Good morning, and welcome to the Cogent Communications Holdings Second Quarter 2023 Earnings Conference Call. As a reminder, this conference call is being recorded, and it will be available for replay at www.Cogentco.com. A transcript of this conference call will be posted on the same website when it becomes available. Cogent's summary of financial and operational results attached to its press release can be downloaded from the Cogent website. I would now like to turn the call over to Mr. Dave Schaeffer, Chairman and Chief Executive Officer of Cogent Communications Holdings. Please go ahead.

Thank you and good morning to everyone. I'm going to start by apologizing for the possible length of this call. There are a number of new topics that we are going to be chatting about, but we will try to be complete in answering everyone's questions. Welcome to our second quarter 2023 earnings conference call. I'm Dave Schaeffer, Cogent’s Chief Executive Officer. With me on this morning's call is Thad Weed, our Chief Financial Officer. Hopefully, you've had a chance to review our earnings press release and our 10-Q for the quarter. Our press release includes a number of historical quarterly metrics that we present on a consistent basis every quarter. We've included a number of additional metrics this quarter related to our acquisition of Sprint, and we will continue to provide these metrics each quarter going forward. Our press release metrics now include corporate, NetCentric, and enterprise revenue, and customer connections and metrics related to the Sprint network assets. Now for a few summaries of round our results. We closed our acquisition of the Sprint wireline business purchased from T-Mobile on May 1st, 2023. This transaction significantly expanded our network, our customer base, our employee talent, and materially increased the scope and scale of our business. We now have an annualized revenue run rate in excess of $1 billion. We acquired a number of large enterprise customers, many Fortune 500 companies. The customers are larger than the typical customers in Cogent’s corporate customer base. We acquired a significant fiber network of owned assets and owned real estate facilities. We acquired a number of right-of-way agreements and relationships. It would've been virtually impossible to assemble the set of assets on our own. We hired many valuable experienced employees from the wireline business. The majority of these employees have been with…

Yep. Thank you, Dave, and good morning to everyone. This earnings conference call includes forward-looking statements. These forward-looking statements are based upon our current intent, belief, and expectations. These forward-looking statements and all other statements that may be made on this call that are not historical facts, are subject to a number of risks and uncertainties and actual results may differ materially. Please refer to our SEC filings for more information on the factors that could cause actual results to differ. Cogent undertakes no obligation to update or revise our forward-looking statements. We use non-GAAP financial measures during this call. You will find these reconciled to the corresponding GAAP measurement in our earnings releases that are posted on our website at Cogentco.com. Some comments on revenue by corporate NetCentric and now wavelength and non-core. We analyze our revenues based upon-network connection type, which is on-net, off-net, wavelength services, and non-core services. We also analyze our revenues based upon customer type. And as Dave mentioned, we now classify all of our customers into three types, NetCentric, corporate, and enterprise customers. Our corporate customers buy bandwidth from us in large multi-tenant office buildings or in carrier-neutral data centers, and these customers are typically professional services firms, financial services firms, and educational institutions located in multi-tenant office buildings who are connecting to our network through our data center footprint. Our NetCentric customers buy significant amounts of bandwidth from us and carrier-neutral data centers, and include streaming companies, content distribution service providers, as well as access networks who serve consumer and business customers. Our enterprise customers generally purchase our services on a price-per-location basis and are typically larger than our legacy Cogent customer base. On revenue reclassifications. In connection with the Sprint acquisition, we reclassified a small portion of legacy Cogent revenue to enterprise…

Hey, thanks, Thad. I'd like to highlight a couple of strengths about our network, our customer base, and our sales force. Now for some NetCentric details. We continue to experience significant growth in our legacy NetCentric business. We're direct beneficiaries of the continued acceleration in over-the-top video and streaming, particularly in international markets. At quarter's end, there were 1,526 carrier-neutral data centers and 56 Cogent-owned data centers directly connected to our network, for a grand total of 1,582 data centers. This is more than any other carrier as measured by independent third-party research. The breadth of this coverage allows us to enable our NetCentric customers to better optimize their networks and reduce latency in connecting to their customers. We expect to continue to widen our lead in this market and project to add an additional 100 carrier-neutral data centers per year to our footprint over the next several years. We also expect to convert an additional 44 of the Sprint technical spaces into Cogent data centers. To date, we have converted one of these facilities. As of today, we are selling wavelengths with a rapid provisioning cycle in 35 carrier-neutral data centers. And with an extended provisioning cycle, we can sell wavelengths in over 200 carrier-neutral facilities. By the end of 2024, we anticipate being able to sell wavelengths in 800 or more or more US carrier-neutral data centers with continued reduced provisioning intervals. We have significantly expanded our network footprint with owned fiber and additional IRUs, as well as right-of-way agreements. At quarter’s end, we are directly connected to 7,891 unique networks. This collection of ISPs, telephone companies, cable companies, mobile operators, and other carriers, give us direct access to the vast majority of the world's broadband subscriber base and mobile phone users. At quarter’s end, we had a…

[Operator Instructions] Your first question comes from the line of Phil Cusick with J.P. Morgan. Your line is open.

Dave, maybe we can start with NetCentric and just maybe go through again what the organic growth is in NetCentric this quarter, both year-over-year and sequentially, and help us think about size of customers committing and things like that, that's going to drive revenue growth for the next segment going forward. Thank you.

Yes. So, as we mentioned in the call, we saw an acceleration sequentially both in traffic growth as well as a year-over-year improvement in traffic growth. We have seen strong demand organically from customers outside of the US, as well as having two additional benefits in this quarter, that being the benefit of selling NetCentric services to T-Mobile outside of the transit agreement. These are primarily co-location services and VPN services, and we ended up acquiring a customer base of off-net and on-net layer two services with T-Mobile. The traffic growth was approximately 11.1%, or excuse me, revenue growth of 11.1%, and that was from the legacy Cogent business and 19.3%. Even if you netted out the additional services that we sold T-Mobile, the growth in that business was similar to Q2 of 2022, both on a sequential and year-over-year basis.

Thank you. That's helpful. And how do you think about the - obviously, there's going to be another month of movement this quarter, but the organic go-forward growth in that business, is that a sustained double-digit sort of annualized growth from here once everything is through, given all these other opportunities?

Yes. So, you are correct, Phil. We clearly will have an additional benefit in the third quarter because we will recognize a full three months of the commercial services that we are selling T-Mobile. Some of those services will wean away by design, but we think there is a very long tail to our ability to provide some of those non-transit services to T-Mobile. Independent of that phenomena, we also believe the underlying strength in the organic business, which has been about a 10% or 11% revenue growth rate year-over-year, continued in this quarter and will continue going forward. We're very encouraged by strong demand in international markets. To remind investors, our NetCentric business over an 18-year history, has averaged about 9% growth. We went into the pandemic growing substantially below that trend line at about 3% year-over-year growth. Growth skyrocketed at the beginning of the pandemic, all the way up to 26% year-over-year. It has slowly reverted closer to the average, but remains above that long-term average. As I said, we're growing about 11% now year-over-year, and we believe that NetCentric revenues for the combined company will continue to grow in low double digits for the foreseeable future.

If I could follow up on one more. Any change in the size of customers buying? Last quarter it was a lot of larger. This quarter, any shift there?

We continue to actually see some strong demand from some of our large hyperscalers. As Thad mentioned, our average new sale remained flat at about $0.10 per megabit. Actually, our installed base went up on a per megabit basis to $0.28, in part because of the acquired T-Mobile customer base. I think the big change in buying patterns we've seen actually over the last couple of quarters, has been a shift towards people buying for a new application, that being artificial intelligence, and the collection of data. So, what we have seen is some customers who traditionally had traffic very asymmetrically skewed in the outbound direction, now collecting a significant amount of data in the return path. And that has driven more growth from some of the larger software and hyperscale companies that I think are using that data to build their large language models and create generative artificial intelligence applications. We think we're only at the beginning of that trend, and that is an encouraging tailwind for the entire NetCentric business. Now, some of our smaller customers may catch up and start also exhibiting those types of patterns. Most of our growth in small customers has really been in international markets from more regional access networks who have really seen an acceleration in streaming video, much as we did here in the US and western Europe maybe 18 months ago. So, there is a lag. So, you've got really two different things going on. The large customers are increasingly pulling information as opposed to pushing for their AI applications, and the smaller international customers are accelerating the pull of content.

That's helpful. Thanks, Dave.

Your next question comes from the line of Frank Louthan with Raymond James. Your line is open.

Great, thank you. Just want to be clear, can you walk us through kind of what the run rate cash payments will be from T-Mobile going forward in each full quarter? And just to be clear, you're going to be recognizing EBITDA, the cash payments, adding back, not the straight-line recognition that you'd previously thought that you would be able to recognize.

Yes, sure, Frank. So, for the first 12 months of the agreement, we will recognize $29.2 million a month, or $87.6 million a quarter. This quarter, we recognized only one payment, even though we had billed for two. We are recognizing that on a cash basis. When we consulted both with the SEC and with our auditors, Ernst & Young, we concluded, in alignment with T-Mobile, that this should be a bargain purchase transaction, even though we are providing the IP transit services and they do meet many of the criteria associated under 606 for revenue recognition. If they were treated as revenue, we would have recognized that revenue on a straight-line basis because it is a unified contract. The cash payment stream will actually step down in month 13 to $8.3 million a month, and continue for the next 42 months. So, we will recognize $24.9 million a quarter on a cash basis after the first 12-month period. The reason for the change in accounting treatment was that T-Mobile made the decision to treat this as a bargain purchase. At that point, we had a customer saying that they were viewing this as a transactional cost and not as a services agreement.

Dave, if I just correct one thing. T-Mobile didn't treat it as a bargain purchase. That's our side,

That's our side.

Yes. They treated it at the end of 2022 as a loss on disposition of a business. That was their side. So, in terms of kind of matching the account,

Our side.

We had the bargain purchase gain. Sorry to interrupt, but wanted to make that clear.

Yep. No, that's why you're the accountant, Thad. But we will recognize that and add the cash back each quarter because it is cash, and we are reconciling that to our cashflow statement. Was that helpful, Frank?

Okay, great. Yes, no, that's great. And then on the customer size, I think you gave kind of the breakdown. What is the largest customer as a percentage of revenue, and is that a legacy Sprint or a legacy Cogent customer? And then what's the largest Sprint customer as a percentage of revenue, and what's kind of been the trend with that revenue?

Yes, so our largest customer is now a legacy Sprint corporate customer. They are, as a percentage of total revenues, about 4% of revenues. And we did acquire a number of large portion kind of 500 companies, the largest of which I think buy services from Cogent in approximately 1,600 locations around the world.

Okay. All right, great. Thank you.

Your next question comes from the line of Greg Williams with TD Cowen. Your line is open.

Great. Thanks for taking my questions. First question, you provide some nice insight on your Q&A here on the NetCentric business ex-Sprint in terms of customers connection and organic revenue growth. Can you do the same on corporate? How much did the reclass of some of those Sprint customers and revenue impact and what was the true organic growth on the legacy Cogent corporate business? Second question is just, on the waves business, when does that really start to ramp to the 800 data centers and get those 259 sales folks up to speed, and what sort of the opportunity next year and the year after? You have big ambitions there. And just lastly, if I can sneak in a housekeeping question, so if T-Mobile only paid one month to you guys, are we going to see four months of payments next quarter, or is it typically going to be in arrears and it's going to be three months next quarter? Thanks,

I'll take the easy one first. We'll get three months next quarter. They pay us typically on the second or third of the month after we invoice them. And as Thad indicated, they had been extremely prompt in making those payments on time. Now let me take your first 2Questions. On the corporate side, organically, Cogent grew sequentially 0.001%, and year-over-year 0.006%. So, just under 1%. And that was independent of any customers that were reclassed from the Sprint acquisition. And as Thad said, going forward we will treat all customers by type equally. So, they'll either be corporate NetCentric or enterprise, independent of whether they were organically sold by Cogent or acquired. We had customers moving in the other direction as well as we took a small number of Cogent customers and reclassed them as enterprise. And this would give us, I think, a good baseline so we can report consistently going forward. But that sequential growth rate and year-over-year growth rate in the corporate business is an improvement from last quarter and indicates the last four quarter’s trend line of slow but fairly consistent improvement in the corporate business. Now for the wavelength question. All 259 of our NetCentric reps, as well as our enterprise and our corporate national account managers, can sell wavelengths. The vast majority of the wavelength market is NetCentric-type customers. The Cogent sales force has been engaged with existing wavelength customers, existing NetCentric transit customers and wavelength opportunities that we have not had a previous commercial relationship with. I would say that our sales force has already engaged with the vast majority, probably 95% of the potential market in discussing the wavelength opportunity. The challenge for us is to get the Sprint network connected to enough locations where that demand can be fulfilled. And…

Great. That was very helpful. Thank you.

Hey, thanks, Greg. Thanks for hosting me. And by the way, I heard one of our wavelength competitors at your conference indicate they were going to have wavelengths available on a daily basis. The only way you can do that is pre-provision.

Right. Good color. Thank you.

Your next question comes from the line of Walter Piecyk with LightShed Partners. Your line is open.

Thanks. Dave, I want to go back to corporate. I think you've mentioned maybe twice on this call that it was up sequentially. I just want to make sure that we're talking apples to apples, because if I took the number pulled out, I think it's 12.9 for the Sprint contribution. Obviously, didn't look at non-core and took out USF, it looks like it was down like 3.7% sequentially, the corporate business for a true legacy organic number. So, I'm guessing my math is wrong and I'm just hoping you can give me, if we're not looking at the benefit that you had from USF and we strip out Sprint, did it grow sequentially?

It did grow sequentially. The USF benefit was actually very minimal. Most of that benefit came from the increase in USF for the Sprint customers. And again, to remind you, corporate includes both core and non-core products. So, I'm not trying to be coy here, but every product or every service gets four classifications. It gets on-net versus off. It gets customer type, which is corporate, NetCentric, and now enterprise. It gets geographic, which is US and rest of the world. And then finally, it gets product. And the products include internet access, VPN services, co-location, now wavelengths, and then finally non-core services. When you look at the organic Cogent revenues, for corporate customers, they grew sequentially 0.001% and …

Seven tenths sequentially.

And six tenths year-over-year.

Yes, six tenths year-over-year.

Got it. So, what was specifically the USF contribution from Sprint? Because I know last quarter was 4.2. So, what's the comparable for the legacy business on USF

Sprint USF contribution was about $7 million. It was the majority of any change in USF.

Yes, the actual USF rate flag.

Got it. Okay. So, then, which goes to the next question. Dave, you haven't grown corporate sequentially since I think it's like September of 2020. That was a long run that's now finally inverted the opposite way. Can you speak to the issues on why that happened? I know I have friends - there are companies that there's obviously a bit of a crackdown finally maybe going into Labor Day, and maybe this is the finally the Labor day that your prediction will hit in terms of getting people back in the office. I mean, can you just speak to the issues on why corporate has finally inverted positive and then how you expect on a sequential basis things to kind of play out over the course of the end of the year and into early 2024?

Yes. So, I do believe that many companies are taking a more proactive approach on getting employees into the office, at least on a part-time basis. Secondly, we have seen market by market, a great deal of differences. And I think there is a trend that when we've seen markets like South Florida, Texas, Phoenix, kind of perform as if there was no pandemic, we think that kind of improvement will continue to spread to other markets. Now, if we looked at our worst performing market, San Francisco, it's almost like there's been no recovery from COVID. So, it is very geographically unequal. However, what we saw happen starting in South Florida, has now spread into Texas and Arizona and markets like Atlanta, are improving. They're not still quite to pre-pandemic levels, but we're tracking this on a market-by-market basis. I think the second thing that's happening is companies can only procrastinate so long in making decisions as they need to deploy new tools and modernize. And I think the three-year hiatus in decision-making is starting to wane. And we appear to see improved sales efficacy and improved funnels in markets where we hadn't seen that maybe six or nine months ago. So, again, you don't know until the customer actually signs the order, but it does appear that we're getting a lot more corporate activity and we are on a path to recovery. But again, 0.006% growth year-over-year is not the 11% that we've historically looked.

At least it's not negative though, Dave. So, that was a good trend to finally break. I guess my final question is more kind of a disclosure question, which is, on a prior call, you kind of contested some derogatory terms I had for Sprint's business that you purchased, but I mean, you didn't pay anything for it. You paid $1 for it. So, I don't think there's any question that this is maybe not a highly valued business. And yet corporate historically has been that kind of workhorse engine until we had this kind of COVID - multi-year COVID dip. So, why wouldn't you provide investors? And now that it's finally inverted positive, why wouldn't you provide this kind of legacy stat going forward to have people say, okay, let's look at, they bought this business, this not highly valued business for nothing, right? You're not getting able to print the revenue. Let's put it in a side bucket and see what expenses Dave can squeeze out of it and maybe get some revenue synergies, but then look at the core thing, really what drives your ability to grow the dividend and see how that's returning to growth. I don't understand why you wouldn't mash that stuff together unless - like other companies do that to hide a bad story. Like if this is now a good story, why not present that to investors going forward?

Well, first of all, I think we actually have multiple good stories to tell. So, let me kind of disaggregate your question and statement. Let's start with the asset that we bought. We bought two different things. We bought a physical network. We actually hired a third-party, KPMG, to come in and evaluate that. Arm’s length, we had no previous relationship, never used them as an auditor or for tax work, and just said, come in and do an appraisal on this asset. They realized that that asset was worth $1 billion. We paid $1 for it, but it's an asset that T-Mobile had no use for. It was a fiber optic network and a series of switch sites that were not strategic to their business. So, they viewed them as a liability. They had to pay taxes on them. They were literally sitting empty, much like an office building that would have no tenants in it. And we saw value in that and said we can repurpose that and create a growth business in selling high capacity optical transport services and by selling co-location in that footprint. We are in the process of making that conversion practical. We're connecting it to the rest of the world. We're putting in the correct transponder equipment and OADMs to be able to provision wavelengths quickly. And we've actually - I wish I could tell you I was so smart that I would've known the AI tailwind for optical transport and co-location was coming, but the reality is there's been a huge uptick in the short-term for demand for both of those services because of the need for high compute and high data transfer that is not easily done on the internet. So, we were at the right place at the right time.

David, let me just interrupt you though. I mean, I respect the fact that you can argue that T-Mobile sold a $1 billion asset right before was going to take off because of AI. I certainly appreciate that, but even if that's true as a narrative, then why not put that in a separate bucket and show what you did with this business you bought, and then continue to show how the legacy business is doing, rather than mashing them all together?

And I'm going to get there and answer the rest of your question. So, the second part of your question is, we acquired a customer base that had a bunch of unprofitable services associated with them. When we looked at this business, initially it was burning $300 million of EBITDA and $30 million of CapEx, nearly $1 million a day. That had nothing to do with the network. These were the services. The network was virtually fallow, sitting there empty. To give you a sense, the data center or the switch sites that we acquired had 22,500 racks of dead equipment that hasn't been in service for at least a decade sitting in them. We have to clear that stuff out. But on the business we bought, we knew we needed a additional stream of revenue to at least give us the time to end-of-life products and to move that traffic on that and fix it. We convinced T-Mobile to buy $700 million of transit services from us with the payment stream that we disclosed in order to mitigate the burn in the operating business. The third point is, we acquired a customer base. We were actually told that that customer base was enterprise. We only saw five of the 1,396 customer names prior to closing. That's fair. That's the way the FTC wants things in non-consummated transactions. Once we unmasked those customers, we quickly realized there were some corporate and there were some NetCentric customers in there, and we classified them appropriately. We also said, we should go the other way and maybe there are some Cogent corporate customers that really fit the definition of enterprise. We've trued that up. We went through a customer-by-customer, port by port reconciliation for the 151,000 customer connections and made that true up.…

Roger that. Thanks, Dave.

Thanks, Walt.

Walt, just real quick, I do want to give you the actual USF numbers before we close with you. So, Sprint was $7 million of the USF sequentially, and total USF went up $6.8 million. So, Cogent classic USF was down $200,000 sequentially.

And is there a mix of corporate and enterprise in that or is that just - you’re just giving that total number?

No, that's all corporate.

That's all corporate. Thank you. Got it. Thanks, fellas.

Your next question comes from the line of David Barden with Bank of America. Your line is open.

Hey guys, nice to talk to you again. Thanks for all the detail, Dave, as always. We've spent a lot of time kind of noodling on the wavelengths and the lit services and some of the cost savings opportunities. I think something you kind of have referenced today, these 40 technical facilities, one of which you converted into a Cogent data center, could you elaborate a little bit on this opportunity? Is this purely a way to expand your reach with a more on-net approach to some of these products we've already talked about? Or is there an actual distinct data center strategy where you're going to be putting X capital to work per technical facility to generate X megawatts of capacity that could generate Y incremental dollars of revenue? Could you just share with us a little bit on your thinking there?

Yes. Hey, thanks, David, for the question, and the only thing I'm going to dispute is the need to spend incremental capital to do this. So, Sprint built a fiber optic network that terminated in tandem switch sites. These switch sites were designed to allow connectivity to ladders. This was a design back in the ‘80s and early ‘90s. That network carried exclusively voice traffic until the late ‘90s and then carried some proprietary data and a little bit of internet traffic. That network is virtually empty today, almost no traffic on it, and it terminates in these former switch sites. There are 482 technical buildings owned fee-simple. There's actually 1.6 million square feet in those facilities. In addition to that, there's nearly 300,000 feet of leased technical space that we will be exiting as part of our cost savings initiatives. And that project is well underway. As we can exit those leases, that is a big part of the way we can save that $180 million run rate in North America, the $25 million internationally. But for the owned facilities, we look at the largest of those. There were 45 of them that comprised 1.3 million square feet and already have 160 megawatts of power to those facilities that are today generating no revenue, sitting empty, with no connectivity to the rest of the world other than to the Sprint backbone and to ILECs in their territory for TDM interfaces, which are no longer applicable. So, we are doing four things. One, we are physically connecting those networks to our metro footprint. We have two reasons to do that. One, to make the data center marketable. Two, we need to extend the Sprint network into carrier-neutral data centers so we can sell wavelengths. The second thing we're doing is…

Thank you. Yes, perfect.

Your next question comes from the line of Brett Feldman with Goldman Sachs. Your line is open.

Hey, Dave, 2Questions if you don't mind. You talked about this target of getting to $700 million of wavelength revenues over a multi-year period. I think that's higher than what you had discussed before. So, correct me if I'm wrong, but I'm just curious, what's giving you some confidence in that outlook? And then obviously, the business is positioned to delever going forward. You’re clearly above your current target range. How are you thinking about your current comfort level with operating closer to the higher end or lower end of that range? And then how does your trajectory of delevering ultimately factor into evolution in capital returns, meaning, when let you start raising the dividend at a faster pace, or or how would you think about buybacks as an alternative? Thanks.

Yes, let's start with the wavelength market. We took the initial view that the market was going to be static at a $2 billion scale. There have been independent third-party studies that indicate that market is going to grow at about 7% a year. We have said that we will get to 25% market share over a seven-year period. The $700 million number just represents the same 25% market share of a slightly larger market. What appears to be driving the growth in that market seems to be these AI applications for large data replication that do not sit well on the public internet. Let's be very clear, a wavelength is less flexible and higher cost per bit than using the public internet. The public internet is the majority of traffic in the world, and it's where the majority of growth is coming from. But there now is a new driver for this premium service of wavelengths and we feel uniquely positioned to be able to capture that. And we think our growth is going to be relatively linear. We're going to get to a 25% market share. And right now, we believe that market appears to be growing, and that's a positive for us. Now, to the leverage question.

Hey, Dave, sorry, if you don't mind, just a follow-up on the wave question before we get to leverage. Can you just give us your update, how are you pricing in the wave market right now? How much of a discount? So, you talk about how it's a premium price market, but are you coming in saying, well, it's not premium priced if you buy with Cogent because you have such a low cost base?

So, we have four distinct levers to pull in winning business. I think the most important is the uniqueness of the routes. They weren't available in the market before, and people want routes that we have for diversity. Secondly, it will be the ubiquity of the number of data centers that we are offering services. Part of Cogent’s strategy is to go wide and have the ability to provision quickly transit, and now wavelengths, in every data center where there is demand. The third thing that we have is the ability to offer an end-to-end solution. In many situations today, a customer has to buy a metro wave from one vendor and a inner city wave from a different vendor. We will have a holistic end-to-end solution with seamless provisioning. And then the fourth thing, which is critical, is being able to actually deliver what you sell. And what we have heard consistently from the marketplace is that many of our competitors fail to deliver or don't deliver in a timely manner. It is critical that we replicate our service delivery quality for wavelengths as we have delivered on transit. Now, with those four advantages, we will use price. My goal is not to destroy the market, but it is to capture market quickly. We will discount, but we have an established brand. We have an established sales force. And because of that, I don't think we will need to be as aggressive as we were in the transit market. We will, if necessary, go there because as Walt said, I got a network for a buck. That's great, but I don't want to sell it cheaper than I have to. But if it takes price to clear the market, we'll use it. And we feel very comfortable that our…

Thank you.

Your next question comes from the line of Tim Horan with Oppenheimer & Co. Your line is open. Tim, your line is open.

Sorry. You guys have had a major improvement in NetCentric pricing. Do you think that's sustainable and what do you think is driving that? And then I just had some basic questions on the go-forward guide.

Yes. So, in terms of NetCentric pricing, there are really four factors that weigh into that pricing, how much a customer buys, how long they buy for, and in what geography they are buying. You layer on top of that existing customers that tend to want to extend their commitments each and every quarter. And we typically see about 2,500 customer connections a quarter that end up doing a reprice and extend. This quarter was no different. So, I think when we layer those overall pricing disciplines onto our customer base, we are going to see a slight moderation in the rate of price decline for NetCentric. More of the growth is coming from international markets. More of the growth is coming from smaller customers in general. We are seeing, however, large customers also having this new use case that they didn't have before and driving some growth. When we kind of layer that together, the 23% long-term decline in price per megabit is probably moderating a little bit. It may be reducing down to a 20% rate of decline, but the price of transit is going to continue to come down for two reasons. The market is sufficiently competitive. We are not a monopolist. I wish we were, but we're not. And then secondly, that the underlying technologies to produce route of bit miles continue to improve at pretty consistent rates. Whether it be optically interfaced routing improving at about a 40% per year CAGR, or wave division multiplexing improving at about an 80% compounded improvement, those underlying trends are going to continue. So, I think it would be naive to think that NetCentric prices will ever plateau, but I think the rate of decline will moderate.

So, the rate of decline is 500 basis point improvement, but you're seeing a lot of improvements from geographic diversification, from customer diversification and now new use cases. So, it'll look better than the 20% decline for a while because of these.

That's right, Tim. That's exactly right.

And then just on the guide, the guidance for the 6% revenue growth long-term and 100 basis points improvement, can you just give us the base that that's off of? And I guess related to this, can you give us some sense of what the revenue and EBITDA are going to look like for the third quarter? I know you don't give guidance, but we have a ton of moving parts here, and I know both, I think, are related. Thank you.

Well, this is a much longer call than I think most people would like, but I think it is critical for foundational reasons to give all this background. So, to be clear, the revenue guidance is a range of 5% to 7%. You pick the midpoint of that over the next multi-year period. The base that is off of is the combined revenue of the company. And within that base there are multiple components. There are non-core revenues that we want to go away and go down. There are corporate and NetCentric revenues, to Walt's question about reporting, that are now combined. There are enterprise customers that are combined, and then there is the new product of wavelengths that will be reported independently. The wavelengths, like any other on-net product, carry the absolute highest incremental margin. We also have T-Mobile as a customer above and beyond their transit purchase from us for other services that we know will decline over time as they wean their wireless network's dependence on the transport network that we acquired. These were intermeshed assets. When we put all of those pieces together, we will be building off of a revenue base of about $11. billion, $1.2 billion a year, growing at 5% to 7%. When we look at the mix of products that we will have on-net and off-net, we will be able to deliver about 100 basis points a year of margin expansion. In the third quarter, because we will have three payments from T-Mobile, not one, and in Q2 we had two months of expense and only one month of payment from T-Mobile, in Q3 we’ll have reduced expense in three months, we should have margins, inclusive of the T-Mobile transit payment, in the mid-30s.

And so, the 100 basis points guide, is that off that mid-30s range?

It is, but remember over a multi-year period, that T-Mobile payment for the $700 million transit agreement is going to step down. So, yes.

But I mean also - I mean, 10 years from now, the goal would be 45% off that 35% roughly, extremely roughly.

Yes, I think that's a fair way to think about it. Tim,

Great job. Thanks, Dave. Good luck.

Your next question comes from the line of Nick Del Deo with MoffettNathanson. Your line is open.

Hey, morning, Dave. I just want to confirm that the revenue from the commercial agreement with T-Mobile is being allocated entirely to NetCentric.

Yes, it is entirely NetCentric. It is predominantly on-net. There is a small component of it that is off-net, and it is for two primary services, layer two VPN services and co-location.

Okay. Now, I think you said earlier, I forget who asked the question, but I think you said that if you deducted the commercial agreement from the NetCentric results in Q2, you got growth rates that were similar to Q2 of a year ago. When I adjust for it, I get numbers that are slightly up sequentially. Just want to confirm what you said because I'm trying to figure out how to allocate these changes appropriately.

Yes, I think that's right, Nick. So, last year in Q2 of 2022, I think our sequential NetCentric growth was roughly 0.003%. It tends to be not the biggest traffic growth quarter because of the heavily dependence on students and the fact that people start to go outside more. So, I think it clearly was much better sequentially because of the sale to T-Mobile of these non-transit services predominantly on-net, and they are counted as net central.

Okay, got it. Thanks for that. And then second, just thinking about EBITDA, can you talk a bit about costs that may have weighed on the bottom line this quarter that are deal-related that you're not specifically breaking out? Like I remember in his remarks, and I think in the Q, you mentioned some bad debt expense for Sprint. That seems like it's sort of a one-time initial thing, but I'm wondering, is there severance or termination costs you’re not breaking out over time, other inefficiencies, stuff like that that we should consider when looking at the bottom-line results?

Yes. So, Thadd did mention the need to reestablish a bad debt reserve for the acquired business, because there wasn't one previously in our accounting, and that was a $3 million SG&A hit that was one-time to this quarter. We also have the inefficiencies of processing payments through T-Mobile under the transition services agreement. The $118 million that we have to send them is primarily for them to send to vendors. And while those payments are handled remotely, we will not have the ability to be as disciplined as I think we are in terms of auditing them and being very aggressive with our vendors. We fully intend to bring that in-house quickly. As it turns out, we were expecting to transition the IS infrastructure from T-Mobile to Cogent and run it in parallel for a year. We have been unable to do that due to some of the security concerns that T-Mobile is dealing with. So, we have accelerated our timeline to move all of the customers, all of the network tools and monitoring into Cogent systems in the third quarter. So, by year-end, the legacy Sprint and T-Mobile systems, the roughly 220 software tools that they use, will only be for archival purposes. With that, we absolutely expect to get some additional benefits and cost savings that we have not fully quantified, that we just said they are probably better than our models project. We also, in terms of headcount, when we started looking at this business, there were almost 1,800 employees. When we signed our purchase agreement last September, there were 1,320 employees. We ended up acquiring 942 of those employees. We understand that we will probably have some more headcount that doesn't fit well in the Cogent model. We, as part of our total agreement with T-Mobile, have the ability to pay severance to those employees and have that severance funded by T-Mobile. So, we don't anticipate that being an additional drag. And we do expect that there will be, both through realignment retirements, as Thad said, the average employee's been here 22 years, we'll probably see another 100 to 150 people on the operation side eventually exit the combined company, resulting in additional incremental SG&A savings. We have office consolidations underway that are part and parcel of those savings numbers. So, I think there are some additional SG&A benefits that are not fully baked into our model. We kind of focused on the network first.

Okay. And just to be clear, with respect to severance, from an accounting perspective, is that reported on a gross basis when you get reimbursed by T-Mobile, or is it reported on a net basis? How should we think about that?

The entire cost of the severance would not hit our P&L since it's reimbursed by T-Mobile. So, it would be our cash out the door when it's paid and then billed to them on the next month, and then we would be reimbursed. So, when you see the amount due from T-Mobile under the RTSA, a lot of that will increase after June because of severance amounts that were paid. But we have no P&L impact on that.

Right. We can't report that as revenue. It wouldn’t make sense.

It's not revenue and it's not expense. It's just a cash transaction.

Okay, got it. And last, just kind of quickly, sorry for dragging this out, what was behind the decision to treat the transit payments in EBITDA? Or rather to include them based on when they're paid rather than when they're billed or accrued?

We had to reconcile to the cash flow statement, and you can't reconcile accruals to that. EBITDA is a non-GAAP measure. We felt that it was critical that investors understood this money was coming in and how it impacted our ability to pay our bills and meet the cash burn of the acquired operating business. And it made the most sense to reconcile it to the cash flow statement. And as such, you can only recognize it as you receive it, not as you bill it.

Okay, got it. Thanks guys.

Your next question comes from the line of Brandon Nispel with KeyBanc. Your line is open.

Hey, guys, it's Evan on for Brandon. I guess first question, with the 100 basis point margin expansion, is that based off of the mid-30s that we're at from 2Q? Just trying to gauge where we're going to end up for the EBITDA at the end of the year. And I was just wondering if you guys have had any more learnings about dark fiber. It seemed like you were still not quite sure what the market was looking like or what the demand would be like. So, just wondering if there's been any developments with that. Thank you.

Hey, Evan, I'll take those in reversal orders. So, we are absolutely committed to monetizing the assets we have, and that includes selling dark fiber. We need to better understand the inventory and the demand set on a route-by-route basis. We have had a number of requests for dark fiber. We are not in a position to start selling that because we don't have a complete enough view of the demand set, the pricing we could achieve and the overall inventory that we have. We will be in a position probably in about a year to really consider those dark fiber sales. They are not baked into our model at all. There is nothing in there. So, it is, again, something that is 100% margin and completely additive. But we want to be, I think, thoughtful before we sell that long term asset or lease it on a long-term basis. To your question around margins, the overall trend is to see 100 basis points a year of margin expansion over say the next decade or so. The inclusion of the payments from T-Mobile are in that, but we also know those payments will be stepping down. So, I think you need to look at it on a long-term basis rather than try to use that 100 basis points and divide it into a 25-basis point improvement every quarter sequentially. As I said, over a 10-year period, in answering Tim's question, we will achieve those types of goals, but there's going to be a lot of moving parts here, whether it be these incremental sales and margin opportunities that we've discussed, whether it be incremental savings, and also the decline in both commercial services to T-Mobile, as well as ultimately the sunsetting of their transit payment to us. So, really think of these as long-term. You're not going to trick us into giving you a year-end EBITDA number.

Okay. I had to try. Thanks.

And your final question comes from the line of Michael Rollins of Citigroup. Your line is open.

Thanks. And good morning and thanks for fitting the questions in. Two more if I could. First, when you look at the contracts of the acquired customer relationships, what's the pacing that you'll be able to manage each of these relationships over to either push them into a product menu that's good for Cogent, or ask the customer to consider alternatives, given the direction of focus for Cogent and the acquired assets? And then the second question would be on just CapEx. If you can give us an update of how to think about the heritage pacing of CapEx through the rest of this year and next year, as well as the opportunities to invest that capital you're describing to do the integration and augment the assets that you purchased to enable all the new products and services that you want to offer. Thanks.

Yes. Hey, thanks. Thanks for sticking around so long, Mike, to ask a question. And let's start with the customer contracts. Probably the longest duration customer contracts that we acquire from Sprint run through the end of 2026. I would say the average remaining term on the contract is about 18 months. The contracts, though, are very custom, very bespoke, and they fall into two primary categories. One is for the services as they exist today. The second is for incremental services going forward. And virtually all of these customers are continuing to do moves, adds and changes, so therefore incremental services. So, each time one of those requests for new services are originated, we have the opportunity to have a discussion about modernizing the products. Finally, these customers have been notified by T-Mobile as part of the purchase process, and have been reminded by Cogent about sunsetting certain products and end-of-lifeing them. Those are those 19,000 non-core products and the $8.4 million of revenue in the quarter that came from those non-core products. We will continue to manage those out. There have been some customers said, please give us a little more time. And we've addressed those on a customer-by-customer basis. For the moves, adds and changes, there are three changes that customers are being informed of. One, our desire only to sell services delivered over fiber. So, therefore, we are sunsetting local access loops that are delivered over copper, coax, wireless or satellite. And we're not religious about this in the sense that, yes, we’ll accommodate a customer if there's no other alternative in a short term, but long-term, we are looking for quality of service purposes and scalability only to use fiber. The second big change is international sales in countries that we are not licensed. So,…

Thanks, Dave.

There are no further questions at this time. I will turn the call back to Mr. Dave Schaeffer for closing remarks.

Well, thank you all very much. Hopefully, this was not too confusing. We’ve tried to be as transparent and consistent as possible. I want to thank everyone for their patience for being on this call. And if you had to predict, the one surprise I didn't hear anyone ask is how we reported $24 of EPS in a quarter. But I can assure you we won't repeat that again next quarter. Take care, everyone. Stay well, and we're available to answer follow-on questions. Thanks.

This concludes today's conference call. Thank you for joining. You may now disconnect your lines.