All right. Well, welcome everyone to our next session this afternoon. It is my pleasure to welcome you to the first-ever version of the Communacopia + Technology Conference, a long time participant in Communacopia Classic, Dave Schaeffer, the Chairman and CEO of Cogent Communications. Dave, thanks for being with us.
Hey, Brett, thank you for hosting me. I'd like to thank all the investors for their time, and as always, thank Goldman for a great venue.
It's always exciting when right before the conference, you give us something new to talk about. On September seventh, you announced that you are acquiring T-Mobile's wireline business, representing your first acquisition in 16 years. I thought maybe just to get started at a high level for the people here, just give us an overview. You know, what are you acquiring? What are the assets, the revenue mix, the product offering? How is this transforming Cogent?
Yeah. Thanks for the question, Brett. We are acquiring what was formerly known as Sprint Global Markets Group. This is the nationwide long distance network that was built by Sprint in the late 1980s and early 1990s to support their long distance business. That is approximately 19,000 mi of owned fiber along unique rights of way, intercity, another 1,300 mi of metropolitan fiber, and then finally, about 16,800 miles of IRU dark fiber. In addition to the physical network, we're acquiring 1.3 million sq ft of fee simple owned real estate technical space, about 150 megawatts of power spread across 1,300 total locations. Forty-seven of them are significant locations that can be converted into colo facilities. We're acquiring 1,400 large enterprise customers.
Today, those customers are generating about $560 million run rate in revenue. We've identified approximately $110 million of that revenue as non-strategic and non-core, and T-Mobile has begun the process of end-of-lifing those products for those customers. At closing, we anticipate having 1,400 customers and approximately $450 million in revenues.
Why are you doing the deal? What does this ultimately do for Cogent? Because you're buying into business that, as of right now, is declining and not generating EBITDA.
We're being paid to acquire that business. You know, first of all, to remind investors, Cogent has done a number of previous acquisitions, 13 acquisitions in the 2001 through 2004 timeframe that allowed us to build our network, buying companies as large as PSINet, Allied Riser, which was actually a Goldman IPO, back in its day, and dismantling those businesses, repurposing those assets into the Cogent business model. We actually view this transaction much the same way, taking an asset that is dramatically underutilized and repurposing it for modern products and serving customers with a better product mix going forward. Cogent today has built its network 61,000 route mi of inter-city fiber and 18,000 mi of intra-city fiber through IRUs from 303 different underlying suppliers.
What this transaction allows us to do is exit one of those large IRUs for about 14,000 mi of intercity fiber, allows Cogent to save about $15 million in maintenance on that fiber. It also allows Cogent to take these assets and repurpose them for the internet, allowing us to take surplus capacity and begin selling wavelength services, an adjacent market that Cogent does not participate in today. It's about a $2 billion total addressable market for North America. Sprint today has, or T-Mobile has, less than $8 million run rate in that market. With this cost basis and our technology, we think we can gain market share quickly, and we can also enter the market of selling dark fiber.
We reduce our costs, gain scale, gain a new customer base, large enterprise, which we have not really focused on previously, and we've reutilized assets far more efficiently.
If I think about the historic profile of Cogent with the investor community, there were a couple of things that were really quite differentiated between your business and essentially every other telecom business that we look at. Very simple, in terms of the end markets and the products that you were targeting. Overwhelmingly transparent for an enterprise-focused telco. Of course, giving more than zero information makes you transparent in this sector, but you guys have been pretty good about it. No headwinds related to legacy revenue streams that are in structural decline. At the moment when you take on this new business, some of those attributes aren't gonna be present across the whole company anymore because you're buying a business that does have some legacy issues associated with it, maybe as it operates today isn't quite as transparent.
A question that we've already gotten is, what's the path to getting Cogent back to a point where with this infrastructure, you once again have many of those same attributes, that you've historically had? Because it seems like that's a very big and important part of how people think about the valuation framework for your business versus most other telcos.
Let me start with something that's an easy commitment. I can assure investors that we will be as transparent going forward as we historically have been. Second, we are jettisoning a large group of products that are non-strategic, and quite honestly, I would have difficulty understanding or explaining. Most of those products are gross margin negative products. Going forward, we are picking up two major groups of revenue from the T-Mobile wireline business. The first being dedicated internet access services and transit to large enterprises. That is roughly about $220 million of the $450 million of revenue. For those customers, they will see a value proposition, not all different than what Cogent does today. To preserve those customers, we intend to immediately increase their throughput by 10x and keep their revenue approximately flat.
Does that align up with how you typically would be pricing that amount of connectivity in the market? That's the idea.
That's correct. It would line up with Cogent's current pricing model, replacing the customers by effectively increasing the throughput. To give you a sense of scale, the MPLS and IP traffic that is being carried on the Sprint network today is less than 1% of the amount of traffic Cogent carries on a typical day. What we will do is migrate all of that service revenue onto the current Cogent network. The second major group of revenue comes from selling VPN services based on MPLS, a technology that is end-of-life and in decline. Cogent has developed technology that it uses today regularly for our customers to migrate them from MPLS to VPLS without a disruption in service. It also allows the customer to remove the piece of customer premises equipment that they had on site to support that MPLS circuit.
We will do the same thing for these customers actually during the transition period. While we are not taking over operational control, we, as part of our agreement with T-Mobile, are providing this technology to help them migrate the customers. What the customer will see is two things, a transition to a more modern platform, and a similar 10x increase in throughput for the same price point. That revenue stream is also about $220 million. The remaining $10 million, $8 million is in a nascent wavelength business that we will end up growing rapidly off of the very small base in place. Finally, there's only about $1 million in colocation revenue across that 1.3 million sq ft.
While colocation is not a primary business at Cogent, it still represents about 3% of our revenues, or about $18 million a year, and we expect to grow that within the combined company. Coming out of the transaction, we expect organic Cogent's SMB customer base to continue to grow at its historic growth rates. Now, we have been growing at half of that growth rate due to the pandemic and the impact in Cogent's corporate business. What had been a 10% organic growing business for Cogent for 14 years turned into a 5% growing business due to the fact that our corporate business, which is heavily exposed to central business districts, has grown more slowly. It's basically declined during the pandemic and now back to flat. Our net-centric business has continued to grow and actually accelerated through the pandemic.
We expect that net-centric growth rate to eventually revert to long-term averages of about 9% growth from the 16% today. You then layer on the $450 million of revenue that we're acquiring from these 1,400 large enterprise customers, DIA and virtual private network services, and we expect a couple % growth in that segment of our business. Then finally, the fastest growth, but from the smallest base is our wavelength business and our ability to sell dark fiber. The total product set for Cogent is going from three products to five. All adjacent to our core strategy. We also know that Sprint has historically underinvested, and T-Mobile has continued that with the sales efforts. Of the 1,320 employees in that business, only 60 are in sales.
Compare that to Cogent, which has 1,050 employees, of which 750 are in sales. We intend immediately upon closing to utilize our sales model and our sales force to help rapidly grow these product sets. The combined company off of a larger base should be able to grow at between 5% and 7% going forward. We will not be burdened with the legacy product mix of a historic telco, but rather this relatively concentrated portfolio of five products.
What will give you confidence? Because basically what you're saying is, by the time the deal closes, which probably is later next year, just based on the regulatory timeframe you've outlined, that you'll be at this $450 million run rate and you'll have transitioned all of those revenues to things that are structurally growing products.
Rapidly growing.
Rapidly growing. That's a lot to get done. It sounds like you're relying on T-Mobile using your technology to accomplish a lot of it. Why do you have confidence that the business will be positioned that way nine to whatever, 15 months out or whatever that regulatory time period is?
First of all, we are not taking control of the business-
Yeah.
Until we get final regulatory approval. What we have done is entered into commercial relationships with T-Mobile, where we can provide resale service, sales help, and technology help. Secondly, they are continuing to burn cash in this business, and they're motivated economically to reduce that cash burn as quickly as possible. Now, from T-Mobile's perspective, they had three primary objectives. The biggest one being they wanted no damage to their brand, and they wanted to treat customers and employees fairly. We're committed to that. They're committed to that during this period. Secondly, we're looking to preserve the revenues that make sense, and it is these two major groups. We think that the vast majority of the end-of-life products will have been terminated, maybe not all by closing, but that means revenue will be slightly higher than the $450 million.
Secondly, of the revenues we wanna keep, we think the vast majority will be transitioned. Many of these customers have had decades-long relationships with Sprint and now T-Mobile. Many of them have actually gone through a previous transition from Frame Relay to MPLS 20 years ago and are now just going through a second transition to VPLS. We intend to keep that legacy sales force in place that many of them have these long-term relationships. The average tenure of an employee that we're acquiring is over 20 years. The most tenured employee still at T-Mobile, formerly Sprint, has actually 49 years of service at the company. There is a lot of institutional knowledge and relationships.
Now it's true, some of those employees may not survive ultimately through an integrated company, but our goal is, and T-Mobile's goal is to use those employees effectively to reduce burn and to ease any disruption to customers in their transition.
You said on your conference call that the business is generating negative EBITDA today, and you expect it to have less revenue by the time that it closes, and you want to increase the amount, the size of the sales force. That's an increase in spending on sales. How do you think about the path to generating positive EBITDA? Where are you able to get costs out? Is it not about getting costs out, it's about getting onto a growth curve?
It's actually a little bit of both. Today the business is burning between $280 million and $300 million of negative EBITDA. By end-of-lifing these less efficient products and by utilizing the technology that we are making available, we think we can get that burn rate down to, at closing, about $180 million of negative EBITDA. We are getting a series of payments from T-Mobile. They are entering into a take-or-pay contract for on-net transit services, which carry 100% gross margin. At this point in time, they may not even utilize those services. In the first year post-closing, they are committed contractually to paying us $350 million or approximately $29 million a month.
During that period, when we're burning $180 million of EBITDA, now this is absent the $350 million-dollar contribution from T-Mobile, we will also need to spend about $50 million of one-time expenses to physically integrate the networks together and an additional $30 million in CapEx. Our determination of the payments from T-Mobile were based on Cogent's requirement that at the end of every month, we were confident we were in a better position post-transaction than we would have been if we had not done the transaction. For the first 12 months, we get $29 million a month. Pure margin. Then that steps down to $9 million a month or $100 million a year for the next 42 months, for a total subsidy period of 54 months.
We think that the majority of the negative margin, gross margin, negative customer products will have been eliminated. We'll complete that work during that transition. We'll also get the benefit of reducing costs within the Cogent network and within the acquired network. We will align the teams and come to the right headcount, and we will also be able to deliver a better product to customers by focusing on modern technology. We think that in year one, we'll be at - $180 million. By year two, that negative burn will be down to somewhere between $70 million and $80 million, and we'll be at EBITDA breakeven in the acquired business by year three.
Exclusive of the T-Mobile payment.
Exclusive.
Basically, by the time that you're done receiving them, the business will be breakeven or better on its own.
Yeah, we'll be better off long before that. The payment stream from T-Mobile is asymmetrically loaded to the front end, and we will have received $550 million of subsidy and sustained probably $250 million of negative EBITDA during that period. The idea is that we're solving a problem for T-Mobile. We're taking advantage of the opportunity for Cogent to drive down costs as well.
Okay. A hallmark of the investment returns that you deliver to your shareholders is this incredible track record of sequentially raising your dividend every quarter. You have said that remains the intent and the expectation for Cogent. Listening to how you talk about the contributors to growth for the business, looking ahead, it seems like you expect the existing Cogent businesses to ultimately revert to the mean in terms of their historical growth rates. Those historical growth rates, I think, would have been sufficient to keep you on the dividend trajectory that you've been on. What is it about doing this acquisition that gives you confidence you're at least as well off, if not better off, in terms of what you can deliver to shareholders versus just having kept things simple?
Yeah. Cogent has bought back 22% of its outstanding float and has now grown its dividend sequentially for 40 sequential quarters, quarter over quarter. We expect that we will be able to continue into the foreseeable future to grow that dividend. Prior to this acquisition, we noted on our last earnings call and have reiterated at investor conferences the fact that our aggregate growth rate has slowed for organic Cogent from 10% to 5%. With that slowdown in growth, our rate of annual EBITDA margin expansion has gone down from about 200 basis points to 100 basis points, with pretty much direct linear contribution. Our dividend growth rate today is at about 12%, 12.5%, growing $0.025 per quarter sequentially, and our cash flow growth is between 5% and 6%.
Eventually, those two numbers need to come into alignment. We have allowed our net leverage to tick up to 3.7 x, in part by our desire to disgorge excess cash on the balance sheet. That still remains a goal of Cogent, but we also are recognizing the reality that the return to office and reoccupancy of offices in central business districts has gone slower and taken longer than we expected. With that, absent this transaction, what we have tried to message is that we are absolutely committed and feel confident we can grow the dividend sequentially, but the pacing of that dividend growth may be adjusted. As we look at this transaction, we actually see it increasing the likelihood of faster dividend growth over a medium to long term. Why? Because we have greater scale, and we have the ability to improve margins off of that greater scale.
What we have guided to is a combined company growth rate of 5%-7% annually, all organic, because there'll be no more acquisitions embedded in those numbers, and we anticipate about 100 basis points a year of margin expansion. We think in five years, by 2028, we will be at a $1.5 billion run rate, EBITDA margins in the low-to-mid 30s. Compare that to the 39.4% that Cogent organic did last quarter, and we will be able to grow the cash flow at better than $0.025 per share. In the interim, if we see a recovery in Cogent's corporate segment, we can maintain our growth rate. If our corporate segment continues to grow at a slower rate, which is what we're seeing now, we will probably adjust that growth rate.
In the dividend?
Yes. Growth rate in the dividend, but still committed to sequential growth in dividend each and every quarter.
How long could you remain below trend before the board would say, "Listen, we don't know what's gonna happen. We've reached a point where we probably do need to have a slower dividend growth rate, at least for now."?
It's actually something that the board debates. Every board meeting is probably the number one topic that the board focuses on. While we have enough cash on the balance sheet to keep the dividend growth as it is with the current growth rate, probably for five years, it's not clear that's the prudent decision to make.
Is it something that you think you need to have a view on before you close this acquisition?
You know, it's something that we'll evaluate. I think it's much more dependent on the re-acceleration in the corporate segment of our business than it is the acquisition. We were extremely encouraged with the pace of rebound in our corporate business, Q4 to Q1 of this year. We were then disappointed from Q1 to Q2, and that rate of improvement did not continue. It basically stayed flat. In surveying customers, we're definitely seeing an improvement in the corporate environment in terms of proposals, spoke to's, and orders, but we are seeing a bifurcation in the corporate customer base. There's a segment of the market that's ready to put the pandemic behind them and re-architect their networks, much like Goldman Sachs is ready to put the pandemic behind it.
If you walk around the streets of San Francisco, it's a very different place than the streets of New York. In a sense, it feels like a ghost town. I think you're seeing a different segment of the marketplace saying, "We're not sure. There may be another surge. We may go to a greater percentage hybrid. We're gonna take a wait and see." With that, we're just trying to be realistic that the rebound in corporate growth, while it will get back to that 2% sequential growth rate, is not in a place where I'm prepared to say it's only a couple of quarters away.
Okay. Dave is willing to take questions from the audience. We have a few minutes left, so I think we've got one down here.
You mentioned going from three to five core products. How do we look at that in terms of growth and margin contribution going forward?
I'm just gonna repeat that just for the webcast. The question was, you're talking about going from three to five core products, and how do you think about the gross margin contribution of those going forward?
The two additional products that we will be selling will be on-net products, that is dark fiber and wavelength, or optical transport services. They carry 100% gross margin contribution. The wavelength market is pretty well understood. Today, it's about a $2 billion market with about 90% market share in Lumen's hands, about 7% in Zayo, and then a couple percent spread between Windstream and Uniti. That's pretty much the market, and then a few minor regional players. With our national footprint and our negative cost basis in this asset, we will be in a position to be extremely aggressive in that market, just as we've done in transit, selling 100% gross margin product, that will allow us to quickly generate market scale and share. In dark fiber, the market is a little less well understood.
There's not as great of a database of market comps because there's fewer transactions. What makes the Sprint network particularly desirable are four key attributes. One, they go to all of the major city pairs where customers want to buy services. Two, the routing between those cities is truly unique, meaning it's different than the routing of others. It's on different right of way, which has a real value to hyperscalers and other service providers. Three, it's a network that is relatively high quality. It was built with direct burial armored cable, so it is not in conduit. The downside is it's hard to pull additional fibers. The upside is that fiber was extremely well protected and has very few splices per kilometer and actually has higher quality than competing fiber.
The fiber was also built with SMF-28, an older generation of fiber that actually is better suited to coherent high-capacity throughput than the non-zero dispersion-shifted fibers that were deployed later in the early 2000s. Ironically, it's got less splices and better. The final point is that the fiber counts are not extremely high. Typically about 20 pairs on a given route. There are some routes that have 200 pairs. Other routes may only have 24 pairs. We think that this is something that is in demand in the market and should be able to help us both generate margin quickly and offset any negative costs. The final point I'd make is Cogent's 210-person net-centric sales team is exactly the sales team that would sell to the wavelength and dark fiber market.
The 60-person Sprint team is geared entirely towards corporate large enterprise sales. By ingesting the right type of salespeople into this market, we think we can gain market share quickly.
Is there another question?
Sam? Let me grab the mic over for you.
In terms of the corporate business, obviously, challenges of folks not being in the office. Are there any opportunities there with companies taking more sites? Or what is it gonna take besides just the business finally getting everyone back into the office to re-accelerate that business? Is there anything you can do in your control or is this one of those it just is what it is?
There are some things that Cogent can control and some that it can't. The good news is that the market indicators are indicating the market has bottomed and is slowly improving. The vacancy rate in our 1 billion sq ft of office space peaked at around 18% and is now starting to slightly improve. two, there is indications that key swipes are improving across all geographies, some more than others. More employees per day are going back. Now we're seeing a great deal of geographic dispersion. A market like San Francisco, we're at about 25% of pre-pandemic levels. In a market like Miami, we're at 95% to pre-pandemic levels. Those are things that are outside of Cogent's control. What is in Cogent's control is the ability to sell larger connections for those remote workers.
Pre-pandemic, the typical enterprise IT department architected their network to 97% work days in office, 3% remote. Today's standard is 60% in office, 40% remote. That actually requires two things. A larger point to VPN concentrator, usually at the corporate firewall, and then secondly, a second location for redundancy with such a high percentage of the workforce remote. Those factors are actually helping Cogent get more revenue from those customers. On the negative side, we're seeing some companies reduce remote offices, so there are less locations to sell to.
On the positive side, we're seeing the average new lease that's being signed in our footprint, 20% smaller than pre-pandemic levels, which would ultimately result in a 20% larger addressable market for Cogent if the building returns to its historic 6% vacancy rate, which has been the class A standard for the past 40 years. there's market forces that are improving and Cogent's ability to sell for companies that need to support a remote workforce. For that reason, we've seen our corporate growth rate, which had been 2% sequentially. It declined to -2% at the worst of the pandemic. Now it's effectively flat, and we think it will slowly improve, and we'll get back to 2%.
It's really gonna take all businesses to kind of have a vision of where they're going, as opposed to companies who are willing to just push the decision out.
Any more questions right now?
I have a follow-up to you were talking before about the opportunity to enter the wavelength space and enter the dark fiber space, and you pointed out that you have a negative cost basis on the fiber that you're acquiring. If I were to think about why is the pricing what it is in those markets, it's because those competitors had to deploy capital. They need to get a return on that capital. The question I would have is, are you looking at the addressable market for your wavelength and your dark fiber business as being limited explicitly to the fiber that you are acquiring? Because once you're in the space, you could have customers, particularly on the dark fiber side, come to you and say, "Hey, I'd like additional routes." Are you willing to expend the CapEx on that?
Because you may not necessarily have the same pricing advantage that you would have selling them something that has a negative cost base.
The answer is no. We are not a construction company. We have no desire to spend incremental capital on transforming our business in effectively a build-to-suit model. I know there are others in the industry who are pursuing that model. We have looked at their performance and think it's better to be a buyer rather than a seller in that market.
For how long do you think you could grow into that infrastructure before you inevitably say, "Hey, we've already sold all the wavelengths we could sell," or, "We've sold as dark fiber we could lease?
A very long time. Let me walk through the steps that are going to happen. Step number one, the physical networks will be tied together. Step number two, all of the MPLS, now to be converted to VPLS traffic and DIA traffic on that network will be ported over to Cogent. It will be less than 1% of the bit volume that Cogent is carrying. That will free up the pair of wavelengths or pair of fibers to use exclusively for wavelength sales, and we will add incremental wavelengths. We will then take a second pair of fibers along every route, utilize Cogent's equipment, light them, and port our entire IP network over to that second pair of fibers. We will be using two pairs out of the average of 20 pairs on a given route.
The remaining fiber will be available for either dark sales or additional wavelengths. Today's technology allows you to go to 160 wavelengths, each wavelength supporting roughly 800 Gbps. Based on the size of the market, just one pair of fibers dedicated to that will last us probably a decade. I think there's a lot of ability to sell our footprint, and we need to have the discipline not to try to get into other geographies.
We're right on time. Thank you so much for being here.
Brett, thank you very much, and want to thank everyone.