits more likely be a couple of years down the road once the build is completed?
From the customer's perception, they'll see it the second that we modernize the cell site that they're touching. So our modernization is not geography market by market. It's not a flash cut, one entire city or neighborhood. It's actually site by site, and it's occurring right now across the nation. It's one of the reasons why, as we develop that strategy, we first tested to learn how to do that with minimal disruption because some of these transformations require taking a sector or a cell site out of service for a period of time, and clearly, we don't want to do that and disrupt the service that we provide customers, but there's going to have to be some level of disruption.
So as we learned how to do that effectively with our vendor partners, we started to gain confidence that we could get this done in this manner of one by one converting these sites over. And we're now at the stage where we're stepping on the gas pedal. We're now ramping up the number of site conversions that we're doing on any given week. And so if you happen to be served by a site or an area that's already received the upgrade, you will feel it today. From an investor perspective, when you'll really see the benefits for us back into our P&L and our balance sheet is as we make it towards the tail end of the conversion when we've got the network converted, and then we can exfil out all of the older wireless architecture and establish our permanent future operating model.
That's a lot leaner and more effective and efficient.
All right. We're going to step off, and John's going to make a few comments to close out the program.
Right? Well, first of all, let me thank all of you for investing the time with us this afternoon and really over the last several years. There's been a lot of explanations that have gone on as we've restructured the business, and it requires a tremendous amount of patience and diligence on all your part to understand them. And I'm pleased to say I think we're getting to the point where we have a business that's a lot more straightforward and simple. It should make your jobs a little bit easier. Hopefully, we'll do an adequate job of explaining it to you. But I do appreciate you investing the time with us today.
I think what I hoped you would pull away from the time we spent with you is a realization that over the last four years, we have made a lot of hard decisions to reposition the business and that we're operating and executing at a better pace, and as a result of that, we have in fact arrived at a different place that now starts to open up some of these choices that you're all asking questions about and that we've been able to lay out, and I think one of the things that's really important, I think we've found the balance in this approach is we want to make sure we're continuing to press our bets in where we have our advantage, and that I believe came through in our discussion today.
And that's about making sure we're scaling this network properly to get endpoints out on the network to aggregate traffic and do it faster and better than anybody else on the right foundation of technology and a belief that doing that, not only acquiring customers, but getting into a position where we can aggregate traffic cost-effectively will be a key determiner of our success over the next couple of years in how this industry shapes up and forms itself. And investing properly in the network in combination with what we're doing and how we're improving our operations with the right technology stacks will be a combination and ability that ultimately results in that business that we just talked about over here. It's one that's built for durability.
It's one that has a sustainable characteristic to it because of the technology foundation that's in place, a more software-driven network, a more software-driven operating structure on the best transport technology that allows us to pick up workloads and move them around. When we do that and we think about what's next for us, here's what you should expect out of us. You should expect that we're going to continue to grow this business in a responsible fashion, that we'll get the right kind of growth. You should expect that we're going to continue to work on our operating structure and our cost models and finish the work that we need to do on some of the legacy side that improves our margin structure, that ultimately drives better returns in cash in the business and growth in EBITDA. That's what we're promising to do back for you.
And we believe that that is largely a set of plays that we get to run inside of our business on an organic basis. And that's what we're going to be focused on doing. So we appreciate your attention. We appreciate your questions. We look forward to clarifying anything we can after we do this. Again, thank you for attending. Thank you for coming to Dallas. For those in the room, right now, we're going to terminate the streams to the external audience. I'm going to ask all of you to hold for just a minute while we're doing that.