All right. Guys, thank you all very much for joining us. I appreciate you kind of continuing to stay with us. The tequila and wine is coming, I assure you. We have with us again, for the millionth year in a row, Cogent Communications Holdings Chairman and CEO, Dave Schaeffer. Dave, thank you so much for joining us.
Oh, well, Dave, thank you for inviting me. You're right, we're both showing our age, how long we've been doing this, how many of these conferences. I wanna thank BofA for a great venue. I wanna thank the investors for their time today, and obviously, for your interest in covering the company.
No, thank you. You know, obviously, you're kinda stealing the show this morning with the announcement of the acquisition of the legacy Sprint wireline business. I remember I said this earlier to T-Mobile. I remember being in college in 1990. I was watching TV, and it was when the long-distance wars were going on, and Sprint was, you know, offering $100 to become a long-distance customer, so I could call my parents. You know, they had the pin drop and the most advanced fiber-
Candice Bergen making a commercial.
Candice Bergen, good old Candy. Fast-forward 40 years, it gets sold for $1. I know you've spoken a lot about this today, and you had a very comprehensive conference call this morning, which is available on your website. I know the transcript just came out, so if people wanna kinda look through all the nitty-gritty, you can do it that way. I guess one of the key questions that has come up, we spoke to T-Mobile earlier, as you know. One of the key questions that's come up is, you know, essentially, you bought it for $1, but they're actually paying you or guaranteeing you $700 million over 52 months.
What is wrong with this company that T-Mobile is literally paying you to take it away from them? I think that that's something that eludes people.
That's a fair question. You know, Cogent built its business on two fundamental pillars. First being a belief that the Internet would eventually be the only network that mattered, and by building a highly efficient network, we could gain market share in a growing market while all other telecom services are being supplanted by Internet-based services. That fundamental premise turned out to be correct. The second pillar was our ability to acquire distressed assets in difficult situations for the owners. Prior to this, we had purchased 13 companies that had raised $14 billion of invested capital and deployed $4 billion in property, plant, and equipment. We repurposed those companies, whether it be PSINet, Allied Riser, Carrier1, FirstMark, Verio, into what became the Cogent Global Network.
That's allowed us to grow organically, generate free cash flow, and return over $1.1 billion to our shareholders over time, which is something that I think few competitive carriers can claim. We have looked at over 800 subsequent transactions from 2005 to today, and this was the first transaction that we consummated. As we thought about the legacy Sprint business, it is a business that actually has roots 100 years ago, in the old independent telcos, United and Brown. It was the first fiber optic network built. At the time, the two market leaders, AT&T and MCI, were using microwave, and Sprint was first to market with a nationwide market, fiber backbone and the pin drop, as you'd mentioned.
You know, to refresh your memory, MCI attempted to buy this business in 2002 for $121 billion and was blocked by the Justice Department. Now, it's true they had some minimal LEC assets associated, but it's really the long-distance network. Fast-forward to today, I think T-Mobile acquired this as part of the larger Sprint acquisition, which had really been driven by their wireless assets, the spectrum, the customer base, and the ability to consolidate towers on that footprint. We looked at this and said, "There are kind of three things here, two of which we want, one we don't want." The first of those is the actual network itself. It is a high-quality, unique right-of-way network with fiber that is optimal for coherent technology. The ability to reuse this asset is effectively sitting fallow today.
We're gonna do that two ways, by selling off excess capacity and migrating the Cogent network to those facilities. The sale will be the sale of wavelength services, so a brand-new product set for Cogent, and the sale of dark fiber. Both products that Cogent has not historically sold. The wavelength market is about a $2 billion North American market as sized by third parties and is growing. It's dominated by two players, and we will be a third player with a negative cost basis and therefore the ability to sell aggressively in that market.
Those two players are Zayo and
Lumen dominates the market, Zayo to a much lesser extent. There are a few very niche regional players like Windstream and Uniti, but it really is a two-player market. There's actually a very limited market for long-haul dark fiber. We today are using Lumen fiber. We have the ability to exit that agreement and save on our operation and maintenance expense and co-lo expense. We also know that a number of hyperscalers and some larger service providers would like to own dark fiber and have been unable to do that. We will monetize this fallow asset by utilizing it more efficiently. The second thing that we would be acquiring is the customer base. I think there was some confusion, at least on the part of some of my questioners this morning, on what exactly was that customer base.
It's really in two categories. There's about $110 million of the current $560 million run rate of revenue that are products that we do not want to support. We think they are not strategic. They are most likely gross margin negative, and Sprint or T-Mobile is subscale in those products. These are things like SIP termination, wholesale voice minutes, and some managed security services. That stream of revenue is being end-of-lifed by T-Mobile beginning immediately, and we are committed to honoring all customer contracts, even if they extend beyond closing for these uneconomic products, but these products are going away. The second and larger category of revenue is desirable to Cogent. It's $450 million of revenue of products that we wish to sell going forward.
Remember, this is a business that has been distilled down from a business that was doing $10s of billions in revenue at one point, and the remaining revenue streams that are desirable are internet access, which today is 81% of Cogent's revenues, and it comes in a form of either dedicated internet access or in transit bulk sales. We will port those customers into the network post-closing and achieve the efficiencies of our scale. Today, we are carrying over 100x the amount of traffic for internet connectivity that Sprint is carrying. There's a great deal of efficiency that can come through the consolidation of that revenue stream. The second subcategory is the private networking traffic. Today, it is almost exclusively delivered using MPLS, multiprotocol label switching, a technology that's 20+ years old, is not easily scaled, and requires expensive customer premise equipment.
Cogent has the technology, the wherewithal, and a track record of converting customers to a much more modern VPLS product. For both our DIA and VPN customer bases, which account for well over 90% of that revenue, we will maintain the revenue by giving the customers a more reliable service with at least a 10x improvement in throughput at price parity. We think this is a relatively straightforward sale, and it's to a customer base that we have not previously focused on, larger enterprises. The third category of revenues is de minimis today inside of the Sprint network, and that is colocation space, data center space. There is 1.3 million sq ft of owned data centers spread across 47 physical locations with about 150 MW of power.
We will integrate that into the 53 data centers that Cogent operates and have well over 1 million sq ft of colocation space available at edge locations throughout North America and couple that with our European footprint. The fourth product that T-Mobile has only recently begun trying to sell and represents less than $10 million of that $450 million of revenue is the wavelength product. We will be able to dramatically increase the marketability of that product by going to 800 carrier neutral data centers versus the 23 facilities that today the Sprint network connects to by utilizing our metro footprint. We pick up a revenue stream that is part of Cogent's core competency, and we pick up a network. Now, the bad news is there are a bunch of stranded costs associated with this business.
This is a business that has operated for a number of years, sub-scale, has a number of uneconomic commitments. We will be able to physically consolidate locations, typically into our own facilities, reducing lease costs for Cogent. We will be able to groom leased circuits. We will be able to eliminate the leased lit network in Europe and bring that all on-net for Cogent. That alone represents over a $25 million annual savings. We will be able to continue to grow domestically, both companies' footprints, and get better scale, and we will rationalize the headcount. However, we are sensitive to T-Mobile's strong desire to treat these employees well. We're actually going to be acquiring employees that in some cases are multi-generational at Sprint. We actually have our longest-tenured employee start date was in 1973. You and I both were young people back then.
You know, we wanna be careful to treat these employees well. In addition to the $700 million payment that we're receiving, there's an additional indemnification and payment to help support severance and termination costs for employees. We definitely wanna treat customers right and employees right. As we look at the business, we tried to tailor the payment streams of gross margin dollars to Cogent to mirror the expected rate of cash burn as we are fixing this business. Now, between signing and closing, the business is entirely under T-Mobile's control and ownership. Through working with them, we have shown them some paths to help them reduce costs. We expect that $110 million of revenue to go away and more than $110 million of cost to go away.
We anticipate that at closing, this $450 million business will be burning about $180 million of EBITDA year one. It will have about $30 million capital requirements and about $50 million of one-time capital expenses. In exchange, we will be receiving a guaranteed $350 million payment ratably spent over that first year. During that one-year period, we will achieve many of these synergies that I have talked about and expect in year two to have a negative EBITDA rate of sub $80 million on the acquired business. We also know that by that point, we will be able to begin growing the revenue streams of both the legacy products and the new products. You know, of the 1,300 employees at T-Mobile's wireline business, only 60 were in direct sales.
Contrast that to Cogent's 1,050 employees, of which 750 are in sales. The key to success here is growing on net revenues that have a 100% gross margin contribution and 95% of EBITDA. Putting this together, we have outlined a stream of guaranteed payments over a 54-month period that every month along the way, Cogent has more cash than it would've as a standalone business. Maybe I'll pause and give you a couple more, Dave.
Wow. All right. I'm trying to do all the math here. Okay. At closing, on a standalone basis, ex the $350 million payment in year one, the closing run rate at EBITDA is $180 million.
Negative.
Negative. By the end of year one, you think you can get that to sub 80 ?
That's correct.
Okay. If I look at 1,300 workers, 60 sales guys, I invert that and I go up and I say, "You maybe need 100-200 of the legacy Sprint people," and we take 1,100 of those people out at a $70,000 fully loaded rate, I can take out basically $100 million, and that's where that's the arbitrage that's here.
I think that's too aggressive on headcount. I don't think we're going to be nearly as draconian here as we were, say, at PSINet or Verio in terms of percentage reductions. In understanding these employees, there's a tremendous amount of institutional knowledge. They're broken into a number of functional groups. 336 were in field services. Today, Cogent has just under 100 people in field services. Probably the combined company needs 250-300 to manage a network of this scale. The second biggest group is the customer care group. We're at about 100. They're at about 245. We will consolidate and probably get down to, you know, below 200 on a combined basis, which I think will be appropriate. You know, we pride ourselves on our customer care.
We have Net Promoter Score in the mid-60s, and it's one thing to sell customers, it's a very different thing to keep long-term customers. We're committed not only for our embedded base but for the T-Mobile base in treating customers with respect and keeping them with the highest quality of service possible. You then move on to kind of back office operational teams. They're split into two major geographic groups. There's about 220 in Kansas City, about 140 in northern Virginia. We will compare those to the roughly 300 operational people in Cogent and try to consolidate where appropriate. I think there will be some reductions in force, but I don't think most of the savings are gonna come solely from that category.
They're gonna come from our ability to migrate off of a network we're paying maintenance to a network where our maintenance is a sunk cost. That's probably in the $15 million range. We have the ability to probably take out $15 million a month of run rate and circuit costs by moving off-net circuits onto the Cogent network because of the breadth of our metro assets and our international network. We also know that we have the ability to take other cost savings in kind of systems, accounting, back office, all of which give us a great deal of confidence that the pacing of the subsidies and the pacing of the service purchase agreement from T-Mobile to Cogent is more than adequate to make sure that at the end of every month, Cogent has more cash available than it would've had on a standalone basis.
Mm-hmm.
That is really the Glimpse test that we solve for.
Got it. Okay, thank you for that incremental color because that did not come, I think, across as clearly.
I was a little guarded in what I could say and was able to say.
Got it. Thank you. That helps a lot. A couple questions. This is gonna be more of a nerd question, but when you and I met back in the day before the 2004 IPO, your business was being built, if I remember correctly, on the pair that you got from WilTel.
Got you.
Which is how you found your way into Level 3, which is how we found your way into Lumen. That would have been 18 years ago. I think it was. There was a part of what you were saying as kind of the rationale for the transaction that that relationship was ultimately gonna come to an end or it was gonna change in a way that maybe made you feel uncomfortable, made this deal that much more attractive to move to your own facilities in a way that you never were able to before. When was that transition likely to happen?
Our original network was built in North America on a pair of fibers, roughly 13,000 miles from Williams on a 20-year IRU with two five year, no-cost extensions.
Got it.
However, those no-cost extensions actually have about $15 million in maintenance and colocation expense associated with them. We-
Can you get off of that then sooner?
We will not most likely exercise our second option, and we will be able to reduce that expenditure and migrate onto the Sprint fiber at the end of the first five-year extension of which we're basically two years into.
Oh, three more years.
Yeah. A little less than three years.
Okay. Yeah. Okay. Perfect. 'Cause that didn't coincide with the IPO. You had that for a few years before you got into the 2004 period. Got it. Okay, so that's one thing. The second thing is, I think that you raised some question marks for people about the $700 million payment that T-Mobile's gonna make. You highlighted that in one scenario, it might be considered revenue and EBITDA, which would, you know, impact the optics of your leverage calculation. On the other, it might be considered the absorption of uneconomic contracts, effectively, an amortizing asset. That was in the morning, this morning when I listened to the conference call.
What we learned around noon, East Coast time, was that T-Mobile's plan is to write off that entire $700 million, because they at least stated in the presentation that they had this morning that it is not their intention to take any of those services. It's a total write-off. Now, obviously, on the one hand, that makes a lot of sense because they don't wanna have to flow these costs through their longer-term EBITDA statement. But on the other hand, it would seem, if it's true, to suggest that there really isn't a question mark about whether you're gonna account for this as revenue EBITDA or as an absorption of contracts. It's not gonna be revenue EBITDA, because they're not gonna take any actual products. How are we gonna make this determination?
Yeah. It's actually not as cut and dried as that. Let's start with the payment. The payment is guaranteed.
The cash payment.
There's no ambiguity about that, the timing and the amount, and I think Peter, I'm sure, was as certain as I was about that.
100%.
There are two additional potential payments, severance payment and the additional liabilities indemnification that are capped at $125 million, that may or may not be needed depending on how much we spend in those two areas. Now, with regard to the characterization of that payment stream, there are basically two different ways to look at it. One would be as a payment of assumption of bad contracts, the other being revenue. We are actually reserving the ports for T-Mobile. They are available day one, and there is pricing on a per megabit basis on those ports. We will go out and get third-party industry experts to benchmark that against market pricing. I think we're pretty confident at the volumes and prices that we have given T-Mobile, these ports will pass that test, meaning they are at current market economic rates.
The second point is, will T-Mobile use the ports? If you have cable service at your house, the cable company sends you a bill, whether you turn the TV on once in a month or not. These services are on a take or pay contract. As long as they are at market price, we're pretty comfortable that some, if not all of it, characterized as revenue. The only reason this bifurcation makes an important distinction is, one, for investors that measure us by EBITDA, the second being for tax purposes. Under the old tax code, pre 2018, you would book this bad revenue as negative goodwill and then amortize it. In the Jobs Act, that tax treatment changed, and you now need to book it as a one-time benefit for the assumption of these bad contracts and then amortize the contracts out over time.
That actually has a negative tax consequence for us. The reason why I didn't answer a question more specifically today is 'cause I don't have the two data points that I need, that independent third-party analysis and the review of that by Ernst & Young, our auditors. I believe the correct answer is gonna be somewhere in between. What T-Mobile has said is today they do not intend to use that, but every quarter they are committing to reviewing their network requirements, and if they need it, they will use the bandwidth. The fact that we have provisioned the ports makes it legitimate revenue for Cogent, provided the pricing is correct.
The way this works is even though they might not be taking it, if you're creating a market price comparable product and it's provisioned and available, even though it's not being paid for in that way, the accountants will say, "Yeah, sure, it's some of it's revenue, at least some of it.
Again, I'm gonna use any kind of regular subscription service that is not dependent on your usage.
Right. Okay.
You have a fixed fee for your phone each month, whether you make a phone call or not. Whether you have one bit of data transmitted, you have a fixed monthly fee.
Got it. Okay. Understood. I think I get it. You don't turn on your TV, you don't use the service, but it's still revenue for the cable companies therefore.
That's correct.
That's the thinking behind it. Okay, that's super helpful. Gosh. We only got 10 minutes left.
We haven't even got to Cogent stock.
No, I know. We haven't talked about the real business. I have one more question, and it kinda relates to that piece of it, and then we'll segue. There was some aspersions that were cast towards the nature of the business that you will be retaining net of the pieces that you're gonna let die off. That $450 that you're gonna have, you know, we heard from T-Mobile that the reason they sold this business, it's in secular decline. The reason AT&T is guiding towards negative double-digit revenue in the enterprise is secular decline.
Why should we believe, and this is gonna come a little bit to the next piece with corporate revenues, but why are we believing that you can grow this business that T-Mobile's giving up on that no one seems to be able to grow?
If the products that I were acquiring were those legacy products, the answer is I could not grow it. I'm not a magician. This is an industry that is in secular decline, whether it's Verizon, whether it's AT&T, whether it's Orange, whether it's Telefónica. Wireline enterprise across the board is declining. The reason for that decline is the substitution of the internet and internet-based services for legacy services. This business has been distilled down to the point where the only revenue that we're acquiring are internet-based revenues. They are IP transit and DIA. They will be VPLS services delivered over the internet. Colocation and a nascent sub-$10 million dollar wavelength addressable market that is in growth.
We are not overly optimistic about the growth of this business, and we have stated that the combined business that is Cogent's legacy business, which has a 16.5-year track record as a public company, organically growing at 10%, should be able to grow 5%-7% combining these businesses. That's implying a lower growth rate for the acquired business, and that's accounting for the two new addressable markets that we will be going after with wavelengths and dark fiber. I think we're being very pragmatic. The second part of that question, you know, was implying the physical quality of this network was somehow deficient. There I'm going to push back strongly. It was the first nationwide fiber optic network. The two leaders, AT&T and MCI, were both using microwave at the time.
Secondly, it's a network along unique right of way. Third, it was built using armored buried cable, which has had very few cuts, so therefore the splice density per kilometer is lower than any of the other major networks in operation. Maybe most important, and this may be a little too technical, and I'm happy to do an offline with any investor wants to chat about it. The network was built with SMF- 28 fiber versus most of our competitors who use LEAF fiber. Originally all transmission systems were using non-coherent transmission. Think of it as you're sending pulses of light down the fiber and they spread out as they go through the fiber. What the fiber was doing was distorting that spread back into something that looked like what was injected into the fiber. That was the non-zero dispersion-shifted fiber.
That was the standard in the 1990s and 2000s. As it turns out to get greater than bits per wavelength, all of the equipment vendors, Ciena, Infinera, Cisco, use coherent technology. You're more than happy to check this with Corning or Sumitomo or any of the fiber vendors. In a coherent system, you think of the light going as a single bundle down the cable, and that attempt to keep it together actually distorts it and lowers its throughput. So the SMF- 28 fiber actually has greater actual throughput. As part of this process, T-Mobile, at its expense, conducted a third-party review by Corning, where they did a span by span loss calculation and a quality assessment. I would argue that the quality of the physical network we're buying for today's optronics is superior to what our competitors are selling.
Single-mode fiber SMF. What about erbium-doped fiber amplifiers? We get some of those. All right, we got five minutes left. Cogent, the real business.
That's the real business.
Um.
$600 million of revenues.
A couple questions. I know we're not going to get. I hope you're having a lot of offline meetings, because we're not going to get anywhere on this. I think that the two big challenges that you've been having, number one has been return to work. I know you're a big landowner yourself.
Yeah.
Real estate owner, you have a lot of insight there. We'd love to hear that. Number two has been getting kids to show up to work and do their job. You know, productivity has gone up for sure, but it's redlining over relative to history. Is that really sustainable? Have you been able to hire people and keep them and train them and make them quota-bearing real people? Those are my first two questions, and you're going to talk so long, we'll run out of time.
Yeah. Let's start with the core Cogent business. The NetCentric segment, where we're selling bulk connectivity, has actually outperformed our historic trend lines, and because of the pandemic, is continuing to perform at exceptional growth rates, 16.5% year-over-year, last quarter, constant currency, 10.2% on a reported basis. Most of that growth is streaming outside of the U.S. I have quite honestly been too pessimistic on that business and expected it to decelerate sooner than it has. We expect that outperformance to continue. That is approximately 40% of Cogent's revenues. 43% and 57% of our revenues come from selling to corporate end users. The corporate end user business had been growing pre-pandemic at a consistent pace of about 2.5% per month.
The pandemic came along. We've had 10 quarters of flat or negative growth because employees are not returning to office. What we are seeing is a stabilization in that business. From a negative growth rate of negative 6% or 7% back to a 0% sequential growth rate. Most companies, 95% of companies surveyed say they will have hybrid work. Pre-pandemic companies architected their networks to 97% of work days in office, 3% remote. Now the standard's 60% remote, 40%. Excuse me, 60% in the office, 40% remote. I think that means companies will take less square footage, but will still have offices. They also need larger connections to support those remote workers. We have seen a significant improvement in the corporate business Q4- Q1.
We saw flat performance Q1- Q2 and expect Q3 will improve over Q2. We think that the corporate business is recovering, but not as quickly as we would have predicted a year ago. Quite honestly, it's been driven by getting people to go to work. To your question about employees, throughout the pandemic, we actually hired at record paces. Cogent today is still hiring at the fastest pace in its history. Our sales force did shrink, and we are looking to grow that, but the shrinkage was coming from underperformance. We were unable to make reps as productive in a remote learning environment as we could in an office. We today have 80 sales teams spread across 45 offices. They go through a month of intense training, mostly online, but then get supplemental training in the office.
With that, we've seen a dramatic reduction in sales force turnover, a growth in productivity, and growth in the sales force. We have no doubt that we will be able to hire enough sales reps to allow us to achieve our revenue growth target. Earlier in the conversation, we talked about reductions in headcount. Now we're talking about hiring challenges. Well, this is a difficult market. The Cogent job is hard, and most people are not successful at it because it's telemarketing, but there are many that try, and our goal is to take those that try and make as many of them successful as possible.
We have to leave it there. Dave, thank you so much for your time.
Dave, thanks for your great questions.
It was a big day for you, so congratulations on the deal and, thanks for being here to share it with us.
Thank you, buddy.
Thank you, man.
Take care.