All right. Let's get started. Good morning. Welcome to day one of our 54th annual TD Cowen TMT Conference. My name's Gregory Williams. I cover cable, wireless, telco, and fiber here at TD Cowen. Joining me this session are Bill DiTullio, SVP of IR and Treasury, and Drew Smith, EVP of Strategic Finance at Uniti and was the CFO on the Kinetic side as well. Gentlemen, thanks for joining us.
Thank you.
Thank you.
While we don't expect the company to comment on unsubstantiated reports on M&A, maybe you can help us just think about value optimization. There's dozens of permutations and types of sales and types of JVs for your assets. Maybe we'll start, there's a potential buyer that has a consistent pattern of 50/50 JVs. I assume you'd prefer an outright sale. Maybe you can help us shed some light or anything on the table that optimizes shareholder value.
Yeah, I'll start with that. Thanks for having us again, Greg, this year. Obviously, there's only so much we can say on that. I don't want to speculate about any kind of rumors, but I think what you've seen us say and do in the past is we take a very open-minded approach to M&A. We have divested assets in the past where we've been able to realize significant value, both some that were non-core, but also could have been somewhat core, I would call. Listen, I think we are open to maximizing shareholder value. If M&A fits that in the near- term and is the plan that best suits us, then that's something that we may or may not pursue, right? Again, just trying to maximize shareholder value.
In that light, the buyers would naturally want just the growth assets. We've seen this in the past on fiber companies, not trying to buy the legacy stuff. Could Uniti contemplate keeping legacy DSL assets similar to what Lumen did with the AT&T deal? Would you prefer the other where a buyer takes the entire plant, including DSL?
Yeah, I can take that one. In a perfect world, the easiest transaction would be for someone to take all of Kinetic. All the fiber that we're building, all the copper that remains, even though it'll be a smaller portion of the footprint over time and all the operations. If the question is, would we be open to a transaction that really isolated on the fiber assets because they have a premium valuation? The answer is yes. If that was the best way to maximize shareholder value, we would definitely be open to that type of transaction.
Got it. Not talking about Uniti specifically, but valuations, both on the fiber to the home side and the commercial fiber side, how are they holding up considering the macro, the Fed, geopolitical risks these days?
Yeah, I think valuations have held up pretty well. If you recall, there was a slide we used to have in our earnings materials going back several quarters, where we did sum-of-the-parts valuation, right? You saw that we put a, based on not only M&A, but also ABS deals that have been done to date. At that time, we had put 10x-20x multiple on fiber assets. I think it was around a 5x-6x on DSL copper assets. Again, blended all in, fiber to the home, like Kinetic, we were showing a 6.5x-9.5x . I don't think those multiples have really moved much. If you look at precedent of more recent transactions, and even there's been more ABS transactions that have come to market, those leverage thresholds have held, right?
I still think you're going to get the upper end of that range if you have premium assets. I think when you look at both our fiber to the home and our commercial enterprise business, those are premium assets.
I want to talk about the non-core assets you guys have been putting in the deck recently. You mentioned $1 billion of non-core assets you can sell over the next, I think, 12-36 months. I guess it's unused fiber, maybe even DSL, real estate, even copper metal. Which assets are the low-hanging fruit here, and how would you contemplate the value of those assets in a potential sale of the company? Could the buyer be like, "I'm not interested in some of that?
I would say the low-hanging fruit is really all of those. I think the one that is probably the most obvious is just excess fiber. We also have the ability to monetize other assets like spectrum. That spectrum was acquired by Windstream about five to six years ago, prior to the Fiber to the Home initiative.
It's a spectrum that is not utilized significantly in our operations today. The other would be real estate. These assets really do not produce any material cash flows, in some cases, really no cash flows. We think our ability to monetize anything there really doesn't impact the long-term strategic options we would have.
You say excess fiber, but you do sell fiber through IRUs, so this would be routes that you would otherwise not do dark fiber IRUs or just overlapping? How does that work?
Yeah. These are really fiber assets that may be outside the major markets we operate. Maybe they were acquired through a variety of acquisitions done in the past that, in some cases, may be metro fibers that were built for other purposes and not the purposes we have. We're definitely building fiber, selling fiber, but there are portions of the network that are outside the core, and that's what we're talking about in this situation.
Got it. To the degree you can speak, what's your working relationship with Elliott as it relates to these potential sales? Any considerations outside of the typical process?
Elliott, they're a major shareholder of Uniti. They've been part of the story for several years. They were a major shareholder of Windstream as well. They've been in the story, and now they're a part of the Uniti shareholder list and also have influence of the Uniti board, and really it's business as usual. No real changes there over the last several years and also right now.
I want to switch gears and talk about the setup for the rest of the year. In the first quarter, Uniti handedly beat EBITDA, but you pre-messaged the one-time fiber sales. I think a lot of us didn't fully appreciate how big the one-time fiber sales would be, because you beat by about $30 million. Interestingly, you did note contemplating a raise of guidance above and beyond the one-time fiber sales.
Right.
You probably had some underlying bullish factors. Maybe you could talk about that. What did you see that made you contemplate raising guidance? Was it on the Kinetic side? Was it the fiber infra side, or both? Why didn't you ultimately pull the trigger then?
No. We got that question. I think, listen, we put out our initial guidance for 2026 in February, right? We got to 1Q earnings, that was only a couple of months later, and we give a range, right? We give a range for a reason, because we could come in somewhere in between that, and we try to manage to the midpoint. Our results in the 1Q weren't a surprise to us, right? I think Kenny did mention that we did see some outperformance in both Kinetic and at fiber infrastructure. We are seeing good momentum there. We are making good progress there. The biggest fluctuation or the biggest impact to our guidance this year is from these, we call sales type lease agreements, right? These hyperscaler deals that are being treated as one-time revenue.
Back in the fourth quarter earnings, we did foreshadow that we expected that to be bookended for the year in 2026. We expected about roughly half the hit in 1Q, which it did, and we said the remainder should hit later this year, most likely in the fourth quarter. When we gave out, when we had our earnings call this past quarter for 1Q, we did provide quarterly guidance to help provide more color on that timing. Again, when we originally gave our guidance, we said that the difference between 2026 and 2025 at fiber infrastructure was expected to be about $145 million, right? About $70 million of that revenue hit in 1Q related to those hyperscaler deals. The remaining half, again, should hit later this year.
There is, like anything else, and we've seen this in the past, there is potential risk that some of that could slip, especially since it's not supposed to hit.
Till later this year. There could be a possibility that it slips into 2027. There's also a possibility that some stuff could get pulled.
Forward.
forward. Given that we only had two months under our belt since we initially gave that initial guidance, we decided to keep our outlook for 2026 as is, keeping the ranges. I think when we get to 2Q earnings, we'll have greater visibility of how 2026 will play out.
Right.
If we think there's any potential risk for that slipping. I think we'll just update it quarterly then.
Is that milestone driven or customers maybe not wanting to take the orders because maybe the data centers they're building?
There's always a build component to some of these too, Greg, right? You recall back when we were doing our large build-outs for the wireless carriers, there could be permitting delays. There's weather delays, although we don't like to blame it on weather and stuff like that. There can be delays along the way. Sometimes it could be the customer's just not ready to take delivery of the route for whatever reason as well. Because of all those, that confluence of factors, we decided we wanted to get another quarter under our belt to kind of see, okay, where do we stand on these deals? If we have, again, we will have greater visibility then, and I think we can make a more informed decision.
Got it. I'm curious, why, from an accounting perspective, do you recognize the sales upfront versus, I think all the peers that we covered over the last 10 years or 20 years, amortize it over the 20-year IRU, for example?
Yeah. The short answer is we're simply following GAAP. When you're following GAAP for these type of transactions, you're really evaluating does it fall under an operating lease, which an operating lease would take any upfront payments and amortize that over the life of the contract, or a sales type lease. Again, you look at certain criteria. I can tell you historically, I think Uniti, the transactions they were doing were more the operating lease treatment. The transactions that we did at Windstream were sales type, and it wasn't that we could choose one or the other, it was simply that we were following GAAP. As we evaluate the deals we're talking about right now, they're falling under that sales type treatment.
Okay. Moving on to Kinetic and Windstream. I thought you guys knocked the cover off the ball in some ways. Solid results, record gross adds, I think it was like 39,000.
Even fiber churn was a record low at 1.45%, down 24 basis points. You mentioned a few drivers there. Customer care, agent tools, outage alerts, communication with late payers. I guess the list goes on, there's still more to go. With all those improvements, where could fiber churn ultimately level out at?
We were very excited with our fiber churn trends. If you look year-over-year, first quarter 2026 versus first quarter 2025, we were down around 25 basis points. For all the reasons that I think John Horobin said on our earnings call and all the reasons you just stated, Greg. The other thing we're seeing is when you look at early life cycle churn, we're seeing improvements there as well. We think that's a good leading indicator of how we think the future will play out. There's always a little bit of seasonality. Typically, the second quarter and third quarter can be a little bit elevated when you compare that to other quarters. We could see some seasonality, but if you look year-over-year, we're still expecting to achieve around 25- 30 basis points improvement in the future.
We think that the mechanisms and insights we're putting in front of the teams are going to allow us to continue to see good trends on churn.
Okay. You'll still see 25-30 bp improvement probably year-over-year over the balance of the year.
That's correct.
We'll see some seasonality natural from the second quarter and the third quarter that we've seen in the past. You did have price hikes, I believe, in January and February. Could we see a small churn bubble from that, or did those price hikes already sort of take in a reflection?
I would say there may be a little bit of a bubble. Well, I don't think it'll be significant. I think a lot of that was in the base that you saw at the end of the quarter.
Got it. Then moving on to ARPU, another good news story. That was up impressively, 5% year- over- year. In late March, we hosted John Horobin. He spoke with us and said, "Don't get too excited. I guess full year ARPU might settle in the 2%- 3% range." Is this still the case even after the solid first quarter you just had, and even with the January and February price hikes?
Yeah, that is still the case. We feel like that our expectations should be churn in the low single digits, 2%-3%. Churn has been higher at Kinetic in the past. I think you're seeing a little bit of that in the first quarter, just year-over-year. We think the growth rate of the 2%-3% is more sustainable. A good example of that is the churn improvements we've seen. Just, I think, a more surgical approach to our pricing is helping with that. Long- term, we do see growth opportunities on churn. That's why we're talking about 2%-3% annual growth. Two things are really driving that. We're seeing a higher adoption of multi-gig at the point of sale.
In 2025, we made investments to increase our max speed from one gig to two gig in about 90% of our footprint. That's going to naturally allow us to sell more speed at the point of acquisition. We still have opportunities in our base. Our fiber customer base today, only about 50% of those customers are on a multi-gig solution. We have an ability to continue to upsell at point of acquisition, but also within the customer base.
Got it. We're hearing a lot of talk on LEO satellites. The company previously mentioned that it sees LEO churn largely in the copper footprint, which makes sense. It's able to recapture some of those customers once you reach them with fiber. I'm curious about the win-back rate here. When you build fiber to those areas, generally, what percentage of the copper churn do you end up recovering after LEO took the DSL subs and then you would, in turn, take them back?
Yeah. That's right. LEO, we see that be a competitive threat, predominantly in copper, not as much in fiber. On your question on win backs, I think one example I can give you is that when we're building in a footprint that really doesn't have a wireline provider there today, so we're building fiber, we're all that there is from a wireline, and so your competitive landscape there is LEO and probably fixed wireless. We're achieving above 50% penetration. A lot of that's in the first year, I think that's the best way you can think about win backs when we're building in those situations.
Do you think the terminal penetration now has to come to a different marker because LEO's now in the mix? It could be 45%-50%, but now with LEO, maybe ratchets that down. Is that a safe assumption?
I think when we talk about our 40%+ terminal penetration, I think we are accounting for LEO. If you think they're going to have a decent share of the footprint, that still leaves us versus cable in the majority of our footprint. 75% of our footprint is one competitor or less, one wireline competitor or less. That's why we think the 40%+ is a good number for us and possibly even conservative when you think about fiber being a superior product. It has the ability to scale, I think, more than the other products we're talking about.
Got it. Would the LEO threat drive an urgency to build faster, or earlier, to reach your goals?
I think we have an urgency to build fast. That was one of the key things that we talked about when Uniti and Windstream came back together, our ability to scale the fiber build quickly. If you think about our guidance this year on fiber households constructed, 450,000- 500,000, that's 2x what we did in 2025. We're definitely focused on scaling quickly. LEO's a reason for it, but also, staying ahead of others who may want to come into our markets and also continue to expand against our cable competitors.
Okay. Is there a business case to build beyond the 3.5 million home target, protect those remaining DSL subscribers? Again, now that maybe you built a business case in the past, but now it's like, okay, well, LEO's in the mix, so maybe our harvesting of DSL is going to be accelerated. Maybe we'd go beyond 3.5 million, or are they just maybe too costly to go out into those rural areas?
Right now, 3.5 million is the target, but we've been open in the past saying that there could be more opportunities. One driver could be additional subsidy builds. Kinetic has been active in RDOF. Kinetic was active in public-private partnerships, and we've also announced some BEAD wins. That could be a way to do it. I think when you look at how to build the rest of the footprint, the economic model, if there's increased expectations in penetration, could move more households into the build opportunity, or a lower cost of capital. We've seen that a little bit with our January ABS transaction on the Kinetic side, where we achieved a pretty good yield on that transaction.
A good segue to that is the cost to build. We ask this all the time about updates on cost per passing. I think Kenny even mentioned on the call. Cost of resin's up, which feeds the conduit. We've seen things like fiber up 75% year-over-year, the actual glass itself, routers, optical equipment, the DRAM that supplies a lot of the equipment. I'm curious, are you seeing any risk to your, I think it's $900 or $1,000 cost per passing at this point?
We've got it to, when you think about the incremental build of going from 1.9 million households today to 3.5, we've got it to a cost to pass of $900-$1,000. When you do all the math of what was the total bill, what was the cost to pass? That would put everything around $800, $900. We're pretty confident in that. We've done things to mitigate cost increases. Historically, Kinetic had leveraged an internal construction team. We still have that available to us today. I know one focus with John Horobin and his team was to really scale our relationships with external contractors, and we've done that over the past 12 months, and that helps us secure our costs for that build.
From a fiber perspective, we've worked with multiple fiber providers to have the right allocations, and all those things we believe can help mitigate against future cost increases and keep us in that 900-1,000 range.
Got it. On copper decommissioning, your peers, AT&T and Verizon, they've spoken emphatically about the huge copper decommissioning savings. What's the opportunity for Kinetic? It sounds like they're doing all the regulatory work to pave a wide road for you guys, too.
There's definitely opportunity for Kinetic, and we've been pleased with how the regulatory environment has given more flexibility to that. We, I think, have been talking about copper decommission for a couple of years but really started to ramp the program in 2025. We had kind of our first market that was fully fiberized. We're doing more markets in 2026. I think naturally, it's going to scale as we do the fiber build, so as we get more and more fiber coverage across our entire footprint. The opportunity is there. From a cost-savings perspective, we see today that the cost to maintain fiber versus copper is a lot less, fewer trouble tickets, fewer dispatches, fewer calls into your customer care, and those ultimately will drive reduced costs and higher margins in the future.
I think another opportunity is more on the infrastructure side when you look at your energy cost and utility cost, when you've got that copper network and distribution cable, but you also have the copper equipment serving those customers, and as you decommission those, you're able to really reduce your energy costs and utility costs.
Sure.
I think another just real quick on that, too, is recall years ago, several years ago, Windstream spent considerable CapEx to build out fiber to the node, right, 95%.
Right.
It's really just that last mile that's copper served, too, that needs to be replaced. It's not like you're replacing it throughout the whole network. It's really just from the node.
Right. They did a lot of legwork early on the fiber to the node to help out.
That's right.
Sure. That's a good point. With the remaining nine or 10 minutes, I want to talk about the fiber infrastructure segments. You're clearly excited about the inference phase and how this plays into fiber infrastructure. I think you noted four to $500 million in recurring revenue.
Which is important, by 2030 and beyond. What have you seen to show that the market is beginning to move towards inference? I know it's hard to see whether this wave is used for inference versus training.
Right.
What are you seeing? What sort of inning are we in as we sort of shift into inference from training?
I think we have a slide that kind of shows the build-out phase, right? There's transition to this lease-up phase, and I think lease-up is where you're going to see inference come. The build-out phase for us, and again, we're building out using existing either conduit or within our network, we're building out these large fiber networks for the hyperscalers. We're putting in 432 strand, 864 strand, even more than that for these hyperscalers. We're putting in multiple conduit systems in these builds, right? One or two conduit may be dedicated to the customer, the hyperscaler customer, and then there'll be additional conduit we can use for lease-up, and that's really to prep us for the next phase, which we do think inference is coming.
I think we're still in what we call the training phase, where these large models are still being built out, and that's where we're seeing a lot of the demand come from. We're in the very early innings, maybe even before the game's even started, in terms of inference. What gives us confidence in that is we have seen, A, once the AI starts to build up where enterprises or schools or healthcare systems, things like that, start to adopt AI more robustly, they're going to need large, dense, low-latency networks to help serve that, right? When you look at our markets that we're in and where we're building out these large pipes today, it's in our southeast footprint.
Again, we are starting to see some customers coming to us to start to ask for some of this stuff, but I would say it's still very early stages. We're still primarily focused on the training phase.
Got it. It's nice to hear that you're building this. Is it empty conduits, then? Sort of like what Level 3 did 20 years ago? Dig once philosophy?
Yeah.
It is.
When you do any construction job, right? The biggest cost is the actual digging.
Yeah.
Construction itself. We're trying to, we've used the term, quote-unquote, "Future-proof these networks." Yeah, to your point, Greg, we're putting in multiple conduit, and the technology on the fiber is ever-changing, right? Now there's microducts where yeah, microducts, you can put multiple ducts through one conduit, and serve customers that way. We are putting additional conduit, too, that we can easily pull fiber through. We may not put all that fiber in the ground day one, but if we need additional capacity, you can very easily pull through additional fiber, at minimal cost.
Sure. The company, again, says that inference is exciting as we believe as well, much lower CapEx, more lease-ups, that's higher margin. On the flip side, there's typically far more competition because inference is mostly going to be like in availability zones in major metro markets. You'll see the usual cast of characters competing against you. While the TAM could be big, could we see price risk with more competition in those markets?
I think in the markets that we're in, we're primarily in Tier 2, Tier 3 markets, right? They aren't highly competitive. There's only a handful of providers today that probably have dark fiber or could put the dark fiber in the ground to serve those markets. Again, we're not trying to be everything to everyone. We're not going everywhere. We're really focused in our southeast footprint where we have existing assets today. As we've said in the past, about 80% of these deals that we're winning or bidding on, 80% of these deals utilize existing assets. It's not utilizing existing fiber because we don't have 432 strand or 864 strand in the ground today, but it's utilizing existing right-of-way or utilizing existing conduit.
Which should, in theory, make you the price leader if you have existing conduit versus somebody.
Yeah. It should allow us to, put it a little different way, I would say it allows us to win our fair share, if not more, because we are well-positioned. Especially with the hyperscalers and what they want to do, time is money to them. They want things put quickly, and they're willing to pay for that.
Since we have existing assets in these areas, that does give us an advantage over somebody that would have to come in and do a true greenfield build, because not only will that take more CapEx, but it also will take time. Right? You just can't get away from that. I think we're well-positioned there. In terms of inference and, again, you mentioned pricing power, I wouldn't say all buyers are basing their decisions on price. Right? They do value uniqueness of the route. Some cases, they want redundancy. They'll buy from multiple providers, and we've seen that. We've seen customers buy from us and from others on the same route or very similar route because they want that redundancy. They want that protection.
Again, we've been talking more on just putting in these large pipes, dark fiber deals, but we're also seeing a pickup too on the wave side. I think when you get into inference, you're going to see more and more of that from enterprises and healthcare and schools and government take more lit services, I think, because not every enterprise is going to be positioned to kind of manage the network themselves. They just want a dedicated wave, a pipe that's lit and managed for them that provides them the capacity that they need. I think, again, we're well-positioned when that starts to come.
Got it. You mentioned the words fair share. Maybe to put you on the spot here, what market share do you think you can get of the waves business? Lumen and Zayo are roughly half the business today. Cogent says they can get to 25% of the market. Where do you guys shake out? Generally, it sounds like you can more dominate the Southeast and maybe not so much in under index other markets.
Well, I think when it comes to waves, when I was talking about the Southeast, that's where I'm talking more about dark fiber.
Okay.
Waves, we have a national capability there. We announced on our last earnings call FastWaves, which is our new initiative, and we have pre-lit some of our 50+ most popular routes. If you look at that map, that is nationwide. Right? Again, there are only a handful of providers today that can offer lit capacity or lit waves. We're definitely one of them. Our market share today is less than 5%. I think when you look at the major players, we think fair market share is somewhere in the 10%-20%. Right? I think we'd be very happy with that. Waves is not a meaningful part of our business today. It's definitely an increasing part, and we are seeing larger orders.
In fact, we announced our largest bandwidth order on our last earnings call that we had a customer, a neocloud customer, take 20 terabits of data of capacity. That was 50 400-gig waves .
It wasn't that long ago when customers were talking about maybe they wanted one, maybe 10 gigs was a lot.
Right.
Now you're starting at multiples of 100- 400.
It's like millions of T1s back in the day.
Yeah. I think it's definitely going to go up from there. We're just cracking the surface there. With this FastWaves initiative, again, we've kind of pre-provisioned some of our most popular routes so we can turn it up quickly. Okay. Under our SLAs, we commit to 21 days, but in reality, we can do it in 10- 14 days.
Okay.
Under two weeks.
On the flip side of provisioning so quickly, are you seeing any customers actually delaying their orders? Maybe the data center isn't built yet, the chips aren't in there, the power's not ready, and maybe they say, "We don't need it so quickly." Are you seeing any delays from customers?
We haven't really seen that, Greg. We haven't seen any kind of delays. Obviously, there have been, on the supply side, on the equipment, there's been longer lead times, but we've managed around that. We do buy in advance. We buy in bulk. We have agreements in place with our vendors, so we have availability of that. I would think for the demand that we're seeing or we expect to see over this year and next year, I think we're well-positioned, and we're already starting to plan for 2027 and beyond.
Got it. Maybe my last question is on the edge, since you guys are deep into the fiber world, and we just got back from Connect (X) a couple of weeks ago, and there was so much talk about the edge again. It almost felt like the 5G edge hype, which didn't really pan out, to be fair. So we were skeptical. Do enterprises really need to spend, or customers need to spend such a heavy CapEx just to shave off a few milliseconds? We're just trying to have that debate. I'd be curious to hear your thoughts on, is AI finally the use case for the edge?
I think it is a use case for the edge, and I think it is coming. That's, again, we think inference is going to spread out more from these larger, longer routes. Again, most of the routes we're building are within our southeast footprint, but some are a town that you may not even heard of, you had to look up on a Google Maps, but usually it's connecting to a more prominent area you would know. Then I think from there, it's going to spread out more to the edge, and that's where the enterprise use case is. Listen, I think at the end of the day, it depends on the enterprise and what they're used for. If you're a high-frequency trader, yeah, milliseconds count. Right? If you're using AI to help do your investment thesis and things like that, those things matter.
I would argue a hospital system or milliseconds can count in terms of getting data. I think it really depends on the enterprise. Regardless of how it shapes, Greg, I think the takeaway is that Uniti is well-positioned, given our enterprise focus, given our Tier 2, 3 markets, that there's a lot of opportunity for us there. As Kenny has said, we do think the inference phase excites us more, even what we're seeing today in terms of training.
Great. With that, we're about out of time. Thanks, gentlemen.
Thanks, Greg.
Thank you.