Started. Good afternoon. Welcome to day one of our 53rd Annual TMT TD Cowen Conference. My name is Greg Williams. I cover cable, wireless, and telco here at TD Cowen. I'm joined in this session by Chris Stansbury, the CFO of Lumen Technologies. Chris, thanks for joining us.
Yeah, great to be here.
I wanted to start off with the AT&T deal that was announced last week. This deal structure is very unique. It's a 1st of a kind, potentially last of a kind. Help us understand why the deal was structured this way. I understand you're going to keep a lot of these wire centers for the enterprise business, and you're going to direct connect to the cloud. Why do you need 1,700 data centers to do this?
Yeah, the reason that the deal was structured the way it was was really for a few reasons. The value on the consumer side is really in the fiber. That's the go-forward strategy. That's the asset that people were most interested in and w e built a great business. And that business can be taken much further. From a buyer standpoint, that's why the focus was there. From our standpoint, that footprint generates a lot of enterprise EBITDA, in addition to some very healthy consumer copper EBITDA that is in very rural areas that will take years to go away. Fiber is very unlikely to be built to those locations.
From a strategic standpoint, as we go forward, the real value in the enterprise offering that we're bringing is the marrying of the underlying footprint of our network with the digitization of the services that we offer. Rather than the typical legacy telecom motion, which is every service having its own infrastructure layer, which is very asset intensive, very expensive, very slow to scale, and ultimately drives very little innovation over time. It's one infrastructure layer, that is digitally enabled where services can be dropped on top of ports by the many. The physical locations, while all 1,700 we may not need, the reality is there's a customer base off of those wire centers, and there's capabilities we can choose to build in individual wire centers that allow us to really untap the value that exists as we go forward on the enterprise side.
That's why we kept what we kept.
Got it. I want to talk about a few of those things. I also want to discuss AT&T. They suggested to us on their follow-up call on that day they could build to another 5 million homes possibly in your territory. Trying to understand how that works. Is there additional service agreements that need to happen, or do they kind of have carte blanche or free rein to overbuild in that area? Can others do the same?
Yeah, so our initial thinking was that, we would build to about 8 million enablements. Their messaging is really going to be closer to the 10 million enablement level, which makes sense, I think, with the wireless offering, the brand recognition that they have, t heir ability to drive, greater penetration and reach is higher. Now, their ability to do that, the fiber assets that they're acquiring sit in 427 of those 1,700 or so wire centers and they can build wherever they want out of that 427 and I think that's going to be their primary focus. If in time as markets grow, there's the ability to go to 428 or 429, then we can have a separate discussion around those additional facilities. I think their core focus will be building out the enablements and penetration around those 427 wire centers.
Got it. When it comes to the copper sunsetting, which sounds like a really good opportunity for all the telcos, you've told us before you have a seven-year glide path roughly with the remaining CSL subscribers and the remaining business in the COs. Is there a fear that maybe you just told an overbuilder right now to come into that territory and cannibalize that business? Are you worried that AT&T will cannibalize it faster? Does it accelerate your seven-year sunset horizon?
No, I think that the seven, and we call it seven to 10 year horizon, I think is a conservative look. The vast majority of the remaining penetration, and again, just to put context around this, 18 million households passed with copper. We got 1.4 or so million DSL customers and a little over a million voice customers. Those are primarily in very rural areas that economically are very unlikely to see fiber. That is why we have confidence in that rundown. And candidly, to the extent that any of those customers are in more metro markets where fiber could be, that is actually good for us because it allows us to turn off those networks. Yeah, we are not concerned about that.
Got it. The 1.4 million DSL subs are so rural that there's probably no overbuilding, whether it's AT&T or somebody else.
Right, exactly.
I want to talk about the opportunity that you could have in terms of AT&T backhaul. AT&T is going to own the fiber from the node to the home and then they're going to have IRUs from the central office to the node. That fiber behind the central office to get to the internet, shouldn't there be additional commercial services from AT&T? Because now they're going to see 4 million homes to go and get that to use maybe your network with. It could be an upside opportunity.
I know. I think there's a possibility there. I mean, the reality is today, we all buy and sell from to each other, right? Because none of us have the network coverage to get to every endpoint that we need to get to. There are very healthy wholesale relationships that exist amongst all the major players. Those agreements exist and obviously as AT&T grows this business, we'll be able to support whatever needs they have. Frankly, they may choose to use their own network in certain locations to do some of this backhaul as well. We'll see how that plays out, yeah, there could be some more there. We'll have to see.
Got it. Also, in the slide deck on that day we announced the transaction, you said the deal was going to be used to pay $4.8 billion in super priority debt, but you're only seeing $4.2 billion in after-tax proceeds from this deal. Help us reconcile the $600 million gap from selling it at $4.2 billion, paying down $4.8 billion. Isn't that cash coming from the hyperscalers that's supposed to be used for dark fiber buildouts?
A few things. I think the $4.2 billion is conservative. We may be able to do a little better on the tax side. We'll have to see. There's obviously a lot of complexity in that. The other parts of this, though, are we were very clear when we signed the PCF deals that we had a cash contribution margin that approximated our EBITDA margin. There is margin there. We're driving a lot of EBITDA and cash savings from our modernization and simplification program. Again, $1 billion out by the end of 2027. That's going to be contributing. I wouldn't point to any one thing as the source of cash. The reality is, as the business is turning around, we're performing really well. We had a great first quarter. We will hit a point of inflection on our EBITDA this year.
We'll call that when we see it. The business is getting healthier. Our ability to drive further cash flow improvements through debt reduction and interest rate deduction or interest payment as we go forward is going to kind of snowball in our favor. All of that kind of fuels the cash flow that we can use to reduce that $4.8 billion.
With the cost savings initiative, you have a $1 billion cost savings initiative supposed to be realized by year-end 2027.
Correct.
Can you help us understand how this deal fits into that billion-dollar cost savings? Does it help speed it up? Or in some ways, could it help potentially slow it down because you're going to need to maintain or retain some central offices that otherwise might have been shut down?
No, I don't think it slows us down at all. The program itself is off to a great start. We're really pleased. We're still doing work to determine if that's just a timing thing or whether it's actually a scaling faster thing, in which case maybe we can do better. We're very confident in the billion outside of this transaction. We'll make adjustments as we go forward. If anything, we're cautiously optimistic that that might be able to creep a little higher.
Got it. You're consolidating central offices, or at least you've messaged consolidating central offices. A lot of the telcos are excited about copper decommissioning. Does AT&T have a say in this? If they're going to want to expand out to additional 5 million homes, they might say, hey, don't be shutting down some of these central offices. We're going to need them. I think you alluded to it. They have 427 wire centers. If they want to grow a few, just help us with that dynamic.
Yeah. I mean, again, whatever we do outside of that footprint, we can do outside of that footprint. Even within that footprint, if we choose to consolidate, we would obviously make sure there is no disruption of service. There may be efficiency plays that we can drive and still provide the connectivity that they need, right? We will continue to modernize the footprint to meet the needs of today, including the needs of AT&T as a customer in this case. The future of that business and its success, that consumer fiber business, is really important to us. We have got a lot of people that have worked really hard to get it where it is. We want them to succeed. We want that business to succeed. We will be very supportive as we go forward.
As you consolidate central offices, one thing AT&T tried, was it early last year in California saying, hey, we can offer fixed wireless as a proxy, and then we could shut it down. You do not have a fixed wireless offer. What would you do? Would you be white labeling like AT&T or T-Mobile or Verizon's fixed wireless and offering those?
We have the opportunity with this transaction to use AT&T's product, where it will work. We can also use other services as needed. I mean, again, what it really comes down to is how many customers remain on those networks, particularly in those more remote areas. When we get to a point where it just doesn't make sense to keep powering those networks, we can convert the customers to other services. We do have wholesale relationships. The bigger piece of it, though, is the regulatory environment is getting easier, where we have to be allowed to do that. A number of states have already eliminated their carrier of last resort burdens and requirements. I think more will follow. There is a lot of energy around that.
Great. Another element of the cost savings initiative is the network integration, turning essentially four networks into one. It's Time Warner Telecom, Global Crossing, CenturyLink, Level 3. Help us understand what inning we're in on flattening out that network. What else needs to happen to achieve this? Just anything in terms of what goes into it and when do we realize the benefit?
Yeah. We will, on a go-forward basis, be almost entirely on one network by the end of this year. Think of that as all future services will be sold on that one single network. We will not be selling on the four old networks anymore. The part that we will do, though, is that any legacy revenue that exists on those other four networks will maintain as those businesses slowly decline over time. As we convert customers to newer services, they will sit on the new network. We are not going to try to consolidate a bunch of old legacy revenue to a new system and new IT stacks. That would be, I think, a failure.
Care systems, billing systems, they'll still stay separate until you sunset them out of.
Exactly. Exactly. That is how that is going to work, we are making great progress. We will be substantially there by the end of the year. Again, the vision as we go forward, and we have talked a lot about it on earnings calls, is, as I said earlier, it is really one network infrastructure with digitally enabled ports. We call them fabric ports. You will be able to layer a voice service, a connectivity service, a security service, whatever it is, on top of that port. It is a very scalable and, frankly, very customer-friendly way to deliver services and for the customer to be able to self-provision. Fewer truck rolls, more immediacy, more flexibility in how customers want to consume networking today.
Right. So it's not just cost savings, which are obvious, but it's this revenue opportunity to build on so much on these ports by doing this. When can we start to see this both on the cost side or even the revenue side? When does this happen? Is it gradual? Is it a flash cut? When does this?
Yeah. Let's talk about this in a few different ways. We've talked about revenue inflecting kind of late 2028, early 2029. The underlying assumption underneath that was, everything that we sell today either grows or declines with market rates. The only two exceptions are, as the PCF deals ramp, that revenue will start to amortize in. That's one. The 2nd is, in that outer year, that 2028 window, we had a few hundred million of digital revenue. That's it. That gets you to a point of inflection. We're already seeing, even though the business is in decline, a much better trajectory than our competition because we've been focused on this for years. Our dependency on kind of those bigger, older legacy buckets of VPN, Ethernet, and voice are smaller than what our competitors see.
Now, what we're looking at, and we're going to give far more color on in September, we're going to have Analyst Day, is really around which services we see scaling 1st because these are going to be kind of J curve deployments. Is there a way to pull in that point of revenue inflection sooner? That's what we're hard at work on right now. The biggest two that are out there right now are really around cloud voice and our ability to retain and move a lot of our legacy kind of copper-based voice customers to a cloud product that we own. The other is our ability to rapidly deploy cloud on-ramps, which in an AI-first multi-cloud world are absolutely critical to enterprise. It's much lower latency than the alternate today. It's also more effective from a cost standpoint.
The value to enterprise is there. The more that we share this with our enterprise customers, they're asking us, can I get it faster? It's not, why are you doing this? It's, can I get this faster? This is exactly what we need.
Great. I definitely want to talk about the cloud on-ramp opportunity as well. Before I do, just to maybe reiterate, what you're saying is the revenue inflection's not really Herculean. It's algebra. It's saying, what's the growth rate of your growth products? Put that on a market level. What's the legacy product shrinking at? Put that at a market rate. You're outperforming on both those ends. Possibly layer on cloud voice, layer on some sort of cloud on-ramp opportunity. You can come in actually a little bit sooner. Is that right?
That's what we're really modeling out right now, right.
I wanted to talk about AI. Obviously, the external opportunities with PCF are there. I mean, before we get there, just talking about internally, any updates on what you're doing in AI inside your house?
Yeah. We're being very aggressive with it. If I think about my own function in finance, there's a lot that we have to touch physically because of old IT systems and whatnot. We're moving to update one ERP with SAP that will go live in a few months. That'll unlock more capability. We're also using services like Palantir to help us navigate the legacy systems that we have today to get to things like customer dispute resolution in minutes rather than hours or days. It's making us far more efficient. More to come on that. We're leaning into AI hard. We're starting to see some real benefits.
Got it. Let's pivot to the external opportunities, the PCF wins that you've had and the pipeline. Glad to hear that the business remains on time and on budget. Anything you want to call out here? Is your build engine fully ramped at this point? Is that going to take a few more months?
It's fully ramped. Teams are doing really well. We actually, two weeks ago, delivered our 1st complete segment to a customer. It was ahead of schedule. We actually got a bonus payment for that. The teams, and I mean not just Lumen, these are our partners that are helping us with the construction, are doing fantastic work. We are very pleased with where that sits.
You mentioned a bonus payment. My next question is, can you get an update on your cash cadence? Because you have a lot coming in from milestones you hit. You have a lot going out from the CapEx. You eventually have to pay taxes on it. A lot of large lumpiness. Anything you can provide in terms of updates?
I'd say really kind of through 2027, we're going to receive, I think, just under $8 billion or so.
From the August announcement from last year to 2027?
Yeah. Yeah. We'll have CapEx and taxes that trail for a year after that. All of the things being equal, if that was our only business, you'd have a negative outflow, obviously, in that 2028 time frame. Obviously, with the cash benefits of the sale of consumer fiber, just the underlying improvements in the business, the modernization, simplification savings, I think those all bode well on the cash side.
Got it. On the $3.5 billion PCF funnel, I know you do not talk about updating your funnels anymore. Just the general characteristics, what you have said in the past, it might require more greenfield and sort of more build-out, if you will. Any particular verticals or use cases as you think of this $3.5 billion funnel that is different than the $8 billion funnel, which seems like it is more training?
This still is really around the training.
Phase one.
There is a really important tell in this. If you think about our network coverage today, we can get from pretty much anywhere to anywhere in terms of the major metros where those training environments exist. We may have to go through an extra node or two in our network to get from A to B. What the hyperscalers are saying is, latency is bad for AI. I need a direct city-to-city pairing between A and B. These are big, complex, and really challenging builds that we have to get really good cost estimation on. We have done a lot of work around that. Hopefully, we will have an update soon on a piece of that. The conversations continue. The reason I emphasize the driving down latency is because that is exactly where large enterprise is going. Latency is bad for AI. It matters.
What we're doing with things like direct cloud on-ramps are to help enable that latency reduction.
Definitely want to talk about the direct cloud on-ramps next. I just want to talk about, you mentioned most of this is the training phase or phase one. It doesn't sound just so discreet anymore. I came back from ConnectX not too long ago. We're talking about some of these deals are a little bit more nebulous. It's training with clearly some inference in there, whether it's waves to availability zones. Are you seeing that at the RFP table where the deals are?
We are starting to see more around inference. Companies are talking about it. It really does matter. What they're really starting to push for is inference at the edge. Because again, it's about.
I'm stopping you there. What's the definition of edge or the availability zones?
Really, edge means that you do not have to bring it all back to your own on-prem data center to do the data manipulation you need to do. Keep the data kind of further away from home so that you are not dealing with latency. Again, what the demands of the network today are is to take a server that sits 2,000 miles away from another server and make them behave like they are two feet apart from each other. That is where AI is pushing this. We already have an edge of network compute environment that is built. It covers 96% of U.S. businesses within four milliseconds. That sits largely in our metro rings that already exist in our central offices. When you pair that with our ability to get to cloud faster, that is enormously powerful. That is all about inference.
There was an announcement a few weeks ago with IBM and Watson and inference at the edge. That is really what that was about.
Got it. On the Direct Connect to the cloud, it's an interesting concept because typically you'd have to go to an interconnect facility, as in one of those cloud on-ramps, two dozen or so in the U.S. It doesn't sound like you're doing some sort of parasitic tethering going into those interconnect facilities. You're actually bypassing.
That's right.
Interconnect completely because your fiber is going to the CO. Then your CO will connect to, say, a Google Cloud directly.
That's right.
You bypass it. Curious how the interconnect data center companies would respond to this. Could you see some retaliatory response, meaning they'll price you up anytime you're in their facilities? Or they'll price you or one of your customers that's riding on you to do this because they can see this as a threat?
Look, I'm sure there will be a competitive reaction. What is really exciting to us is there's finally innovation in this space. If you think about the world of compute and networking and AI and cloud, there's been enormous innovation in everything except the network. If you were sitting at a table today with a bunch of really smart engineers and you were designing the network to support an AI-first multi-cloud world, you would not draw what we have today. What exists today exists because legacy telecom has not innovated. We've all been competing on, we call them dumb pipes. It is really the commoditization of just simple connectivity. Carrier-neutral facilities have become the de facto network for cloud because there's been a lack of innovation. There is a real need for those carrier-neutral facilities in the future.
There is also a real need for innovation and a better way that slays latency. That is what we are bringing. We love competition. We would love to see more competition. I think this is an environment where everybody wins. It just so happens, though, that in those applications, in those born-in-the-cloud companies that are being born there because they cannot have latency, the way that it works today is untenable. It just does not work for what their needs are today and what their future needs will definitely be. I think the reality is this is new TAM. This is meeting a need that no one else can meet today.
It's not only lower price because the interconnect facilities charge higher interconnect fees. You can bypass those. It's also a latency thing.
Latency is the main thing. I would say that the cost piece, while important, is secondary. It is latency that kills AI. AI, again, you have got machines talking to machines. If there is a delay, that kills that dialogue. You cannot have that. Again, we are all about innovation. We are going to bring our piece of it. We hope others do as well.
Right. You're working with Google on this Direct Connect cloud onboarding. That begs the question, is it safe to assume you're looking to strike deals with AWS and Microsoft?
I think you'll see that continue to expand. Stay tuned.
OK. Will do. You do have an Analyst stand in September.
Yeah, exactly.
I want to talk about margins and economics with the whole network as a service that you're providing here. How does it compare to your legacy business? Are there revenue shares with this sort of model? Partnerships? Are margins higher or lower?
I think we could see all of that. The margins will be very high because, again, it's digitally delivered. And frankly, in many cases, will be self-provisioned by the customer. Fewer truck rolls, w e'll still have a need for, obviously, the network maintenance, but the ability to drop multiple services on one port is a very scalable motion. That's very margin-friendly. Frankly, over time, while we haven't proven this out, we would also expect it to be less capital-intensive per dollar sold.
That sounds interesting to me because if you need to connect from a central office to a Google Cloud point-to-point, that's a brand new connection. How is the CapEx?
It is largely leveraging connections that already exist. In some cases, we are expanding the capacity of those connections. That is already in our capital budget. Again, if you think about a legacy world where every service had its own infrastructure and a new world where multiple services share the same infrastructure, it is just much more scalable. We would expect the capital intensity to be somewhat lower as we go forward.
OK. I wanted to move to the waves business. Cogent was here earlier today. They're off to a slower start, but they're ramping their wave business. Not to speak to them specifically, but they do say they can provision waves in 30 days or less. Are you doing the same? How do you differentiate? How do you defend share? Is that a real competitive advantage?
Yeah. So we're growing our waves business with market a nd we're pleased with that. I would say that some of our competitors actually do a better job than we do today in terms of speed to quote. That's something that we can continue to focus on and will. Bigger picture, we're not going to play the game the same way that our competitors are. Our competitors are still very focused on price as the selling vehicle. Again, great. If you want to buy cheap waves to get into a carrier-neutral facility, which is going to have much higher latency and a higher cost because of cross-connect fees, then great. Do that. We may or may not participate in that. That's not what we see as a winning formula.
We see the winning formula really being to put services in the hands of the customer that they can self-provision and to provide them with faster, more secure, and lower-cost connectivity in total than they have today. I think we're very confident in that strategy. I think we've got a winning strategy. Everybody else is a long way behind us right now. We will see how that plays out over time.
With about a minute left, the last topic I want to mention is just the balance sheet management. You've got some large, longer-dated maturity walls. What's the idea to solve those? Is it that you'll have your revenue and EBITDA inflected ahead of time so then thereby the capital markets open up and then you'll refinance that debt? Is that the idea here? You do securitization on some of these PCF wins?
A couple of things. We would never securitize PCF because we already have the cash. That one does not make sense for us. The transaction with AT&T allows us to wipe out our most restrictive, most expensive debt. It will put our leverage down below four. With the cash savings in the coming next few years, we will be a lot closer to three. That will unlock what we believe is more ratings improvements, the ability to refinance what is there and get to a much more normalized maturity curve. I think the balance sheet is very soon to be a thing of our past. It is all about execution as we go forward.
Great. Thankfully. And with that, we're about out of time. So thank you, Chris.
Yep. Thanks a lot, Greg. Thanks.