Welcome, everyone. My name is Chris Schoell. I'm with the communications and media team here at UBS, and today we are pleased to have Dave Schaeffer, President and CEO of Cogent Communications. So Dave, thank you for being with us today.
Well, thank you for hosting me. Always thank UBS for a great venue, and most important, thank investors for taking time out of their day to spend with us.
Awesome. So given the timing of the conference, we always ask, maybe we can start off by just recapping the year and what are the priorities as you look out into 2024.
Well, definitely 2023 was an eventful year for Cogent. After an 18-year hiatus of not doing any M&A, we ended up acquiring the Sprint Global Markets Group or Sprint GMG, from T-Mobile. That transaction represented almost 80% of our revenue scale, and we acquired a business that definitely needed some additional focus and attention. It was a business that was doing $565 million in revenue, negative EBITDA of $300 million, and $30 million of CapEx. So a business that was burning nearly $1 million a day. You know, we closed that transaction in May 2023. We are in the midst of that integration, and while it sounds straightforward, it actually is two completely separate transactions that are occurring contemporaneously.
The first of those is the acquisition of an enterprise customer base that was basically independent of the network that Sprint had built originally to carry voice. 93% of the revenues were completely Off-Net. The customers purchased many products that were actually gross margin negative. We were paid $700 million by T-Mobile to take over that business, $350 million in the first 12 months, the next $350 million spread over the next 42 months, and then in months 55 through 58, a true-up on the working capital deficit. So we'll be receiving payments from T-Mobile for a period of 5 years. In acquiring that business, we had to end the life of 24 of the 28 products that were being sold.
We had to migrate as much of that traffic on-net as we possibly could, and we are targeting to bring about 50% of that traffic onto Cogent's network. The second part of that transaction was the repurposing of the physical network that Sprint had built. It was a network constructed between 1982 and 1989, at a capital cost of $20.5 billion. The network is comprised of 19,000 route miles of innercity fiber, 1,200 miles of metropolitan fiber, and 482 switch facilities that comprise 1.9 million sq ft. This network was virtually unutilized. It was built to carry long-distance voice, and while the network was maintained, there was little or no traffic on it. We purchased that asset for $1.
It was appraised by KPMG at $1.15 billion. As part of that acquisition, we were required to take an uneconomic capital lease that represented a liability of about $147 million. So effectively, we got $1 billion of net asset value for a dollar. Our challenge is to try to turn what was a liability of a network with no revenue into an asset. To do that, in 2023, we began the effort of implementing four key initiatives. One, extending the Sprint network to meet our metropolitan network in each of the major markets. Two, reconfiguring Cogent's metropolitan networks to segregate the carrier-neutral buildings where we will sell wavelengths from the multi-tenant office building footprint that we have. Three, to equip each of 800 data centers with a transponder shelf to ensure quick and seamless provisioning of wavelengths.
Then fourth, deploy reconfigurable add-drop modems or multiplexers at the endpoints of the Sprint network, where multiple degrees or directions of traffic come in. So we could be able to provision a wavelength within a two-week window. We're about 40% of the way through these reconfiguration projects. So a major focus for Cogent in 2023, and continuing into 2024, are these two initiatives: one, improving the profitability of the acquired enterprise business, and two, facilitating the repurposing of the acquired network.
As you think about stabilizing that base of enterprise business that came over, I know you're harmonizing the product set. You're probably exiting some of your unprofitable products. Where does that stand, and at what point do you feel like you can get the stability there?
So at the point of acquisition, we outlined to investors that this was a four-year project. We are slightly ahead of the schedule that we had laid out. We, I think, surpassed expectations on cost reduction and EBITDA in the acquired business in the first full quarter that we operated the business, Q3, and I think we'll continue to show improved progress. However, we acquired a number of take or pay contracts that have fixed term on them, so even if we complete the migration of the customers, we are left with those costs. We will not achieve that positive 20% margin on the acquired enterprise business until four years after the transaction closed.
As you think about leveraging the asset for new growth opportunities, Wavelength is something that you, you've talked about. Maybe just discuss the TAM there and what the go-to-market approach is, and how that might be similar or different to what you've done historically in, on the IP side of the business.
Yeah. So again, just to set the stage, Cogent operated prior to the acquisition a 61,000 route mile innercity fiber-optic network connected to 18,600 miles of metropolitan fiber. That network then touched 1 billion sq ft of multi-tenant office space in North America and 1,600 carrier-neutral data centers in 55 countries. That network carries approximately 25% of the world's internet traffic. It's a network that we built up over a 23-year period and are the largest carrier of internet transit, and we established a reputation of both high quality and low price, coupled with quick provisioning. In acquiring the physical assets of Sprint, we felt that the best way to monetize those assets was to repurpose them. Since there is no long distance demand for voice services, repurposing the network to sell optical transport services made the most sense.
In doing that, we're doing much the same thing we did 20 years earlier when we bought companies like PSINet and Allied Riser and Carrier1 and FirstMark, in that we are repurposing those assets. In looking at the network that we acquired, there are two key components. One being the physical fiber that I've just touched on, the second being the facilities that need to be converted to data centers. We are working rapidly in both of those areas. In both instances, we have substantial advantages over our competitors. Let's start with the fiber optic network. 90% of the routes are unique, meaning no other carrier shares those right of ways. Two, we serve all of the major city pair combinations that any customer would need. Third, the network was built by a single provider, not a roll-up.
So we have very accurate GIS, Geographic Information System, data that allows us to map the network with 1 meter accuracy, which is highly valued by customers. Fourth, by connecting the network to Cogent's metropolitan footprint, we can deliver services in 800 carrier-neutral data centers, more than any other provider. By standardizing the architecture, we can bring the provisioning windows down to comparable windows that we've achieved on selling transit. So when Cogent entered the transit market in 2001, the average price per megabit was $300 a megabit, and the average time to install was 90 days. Cogent came to market with a $10 per megabit price point and a guarantee of 17 days to provision. In fact, we have achieved an average of 9 days over that 20-year period in terms of actual time to provision, and the average price per megabit-...
has declined to about 10 cents. We have been able to capture market share by undercutting our competitors. We guarantee to sell transit at a 50% discount to other providers. That strategy has made Cogent the largest transit provider in the world. In the wavelength market, we have the qualitative attributes that I described earlier, and we have a cost basis of $1, so we will be aggressive if needed. We believe we will be able to capture 25% market share over a 7-year period. So the market today for wavelength services globally is approximately $7 billion. It's actually five times larger than the transit market. $3.5 billion of that is outside of the U.S., $3.5 billion in the U.S. Of the $3.5 billion in the U.S., we are targeting the $2 billion intercity market.
There is an additional $1.5 billion intra-city or metro wavelength market. While we will participate in that, it is not our primary focus. Today, many customers who have a data center to data center wavelength requirement may be forced to buy two independent wavelength products and cobble together a solution on their own. Because of the ubiquity of our metro footprint, we're going to be able to offer a single provider, end-to-end solution with that 17-day provisioning window. So I think we should be able to replicate the 25% market share that we've achieved in transit in a much shorter timeframe. We're targeting 7 years
In that innercity market, who are you predominantly competing against? I'm sure they're not standing still knowing that you're now targeting this space. What sort of competitive reactions are you anticipating there?
Listen, you should always be paranoid about your competitors and never be complacent. But our two competitors have multiple lines of business. One of the reasons why Cogent has survived in a very commoditized portion of telecom has been our maniacal focus on a limited product set and our insistence on being the low-cost producer, understanding what we sell as a commodity. So 25 years ago, the two wavelength providers were actually AT&T and MCI. They were the only companies that would sell transcontinental wavelengths. Fast forward, after the dotcom meltdown and the telecom crash, that market was primarily the purview of companies like Level 3, Global Crossing, Broadwing, Williams, and Qwest, all companies that are now part of Lumen. Lumen today dominates that market.
The second player in that market is Zayo, and then to a much smaller extent, you have some regional niche players like Uniti, Windstream, and Crown Castle. Cogent will have a nationwide footprint, more endpoints, and these unique routes. I think those attributes differentiate us and will allow us to compete. We have a sales force of 257 reps who focus on the wholesale market, larger than any other provider. We have existing relationships selling transit to three-quarters of the wavelength buyer universe. We intend to use those relationships coupled with price. You know, from what I can see from the outside, Lumen is more focused on selling enterprise managed solutions and technology packages and not dumb pipes. Just to remind investors, Cogent's motto is: "Smart people buy dumb pipes." That is our way of going to market.
Zayo is a little bit different in that Zayo has many different products. We're a large customer of both Lumen and Zayo, buying dark fiber from both. But in addition to dark fiber and wavelengths, they sell a suite of managed services. What will differentiate Cogent is by architecting the former GMG network to sell only one product, we can achieve operational efficiencies, and coupled with our cost basis, we, I think, can gain market share quickly. So you should always be worried about your competitors, but when they have both operational and balance sheet constraints, they sometimes cannot be as a virulent a competitor as they would have been otherwise.
Given the uniqueness of your routes, is Dark Fiber something that you would look to also sell yourselves?
It is. So, we are happy to sell dark fiber. We are happy to monetize our data center facilities by either selling them or leasing them on a wholesale basis to other data center operators. We, however, have to pick our priorities. Our first priority is to get these assets into marketable condition. Our second priority is then to begin to sell wavelengths. Our third priority is to turn down our 12,900-mile Lumen IRU that we prepaid for, but have a $15 million annual maintenance expense on, and roll that onto the acquired Sprint network. So once we've achieved those three major objectives, we will then look at selling dark fiber and data center space on a wholesale basis.
I think to be successful, you can't be too defocused, and you need to narrowly focus on a handful of objectives to keep your team focused and to actually be able to complete those projects. If you started selling dark fiber while you're in the midst of integration, I think it would be a huge distraction.
I recognize you probably view the business as one now, but maybe returning to you, the legacy corporate business. Obviously, that business saw a lot of change with the pandemic, and faced a number of headwinds. It's still working its way back. Maybe can you just touch upon what gives you confidence that that business can recover back to the growth levels that we had seen historically? And what's the timeline for us getting there?
Okay, so our corporate business is selling to end user businesses located in skyscrapers in the central business districts of major North American cities. Cogent has slightly over 1 billion sq ft in 1,860 buildings. That represents 11% of the total multi-tenant inventory in North America, but only 0.2% of the building count. In those buildings, we have a compelling value proposition. We deliver a service 9 times faster than our competitors. Once installed, it's 3 times more reliable based on our ring architecture, and then finally, it delivers 30-60 times the actual throughput. Our corporate business had grown organically for 18 years. Well, actually, I guess, probably 15 years pre-pandemic, at a compounded rate of 11% per year. When the pandemic hit, that growth rate went to -9% at the trough of the pandemic.
Today, we're back to positive 1%. So we're about halfway back from where we started to trough, back to where we began. That business is improving, but slowly. The return to office phenomena has been uneven and slower than many had predicted. You know, it's really been two and a half years since people were asked to go back to the office, and many companies are still struggling to bring employees back, and the exact amount of time in the office is still being figured out. Our corporate business has also been impacted by the geographic unevenness of this return to office. Southeast, we're pretty much back to pre-pandemic growth rates. In the Pacific Northwest, we're still in a negative growth rate environment. I think over time, companies are exerting additional leverage on employees, requiring them to come back to office.
As a result, we have seen the vacancy rates in our footprint start to decline. Now, it's not meaningful. Pre-pandemic, our footprint was running at 6% vacancy, 94% occupancy. That spiked up to 18% vacancy. Today, it's down to 17%, but it is improving. Secondly, the average new lease that's being signed is about 20% smaller than the lease that was in place. That is actually a positive, because it means that the total number of discrete businesses per building will actually increase. Third, we're seeing companies that had procrastinated on buying decisions now making those architectural changes to their networks and buying from us. So it is improving, both on a sequential and year-over-year basis, but it's probably going to be some time before we're back to an 11% grower, but we do firmly believe we will get there.
... In that market, have you seen anything different from a competitive standpoint, whether it be cable moving upmarket or I, I'd imagine fixed wireless doesn't really cater to the types of customers you deal with, but anything on the competitive front you would call out?
So let me take each of those independently. In the case of fixed wireless, not relevant in high-rise buildings, there are both reception issues and the nature of the tenants tend to have not shown interest in a fixed wireless product, and there's an inability to place any kind of antenna or receiving device on the exterior of the building. So fixed wireless has not been a factor. I do think fixed wireless will have a place in some suburban markets. In terms of cable, they have not generally been very competitive in our On-Net footprint for two discrete reasons: one, their plant typically is not in that neighborhood. They tend to be in areas where there are residential users. And two, they have not pre-wired the building. On average, it costs about $1,000 a floor to pull a cable or a piece of fiber to a customer.
The average building that Cogent connects to is 41 stories tall. So if you just took the midpoint at 20 floors, that means it would cost $20,000 and a non-recurring fee to turn up that customer. The cable companies haven't been willing to eat that. The customers are not willing to pay that. So we have not seen significant pressure from cable in our on-net footprint. Now, within Cogent's corporate segment, 20% of our connections and 40% of our revenues come from off-net services. These are services that began with an on-net relationship, and then the customer had a branch office in a location that was not large enough to justify our capital expenditure. In that footprint, we use last mile connectivity from ILEC, CLEC, and cable companies. In that footprint, we do see cable very aggressive.
These tend to be suburban, campus-type buildings with, you know, smaller buildings. Usually, you know, average building we connect to off-net is about 21,000 sq ft, compared to 553,000 feet on-net. The average building in the U.S., the average office building is only 11,000 feet, so they're slightly larger than average, but nothing like where we're at. So cable seems to be focused on those office parks, not CBD.
And then maybe pivoting to the NetCentric business. That was a business during the pandemic that outperformed expectations, grew faster than your historical run rate, and it's still growing faster. Maybe just walk us through what is still driving that outperformance and how sustainable you think that is as we go into next year.
So our NetCentric business is selling one product, bulk transit, and carrier-neutral data centers around the world. 55% of that business is outside of the U.S., about 45% U.S. and Canada. It is a business of metered services. We are selling megabits on a metered basis. We have approximately 7,900 access networks and about 5,000 content-generating customers. That business had averaged a 9% growth rate between 2005 and 2020. Going into the pandemic, we were actually below that average at 3% annualized growth rate. When the pandemic hit and people went home, there was an acceleration of the transition from linear video to streaming. So we went into the pandemic with 18% of video being streamed, 82% either linearly delivered through broadcast, cable, satellite, or DVD.
Today, we're at about 48% of all video consumption being streamed in the developed world. That pull forward accelerated the growth in that NetCentric business. It peaked at a 26% year-over-year growth rate. Today, it's at about 11%, so it's still running ahead of trend line, but not nearly as material. We are also seeing faster growth outside of the developed world than the U.S. and Western Europe. That is a positive for three reasons: one, we have a stronger presence in those markets, so gaining a bigger market share. Two, it is more likely that the traffic will be dual-sided, meaning we're paid by the sender and receiver. And third, we generally charge more in those less developed markets.
For all of those reasons, the outperformance of the NetCentric business has continued, and we expect it to continue at or near long-term trend lines for the foreseeable future.
... when you piece it all together, the three different businesses we talked about, I know you target about 5%-7% top line growth. Maybe you can just unpack that a bit as to what you assume for each of those areas, whether it be the acquired assets, the legacy corporate, and legacy NetCentric.
So let me start with a qualification. Cogent does not give quarterly or annual guidance. Second, when we announced the acquisition of GMG, we felt it was a requirement that we give investors some guideposts on what we were trying to achieve. So we gave some high-level, top-down guidance. We said, within five years, the combined company would be doing $1.5 billion in revenue and $500 million in EBITDA. That was not meant to be specific quarterly guidance, but rather an overall investment thesis on why the acquisition made sense for Cogent, which at the time was doing $600 million of revenue and $240 million of EBITDA. We issued no stock, we issued no debt. As further justification, we have answered the question you've just asked, which is a granular buildup, and that buildup has four components to it.
It has organic, classic Cogent. That business has averaged a 10% growth rate since going public in 2005. It was 11% going into the pandemic, it's about 5% today. That business is improving. Every investor will have a thesis on when it gets back to being that double-digit grower, or does it ever? The second component is the acquired enterprise business that was shrinking at about 7% a year and burning cash. We have said that we will stabilize that business, jettison these negative margin products, and turn it into a non-growth business of about $450 million, with 20% EBITDA margins and 8% capital intensity. The third opportunity is the wavelength business. The wavelengths carry very high incremental margins. They are On-Net. They carry 100% gross margin, $0.95 of incremental EBITDA.
We expect to gain market share and be 25% of that market within a 7-year period. If you add those three pieces together, you get to a higher revenue and EBITDA number. Just this last quarter, we reported revenues of $275 million, so just about $1.1 billion run rate. EBITDA margins are 47.7% or $133 million, so a run rate of $520 million. Now, that is artificially inflated by the payments from T-Mobile to subsidize the cash burn, but we're picking up those costs as well. The fact that we've been able to reduce the costs faster than originally projected has allowed our free cash flow to improve. I think over the next several years, investors should expect to see the T-Mobile payments decline as we've scheduled them out, the underlying performance of the three businesses improve.
Growth in the organic Cogent business, which carries a blended contribution margin in the 70s on-net and off-net, the wavelength business, which carries a 95% contribution margin, and then the enterprise business, which is not going to grow, with a static margin of-- cash flow margin of about 12%.
Maybe we'll just fit in one last question. Just from a capital, capital allocation standpoint, you've had a long history of dividend increases. How do you balance that with where your leverage sits today and investment levels going forward?
Our leverage is declining rapidly. Because of the improvement in EBITDA, we delevered on an LTM basis 0.5 turn last quarter, just sequentially, quarter-over-quarter. We expect that rate of improvement in leverage to continue as we get a full 12 months of reporting the T-Mobile acquisition and GMG numbers and our numbers. We have a policy to be between 2.5 and 3.5 times levered. We are slightly above that range today. We will be in that range or even below that range within a year. We have 45 sequential quarters of growing our dividend. We expect to be able to continue to do that. We also have the ability to add buybacks to our return of capital strategy. We have bought back approximately 22% of our outstanding float over time.
We have not used buybacks recently and mostly used dividends, but our expectation is we'll continue to increase the amount of free cash we return to shareholders every quarter.
Awesome. That was a great overview. Thank you very much-
Hey, thank you..
For being here with us again, and thank you everyone for joining.
Thank you very much. Thank you.