All right. Good afternoon, everybody. Thank you for joining us. I'm Eric Lubko, Senior Analyst on the Wells Fargo Communications Infrastructure and Telecom Services team. Thank you for joining us. We're very pleased to have Chris Stansbury, the CFO of Lumen. Thank you for joining us, Chris.
Yeah. Great to be here today.
Maybe we could start off kind of at a high level. I think, you know, one of the critical components to your story is kind of the progression to revenue growth in the next few years. You know, maybe you could, at a high level, kind of give us some of the building blocks from the kind of mid-single-digit declines that we have in the base today to get back, you know, that glide path to growing revenues by 2028 or 2029.
Yeah. I mean, the transformation is well underway. It's working. If you look at where we sit today, just this past quarter, in the third quarter, we announced that half of our revenue stream is coming from the items that are actually growing, and the other half is from items in decline. We're already, from a portfolio standpoint, in a very advantaged position versus our competition. That's why you see our rates of decline at roughly half of what you see in the industry.
If you play that forward, if you quite simply just take the industry rates of growth and decline by product, and then you layer in the revenue impact from the PCF deals that we sold, the $10 billion of deals, as well as, you know, $500 million-$600 million in digital revenue, you get to growth on the business segment by the end of 2028. So roughly, you know, $500 million in each of those two buckets, the PCF and the digital. So they're not crazy assumptions at all. Obviously, we'll work hard to go faster, but we feel really good about where we sit.
Yeah. If you think about the relative growth rates or declines in your business, like, how should we expect legacy? Is that gonna be declining at a faster rate than grow in the near future, or do you think you'll kind of be able to eclipse that over the next couple of years when you think about the relative growth?
I think it will continue to decline at rates that are a little faster than grow for the next year or two. It really depends on how fast the digital adoption comes, and it is coming our way. I mean, we're seeing incredible rates of adoption, albeit on a small base, for our network-as-a-service offerings growing at 30% quarter- on- quarter consistently now. It is just a matter of time. You will definitely see the rate of decline improve as we move through that window. We will continue to give visibility around the components of that as well as the adoption rates around the digital product portfolio growth. You will continue to see the growth items each quarter become a bigger and bigger percentage of the business as we go forward. That is all positive.
In the network-as-a-service growth category, from someone who comes from a more traditional telecom background, it's probably relatively misunderstood. I'm sure you guys will, you know, help us along the way to understand it better. Maybe you could just talk about that platform. I think you have north of 1,500 customers today. You could talk about what makes that particularly unique to other players you're competing with in the telecom ecosystem that are primarily really just focused on the connectivity layer.
Yeah. I'm gonna hover here for a couple of minutes and just. Walk people through it. If you look through enterprise telecom over history, its Achilles heel has been that it's not scalable, period. Every single provider, when they provided a new service to a customer, required a level of infrastructure to support that. If you want Ethernet, if you want VPN, if you want Internet on Demand, it literally was truck rolls and boxes and things getting plugged into other things. It's very static. It's slow-moving, and as a result, it stifled innovation. It also generated a lot of cash, which led to dividends, which led to declining rates of sales, which led to borrowing to pay the dividend. We've seen that story repeat itself multiple times. Network-as-a-Service, for the first time, brings scalability to enterprise telecom.
Quite simply, it is one infrastructure layer, one port at a customer location that can carry thousands of services on it. Those services are delivered digitally. The customer can self-provision those services. There are no truck rolls or far fewer truck rolls, and so it's infinitely scalable. That's why it's so important. We're the only ones investing in that infrastructure broadly. That isn't just about consuming services at that location. It's also about the ability, in a world where data is proliferating, both in its magnitude, but also in terms of where it sits, as data centers, to support cloud, to support AI, are chasing cooler temperatures, power, water. The ability to access digitally all your data in all the places and move it through very high-capacity, low-latency pipes is absolutely critical. It's table stakes for AI, and that's why we're doing it. The crazy thing is no one else is. We feel really good about where we sit.
As we think about kind of the P times Q math to get to the revenue run rates that you've talked about, how should we think about that? Is there kind of a base connectivity layer, price that you'll charge? And then as they take more and more services, you can obviously drive up kind of your average revenue per user? How should we think about kind of the interplay of, of the P times Q?
Yeah. It's the P times Q is exactly the way to think about it. We're gonna give some more information on that at Investor Day 'cause we're learning on the fly. We're learning a lot right now. Think about P times Q really being broken down into a number of subsets because there's DC to PREM connectivity. There's DC to DC. There's DC to cloud. There's PREM to cloud, right? There's all those different combinations. Depending on how you wanna access that data and move it around, the P times Q will be slightly different for each of those.
We do think about, at its most basic level, the number of ports per customer and the number of services per port and the price per port. We're still looking at, you know, that initial port. Do we charge for that port, or do we just put that port in that location? Because then we're gonna force adoption a lot faster and people can start consuming services faster. Still TBD, but, you know, more to come at Investor Day.
As you think about pricing that, when we think about the history of telecom, pricing compression was just a way of life. Forever, right? Obviously, you're building a platform that's different that none of your peers, you know, necessarily can or plan to replicate, at least based on what they've said. Does that give you maybe a little bit more stickiness in the pricing you think you can achieve just given the fact your peers are not even trying to?
Yes.
To adopt something similar?
I mean, look, if you think about the PCF deals, what is our durable competitive advantage? It's, it is the network. It's that conduit that was put in the ground 25 years ago that has remained empty because of the advances in photonics. You know, at that time, we were pulling under 100 strands of fiber through each conduit. Today, we're pulling as much as 1,728, and it's faster. There is a lot of scalability there. For anybody who wanted to compete with us, they'd have to replicate that. It doesn't exist. No one has the pipes. By the way, if it did exist, there would be a lot more noise about other people winning a lot of the $10 billion that we've won. There is no more definitive proof that we have an advantage than the fact that we have won $10 billion of PCF deals.
There is more to come, and others have not. When you think about the digitization, it's so important. I think it was Microsoft that actually had a really great clarifying metric that they're very focused on right now. It's referred to as time to first token. That is the time it takes for data to be consumable in AI. All of the economics are driven around how do you shorten that time. The reason that's important is today, a GPU cluster costs about $2,000 an hour to run. In order to earn a return on those clusters, you have to keep them fed and spinning as fast as they will spin. If you are using legacy telecom that, say, is a 10 Gb connection, it would take 222 hours, almost $450,000 of GPU time to move a petabyte of data.
If you're using a 400 Gb connection, it's 6 hours. The services that we're bringing are giving customers, in the power of their hands, the ability to move their data from anywhere to anywhere digitally on demand and shorten that time to token timeframe. The value we bring now isn't about, "Hey, I can sell you a cheap wave to the procurement department." We're talking to C-suite executives because they're rewriting their business models based off of our ability to improve their total value of ownership.
Maybe related to that, Chris, we were talking about this a little bit before, our discussion here. You know, there's been a lot of fear in the market, particularly the last couple of weeks around, you know, an AI bubble forming and, you know, inflated valuations in the sector. There have been, you know, parallels that have been drawn to the current AI craze with the telecom bubble that crashed over 20 years ago. Maybe you could kind of give your perspective on where we're at in terms of that, you know, the AI infrastructure build-out and how it may be a little bit different than what we experienced 20 years ago and how you're positioned relative to your peers.
I'd say three things. Point number one, 20 years ago, when the internet was developed, there wasn't a use case. It was just a thing. The financing that drove that was a lot of private funding and, you know, highly leveraged companies. Today, the checks that are being written are by very extremely well-capitalized, the best-capitalized institutions in the world. This is just cash off their balance sheet. There's not a lot of financial risk in that. AI is real, okay? That would be the second point. I don't know, we don't know, none of us know who's gonna win or lose. What is absolutely a truth is that as a country, we have to stay in the lead on AI or we will fall behind. This is the great way to narrow the gap between nations, and we have to stay in front.
Enterprise has to stay in front, and they're leaning into it hard now because they know it's the next wave of productivity. The third, last, and most important point as it relates to Lumen is, unlike those chasing data centers or GPUs, we're in the networking business. We are the only ones who are building a network ecosystem to support AI and multi-cloud. The network is not getting overbuilt. The big investments we're making in the network are prepaid by our customers. There is no financial risk. What I know with confidence, what we know with confidence is that while we don't know who the winners and losers will be, we know that the winners will be riding on our network.
There has also been some skepticism in the market around the large hyperscalers building data centers, but a new wave of AI model developers, NeoCloud. GPU as a service, companies that have to take on more leverage are not necessarily that profitable today, driving a lot of the demand. I mean, are they companies that you do business with today as well? Do you have kind of different terms and conditions when you do business with them versus a much more established hyperscaler?
It's really interesting because when we started the announcements around the PCF sales, you know, 18 months ago, things like NeoCloud weren't even in the decision set. We couldn't. Have predicted that at the time. Are they customers, potential customers today? Absolutely. Every contract looks a little different, but broadly speaking, they behave the same. I think that's what the mode would be. We've actually said recently, we said it in the last earnings call, we're not gonna try to predict how big that TAM is anymore 'cause we don't know. Right? We wouldn't have predicted this a few years ago. We're at $10 billion. More is coming. We're involved in a number of conversations right now, but the demand is there.
Of that $10 billion of PCF, how much do you attribute that to more AI training and more remote locations versus kind of the next wave that Kate talks about that we hear about a lot, kind of inferencing, more edge use cases that are gonna require more metro fiber, locations that are closer to the end consumers? I know that's gonna take place over a number of years, but where are we in that evolution?
It's a really good question. Really, over the last year or so, we've seen it shift quite a bit. You still have a lot of training going on, and the training will continue. You know, I was at a conference recently, and some of the AI experts were saying that over the next three years, AI is gonna become 10,000 times more capable than it is today. That will continue to evolve. Companies are definitely leaning into it. You've seen our discussion around Palantir and what we're doing. We're definitely starting to see more inference now. That's where our tools become so critical because it's about, you know, self-provisioned access to the network.
I mean, we had an announcement yesterday, and it was really exciting. It's with a smaller company, but it's with Meter. It's really the first example when we talk about ecosystem partners where tech providers are coming to us to build APIs into our network because it makes their product better. It's a kind of a synonymous relationship. They're a LAN company. We have WAN. Today, digitally, you can provision your networking needs that allow you to move stuff to the edge, right, in a digital way. You're gonna see a lot more of those announcements. I think that is evidence of the fact that everybody is preparing for enterprise needing to, in a very flexible way, access those AI algorithms.
Interesting. As we look at the mix of the $10 billion today, maybe you could talk about where you're at in kind of the CapEx cycle to build out, you know, the rest of the infrastructure. I know there's kind of a combination of existing conduit and some new routes that are part of this deal. Has a lot of the capital spent so far really just been pulling, you know, pulling through conduit over pull work, or are there a lot of new route builds as well that we should expect over the next few years?
It's, most of it is actually existing conduit. The CapEx is really about two things. It's about blowing fiber through those empty conduit, and it's about building the ILAs, the inline amplifiers, the sites where the fiber comes out of the ground for the signal to be repowered and sent on its way. We're well on that path. We're meeting all the milestones with our various customers. In some cases, in some of the contracts, we're getting incentive payments because we're ahead of schedule and on or below budget. That's all very good. At Investor Day, we will separate out the PCF cash flows from the non-PCF cash flows.
What I can tell you is that even when you separate out the PCF cash flows, the business actually self-sustains over the next five years. Right? That is because the revenue trajectory is getting better. It's also because, you know, post the debt restructuring we did not yet two years ago now, our, you know, once we close that sale to AT&T of the fiber to the home business, our interest expense will be cut in half. We will be going from $1.4 billion down to $700 million. The balance sheet is becoming a real point of strength, and it gives us a lot of options strategically. We will give you more of that. What I can say for now is that, with all the CapEx that is PCF-driven, with all the cash flows that are PCF-driven, ex-PCF, this business is self-sustaining.
You have also laid out plans to materially expand your metro and long-haul fiber assets as well that I don't believe is really related to PCF, right? That's really just.
Correct.
The core business. Maybe you could kind of outline the CapEx that's being spent in the core business, not related to.
Sure.
The hyperscalers.
If you look at our CapEx spend today and exclude the fiber to the home business, we're spending about $3 billion. You should also exclude a billion from that because it is paid for upfront in the, in the big builds. Right? So there's really $2 billion of CapEx that we finance. You know, $400 million-$500 million of that a year for the next few years is really going into maintenance. There's a lot of sins of the past that we're fixing, generators that haven't been replaced in 30 years, things like that, a lot of tech debt.
The balance is either funding equipment, the old way of selling that we will still do as we're transitioning as we sell deals to customers, but it also includes a few hundred million a year to do things like these metro expansions and rapid routes. That's all about increasing the capacity, for things like DC to DC, DC to PREM, DC to cloud, you know, PREM to cloud. The connectivity, think of it as a mesh, a fabric mesh where digitally the customer will be able to go from anywhere to anywhere on demand in minutes. That is what that is building out. Again, no one else is building it. That is why we feel very good that we are not impacted by a bubble.
What about some of the other products that are within your grow bucket that you do not necessarily strip out every quarter, thinking like wavelengths? You have talked a lot about that with one of your competitors or IP. Like, what are kind of the growth trends there? We've heard about wavelengths, for instance, as growing pretty substantially. Maybe 5%- 10% a year at, at an industry level. I think you're the largest provider in the industry, so I assume that's, you know, been a nice growth area for you. Maybe we could just kind of dive into some of those.
Waves.
products.
Yeah. Waves, definitely. You know, we're growing pretty much with market. And we're seeing, you know, a transformation, right? There's 1 Gb waves that are, you know, pretty soon gonna be in decline. Because people are moving up the stack, right? They're moving 100- 400. Let's remember, the fiber that we're putting in the ground right now, in the next, you know, 18 months is gonna go to 800 and 1.6T. As that equipment comes along. The speeds and the capacities are just gonna continue to rise. We definitely see demand in waves. We see demands for security. We had an announcement with Microsoft today about that. We can talk about. We see dark fiber, so non-PCF, smaller dark fiber deals. With customers. A lot around connectivity, security, edge c ompute, so.
Interesting. You know, you had an announcement a month or two ago about how your provisioning times in Waves have actually come down quite a bit.
Yeah.
I know that was potentially a criticism from some customers that it took so long.
It absolutely was, yeah.
And, you know, it seems like you've kind of maybe fixed one of the pain points. Is that, has that helped drive incremental, you know, conversations, activity with them?
Too early to say. I can tell you though that the response from customers has been fantastic because our ability to stand up waves as these rapid routes, these waves-ready routes, are put into place. We're doing clusters of 16 cities at a time. The first 16 are done, and we're well on our way. On the second 16, you'll see us continue to do that as we move our way around the country. The time to deliver has reduced substantially, and customers have been very pleased.
The mass market sale or the Quantum Fiber sale to AT&T sounds like that's gonna come very soon in the first part of next year. Maybe you could just talk about some of the, you know, cash free cash flow accretion you're gonna get from that sale. Then as we think about what you retain, really, your kind of DSL and copper base, you know, presumably AT&T is gonna try to overbuild that in a quicker fashion. I assume you will probably, you know, enable them to do that given the structure of the deal. Does that potentially give you the opportunity to take some cost out of the fiber network? We have heard about some of the large wireless companies or fiber companies decommissioning copper. Taking out maintenance CapEx.
Yeah, really a few things. Really excited about that deal and excited about the construct of the deal. We, to your point, are selling only the fiber assets. And frankly, we're selling really from the edge of the neighborhood into the neighborhood. We kept those main lines because they're critical to our enterprise strategy. The structure of the deal really worked well for us. We'll use the proceeds from that sale, the after-tax proceeds, as well as some cash on the balance sheet. We will entirely, you know, shortly after closing, wipe out $4.8 billion of super priority debt, which is what we put in when we did the debt restructuring not quite two years ago. That will bring down our total debt from a little over $18 billion to just a little over $13 billion, bring our total leverage down to 3.738.
It'll go down from there. Oh, by the way, you know, over the course of the last year, we've significantly restructured the rest of the stack. It is a much more smooth, normal maturity curve. The team's done a phenomenal job. That's what we'll use the cash for. As it relates to the consumer business, there will be some overbuild, but I wanna be really clear. The bulk of our footprint, we pass, you know, 17-18 million homes. I think we've got a little over a million voice customers and about 1.4 million DSL customers. Most of those sit in very rural areas where I don't think they're building fiber.
The opportunity though is, and we're in conversations, is I think there's a real opportunity between the two companies for us to use potentially AT&T's fixed wireless product. To your point, when we think about areas where there's less penetration, the cost savings associated with, you know, turning that off. Could be substantial. We have to get the economics right, but we will manage that business for cash. I think it'll be a long wind down because it does sit in those rural areas. We're probably talking seven-plus years. There is a great cost takeout opportunity. Quite frankly, I think there's a copper mining opportunity, and that isn't even on our radar screens as it relates to kind of financial outlook at this point.
The maintenance capital you spend on the copper network is a couple hundred million dollars a year, is that right?
Yeah, that's right.
Just keeping.
Just keeping things turned on, yeah.
Yeah. Yeah. Gotcha. Okay. There is some opportunity there. Okay. Great. Maybe we could just talk about broader cost transformation. I know you're in the midst of, you know, a broader cost transformation effort. I think ERP phase one is the latest m ajor initiative you're on, but I'm sure there's a lot more that lies ahead. A lot of this, you know, is within your guide to grow EBITDA in 2026 and thereafter. Maybe you could talk about the main cost areas you're focused on today, maybe any preview of what w e should expect over the next couple of years.
Sure. If you really think about the sector we're in and the landscape, the only way that scale, we'll go back to where I started the conversation. The only way that scale existed in enterprise telecom over the last number of decades is buying other companies. When that happened, none of us, and I mean Lumen, I mean the other big guys, none of us ever consolidated the IT stack. So it is a dog's breakfast of tech debt. It's ugly. The path to freedom isn't trying to consolidate all that.
The path to freedom is selling all the new stuff that is replacing the old stuff on a clean new set of IT infrastructure and then turning off those old systems as those products retire. A big, big piece of that modernization and simplification is really around addressing the tech debt. There are two paths, and we're running them in parallel. One is the big investments that take years to get in place, ERPs, right? The ServiceNows of the world, those kinds of things.
The other path though, and I can't emphasize how much it's helping us, is the use of AI with companies like Palantir, where they can come in and their algorithms will find connectivity and data across multiple systems that actually move us forward much faster. We don't have to wait for those systems to retire. Case in point, when you're, you know, running whatever it is, you know, 16 different order entry systems, you end up with customer disputes, invoice disputes. We're now, after a five-week test, resolving half of those through Palantir's tools. As we go forward, there's gonna be a lot more of that. We're gonna get more efficient, and that's what's driving the $1 billion of exit run rate by the end of 2027.
There have been some big announcements from some of the telcos as well about, you know, headcount reductions, and people are always focused on, you know, the future of AI and what that will mean for the labor force. As you look at, you know, your headcount today and think about the kind of productivity unlock you can get by enabling them with more AI tools, whether that's software stack, whether that's, you know, building new routes, like, is there a lot more of that to come that's beyond the $1 billion?
I think we're, as a society, in a place where that's inevitable. I think what our job is as leaders is to make sure that we're preparing people for the future, right? I think the more that people can be exposed to AI and how it can change their lives, and yes, there will ultimately end up being reductions, those are skill sets you can take to your next role, right? I think that is a reality, right? It's a reality that a lot of us face. I don't know how much of the other announcements across the sector are driven by AI versus just broader cuts. We're focused on doing things the right way and making sure that everything is focused on the customer and that as we improve that customer journey and we can drive efficiencies, you know, that's where that comes from.
Maybe just touching back on the revenue equation again, I know we talked about some of the products, but as we think about the customer end users, so the large enterprise versus mid-market versus public sector, how you kind of expect those to trend over the next handful of years until we get back to this kind of revenue stability and revenue growth.
Yeah. I mean, there's opportunities across the stack. We're very strong in public sector. We'll continue to be strong in public sector. You know, wholesale has been a decliner for the last few years because, again, we sell to our peers, they sell to us. Everybody's pulling back on copper circuits, so that's been impacting that. As we move into the future and we're doing more digital, I think there's wholesale opportunities. In mid-markets, and large enterprise, different, different scale.
I think, I think network as a service is big in both of those. You think about mid-markets and the ability to deliver services digitally to customers, what a great way to reach that audience. In large enterprise, it's that and, and the and is, extremely disruptive stuff like direct cloud on-ramps where, you know, very large users of data that need to feed those GPU clusters, do not want to have to go through, for lack of a better example, county roads to get to the cloud, paying toll charges all along the way when they can get on the autobahn by connecting to the cloud directly through Lumen's network. Those are direct cloud on-ramps, 400 Gb connectivity that is coming now. That is, that is very disruptive to the space.
Maybe, in the last few minutes here, we can touch on the balance sheet. You alluded to it a little bit earlier, but, you know, you've reduced your annual cash interest expense by over $230 million, I believe, year to date. This obviously doesn't include the cash proceeds you'll get from the AT&T transaction early next year. It seems like you're in a very good place. I guess, and if you look at your maturity stack, you really don't have much at all until 2029. You alluded to on your call that there's still things you're looking at. Maybe you could just talk about, you know, what's kind of next on the agenda now that you've deleveraged so substantially.
Look, it's been an extremely complex balance sheet, and I don't think that does any investor any favors. Whether you're a credit investor or you're an equity investor. You will continue to see us simplify the structure both in terms of number of reported entities so that we more closely align the way we run the business to the investors in those entities. You're also gonna see us simplify the structure.
I mean, we have four layers right now, right? You've got the super priority, you have first lien, you have second lien, you have unsecured. We can simplify that too. Super priority obviously goes away. We'll continue to do things in line with that in, you know, in principle. I, my goal is that our balance sheet is boring. I wanna be boring. I wanna be the most boring guy at the credit conference not the one that gets chased around the hallways.
Yeah. No, that would be quite a transformation from,
Yes, it would.
From many years ago, and I guess longer term as we think about, you know, you returning to revenue growth, continuing to improve profitability, and you've talked about declining capital intensity is right, right? And how do you think about returning capital to shareholders? I mean. It's a little bit premature to get into that today, but just along the trajectory.
Yeah. No, and we've talked a little bit about it. I mean, the look at the number one goal is to make sure that we are funding what we need to fund to execute the turnaround. We're doing that. That could be organic, which is where it's focused today, potentially. I'm not signaling anything, nothing on the table today. It could be inorganic if the right thing comes along.
The second thing is to bring leverage down a little more so we've got some extra gunpowder. After that, I think you're looking at share repurchases, to the extent that there's excess cash flow because, you know, I think the stock remains well undervalued. We've seen obviously some nice, you know, valuation uptick. Obviously, the market's pulled back on the AI trade. I think fears of a bubble that don't really exist for us. I think that's where we go next.
You've obviously pruned the portfolio significantly with, you know, mass markets, with, you know, some of the international assets being sold. I mean, do you sit here today, you think most of the kind of non-core asset sales are behind you, or are there still some things on the margin that you can potentially buy?
I think there's still some things on the margin. I mean, certainly there's a copper mining opportunity. Right? And there's a few other little things in there. Our real estate footprint will shrink over time. We will constantly be in the mode of, you know, getting rid of those things. Are they material to our outlook? No. But it is just good hygiene and a lot of cost hide in those assets, and it is a way to manage that.
Yeah. Fair enough. As we think about kind of longer term, where do you think kind of the right leverage level is for Lumen, you know, just based on your current investment?
I haven't said it, but probably low threes, somewhere in there. We'll give more guidance at investor day and more certainty. Look, there's two ways we get there. Right? We get there by paying down debt. We also get there by inflecting EBITDA next year and growing from there. One thing that we haven't talked about, as we close, is the ecosystem layer. I think in our transformation, that may very well be the most important thing. That's where I mentioned earlier, tech partners are building APIs into our network and pulling our network through with their products because it makes their time to revenue faster.
It makes their customer experience faster. The number of tech companies that have shown up and have started to build APIs into the network, it's staggering. You know, we talked about Meter yesterday, the Microsoft announcement today where they're taking our Black Lotus Lab threat intelligence and building it into Sentinel. You know, we talked about Commvault on earnings and automated backup and recovery, Zscaler. It's coming. I think a big piece of how we delever the company is actually by growing the numerator. We're super excited about that.
That's a great place to end. Thank you, Chris, for joining us today.
Thanks a lot.