Good afternoon, everybody. Tim Horan, the Communications Analyst here at Oppenheimer. My pleasure to be hosting the CEO, Dave Schaeffer from Cogent Communications Holdings Inc. Dave, I know it's been a very trying week or a few months for you. Not trying to sugarcoat anything. I know this is the third one you've done this week, and I apologize I did not get to read the full transcripts of the other ones. Maybe we should try to focus on what hasn't been discussed too much. The main thing I'm trying to figure out is, I think, you have guidance for 2028 or 2029 around $1.5 billion in revenue and $0.5 billion of EBITDA, give or take, in that time frame. Could you just maybe update us on that longer-term aspiration and what, what, you know, what's going to get us there?
What are the steps that we should be kind of tracking to reach those targets?
Yeah. First of all, thanks for hosting me, Tim. Thanks, Oppenheimer, for the venue. Thank investors for caring about Cogent and spending some of your time with us. When we announced the transaction to acquire Sprint, we laid out a five-year, multi-year plan that would get us to $1.5 billion of revenue and $500 million of EBITDA. We started out with a little over $1 billion of revenue combining the Sprint and the Cogent business. Cogent represented about 60% of the combined company, Sprint about 40% of the combined company. The Sprint business for the previous three years to closing had been declining at 10.6% a year. The Cogent business during that same period had been growing at about 7%, about 5% at the end of that period because we did have some impact from COVID.
We intentionally accelerated the decline in the Sprint business by grooming unprofitable products and unprofitable locations. As a result, for the two years post-transaction, Cogent has had negative revenue growth but has grown its underlying EBITDA on average $5 million a quarter, sequentially each and every quarter for those eight quarters. While we have some additional cost savings and network optimization ahead of us, most of the growth going forward in EBITDA will have to come from the ability to grow top line, in particular, selling higher margin products replacing lower margin products. We also were fortunate enough in the negotiations to get a subsidy payment from T-Mobile that has been added to our underlying EBITDA. That subsidy payment was $700 million over 54 months, asymmetrically front-end loaded.
Our EBITDA on that $1 billion of revenue jumped from $260 million to $350 million in the first year, $348 million in the second year, and it will be somewhere above $300 million- $350 million this year based on the step down in those subsidy payments coupled with the cost savings that I outlined. For the full year, we will still probably be negative, flat to slightly negative. On a sequential basis, we anticipate our top line revenue returning to growth sometime in this quarter, Q3 of 2025. From this point going forward, growing at a combined rate of between 6% and 8%. Within that revenue growth, there are really three independent revenue streams. There is the legacy Cogent classic business that is growing at about 5% year- over- year, not as well as it was pre-pandemic, but still a growing business with contribution margins of about 100 basis points.
There is an acquired Sprint enterprise business that has been migrated onto the Cogent network. Costs have been taken out, and undesirable products have been groomed. That business represents less than a third of our total revenue and is a flat business with much lower margins. Finally, we have a new business that is effectively a startup in a public company, which is our wavelength business. That wavelength business was zero at acquisition. It has grown to a $36 million annual run rate by last quarter. We were not able to actively sell those wavelengths until the beginning of this year. We anticipate that that business alone will be a $500 million business with roughly a 95% contribution margin on those incremental waves sold.
The combined company, having the new products, the legacy customer base, and the traditional Cogent, will be a business that is $1.5 billion in revenue scale by mid 2028 run rate on an annualized basis and will be generating $500 million in EBITDA. There has been concern among investors that as that T-Mobile final subsidy dissipates, can we both make up the ground that we will lose when that last $100 million a year of subsidies for the next two and a half years disappears and add this additional EBITDA from top line growth? The answer is absolutely yes.
Dave, just to clarify the numbers, to hit a $1.5 billion run rate by mid 2028, you're talking more like on a consolidated basis, like 12% revenue growth the next three years. Is that right? Your guide, I know 6%- 8%, is that absent wavelength or does that include everything? Because if you include everything, if you're going to grow 7%, it's a number much, much less than $1.5 billion. Yeah.
It is a higher growth rate in the near term. The 6%- 8% number that we give is meant to be a multi-year long-term trend line, coupled with the 200 basis points of continued margin expansion. In the near term, to hit our targets, we anticipate aggregate growth rate to be above that and average contribution margins to be higher.
Yeah, I mean, not to put words in your mouth, but yeah, you'd have to have margin expansion more like 300 basis points a year the next, you know, 300 basis points- 400 basis points in the next three, four years. The revenue growth I laid out like 12% the next three, four years. You're still pretty confident you can do that. I mean, a lot of that is driven by wavelength, obviously, which we're looking for basically hockey stick growth at this point.
The answer is yes. With the wavelength business, it is a repurposing of the dormant voice network that Sprint had deployed, connecting it to our metro footprint and selling a new product. We have enabled 938 data centers. We've now sold waves as of the end of the quarter in a little under half, 428 of those facilities. We have less than 1% market share in the wavelength market. We have a compelling value proposition of four key advantages: one, more data centers; two, faster delivery; three, lower price; and four, higher reliability. Based on those competitive advantages, we are targeting a 25% market share, up from 1% over the next three years. The period of time between closing and the beginning of this year was focused on that network enablement. Now we're in the process of converting orders into installed revenue and accelerating the build of the funnel.
That will result in a much higher growth rate in the short term. Remember, we were fighting the headwind of declining revenues, intentionally declining due to the decision to terminate a number of legacy Sprint services. That decision turned out to be correct because we were able to take the underlying EBITDA absent the subsidy payments from less than $5 million a quarter to $48 million a quarter in eight quarters, or a better than $5 million a quarter sequential improvement, which resulted in actually faster EBITDA margin expansion than we're projecting going forward. While we have some additional savings, and that may account for another two or three quarters of improvement in EBITDA, most of the improvement will come from wavelength sales. There will be continued contribution from legacy Cogent services, whether it be transit on net internet, on net VPN services, or address leasing.
The wavelength driver will be enabling us to get from $1 billion of revenue to $5 billion and then getting EBITDA margins up to mid-30s from where we are today. While we are at 29.8% last quarter, that was partially helped by the $25 million subsidy payment from T-Mobile. Those subsidies will continue through early 2028. At that point, we need to be on a run rate by mid-year of $125 million of quarterly EBITDA with no subsidy payments. The progression to get there is selling on net services.
Great, great detail. Thank you. I guess it comes down to, you know, obviously it's taken a little bit longer to get the wavelengths up and running in terms of revenue and actually the number of wavelengths. Can you just talk about how sales are progressing at this point? Maybe there's the most recent kind of few weeks and, you know, how provisioning is going?
Yeah. The first effort we had to do once acquiring Sprint was integrate the networks and convert the Sprint voice network to a wave network. That program was substantially complete at the end of 2024. While that conversion was ongoing, we were validating our demand assumptions by going to the market with a footprint that was not yet installed and a timeline of when we could not guarantee delivery. We assembled almost 10,000 orders during that period. About 1,000 of those were manually installed between May of 2023 and December of 2024. At the end of 2024, the network integration and optimization was complete, and we began selling orders with specific delivery windows. Of that original 10,000, 1,000 installed, 1,000 carried forward, and about 8,000 of those orders disappeared. Going forward, we now have 938 specific locations.
We have a guarantee of delivery, and we've got 296 of our 628 salespeople focused on this market segment. We have built a funnel of nearly 4,700 orders. That includes orders that have been installed but have not yet started billing, those that are in installation, those that are under firm contract, and then finally those in which the specific route, price, and speed have been negotiated, but the contract has not yet had the terms and conditions executed. We anticipate that funnel number to build. We also anticipate that the 200 orders- 300 orders that we installed in Q2 but did not begin billing will bill over the next quarter or two. More orders will come into that window as the provisioning we are doing. We should end this year with about double the number of waves that we had at the end of Q1.
We went from 1,000 being manually installed in that 18-month period while we were building the network. The first two quarters when the network was operational, we installed basically 500 revenue billing and another 250 that were not yet billing revenue. We need to get the revenue build numbers up to 3,000 by year-end, but then continue to grow that. I think as we demonstrate the superiority and the value we deliver, that should be easily done. We have gotten a great deal of comfort from customer feedback and from the competitive dynamic that the addressable market is actually a little better than we expected it to be.
Very, very helpful. Your confidence level in the 3,000 by year-end, you think there's an 80% chance of hitting that number, 90%?
I think 90% is a reasonable target.
Okay. You know, how many wavelengths do you think you can have installed for 2026?
The pace will pick up in 2026. We will have a credible set of reference accounts, and we should be able to triple that number, but somewhere between double and triple the number of wavelengths that we started the year with. While we don't want to give specific granular guidance on a quarterly basis, we want to stick to that multi-year set of metrics. We should be seeing the wavelength revenue growth of GAAP revenue progress. What we want to do is then start to focus exclusively on revenue and not on these other KPIs.
I greatly appreciate that. I guess while we're on the KPI, the ARPU has been coming in pretty healthy. It sounds like, you know, are you pricing at a 10% discount to the market now and parity with the market? You know, how's the pricing been shaping up?
Yeah, ARPUs actually went up in this quarter, and pricing has been stronger than we had initially expected. We typically go to market with a 20% discount to what we believe the comparable route and speed is from our competitors. We built a pricing engine that took into account both the competitive dynamic on a route-by-route basis and any incremental costs that we had. In building that tool, we optimized it for a 20% discount to the market rate. We knew that we would have to discount beyond that. To date, we have not done nearly as much discounting beyond that 20% as we anticipated. To refresh everyone's memory, in our transit business, we typically discount 50% to market rates.
Is that 50% still holding, do you think, Dave, on transit?
Yeah, on transit, I think we will continue to be the price leader. We continue to be at about a 50% discount, although in many international markets, the competitive dynamic is very different. Prices are much higher, and we have not been willing to necessarily be as aggressive against regional local players who are confined to one market. Part of the reason why our net-centric revenue growth has outpaced the Px Q of volume growth and price declines has been our ability to get better pricing in international markets.
The net-centric business this last quarter, it grew pretty healthy, 7% year- over- year?
That is correct. Even though traffic only grew 9% year- over- year and prices continued to decline in the roughly 20% range, we were able to get more of the revenue sales from smaller customers and international. Some of it's by Cogent's marketing and design, but some of it is a shift in the underlying market itself. The internet is growing faster in the developing world than it is in the developed world. The market concentration that has occurred with, you know, seven or eight key names accounting for about half of internet traffic has pretty much played out. While we sell to all of those players, we're seeing a broadening out in growth of mid and smaller players, which results in a higher price per meg.
Dave, why don't you, I mean, just the net-centric business, you know, it's not growing that much. You know, why don't you take the discount over time to 25% instead of 50% or some lower number?
That effectively is happening in part through this international strategy where we have limited competition in a given market and therefore are not as aggressive. If you went to Argentina or Chile, the number of players in that market are very different than going to the Netherlands or to Germany. You know, we are going to be entering the Indian market this quarter. It took us eight years to get to this point in license applications. I don't want to jinx ourselves. We're at the one-inch line. I think when we go into that market, prices will be higher and we will not be as aggressive just because the dominant players are buying from us today out of market and bringing traffic in the country. We will continue to support that business where we will be very aggressive, but within the country, the competitive dynamic is very different.
I don't think we need to be quite at a 50% discount.
Got it. Very helpful. Can you talk a little bit about the legacy, if you don't mind, just what you're seeing in the legacy corporate business, what that's kind of growing at? I know it's a little difficult to parse out.
Yeah. Our corporate business, which is driven by our net-centric business services in large multi-tenant office buildings, had grown for a decade between 2010 and 2020 at a compounded rate of 11%. When the pandemic hit, that growth rate materially decelerated to - 9%. It returned to about 4% growth, and for the last couple of quarters, it's actually been growing at about 3%. We do think there will be gradual improvement. Layer on that the acquired Sprint corporate business, which had almost exclusively been delivered off-net and had a number of non-core products, that business is probably declining at 20%+ . As a result, the entire corporate business is still slightly negative, but the rate of decline in the acquired Sprint corporate base is also moderating as we have worked our way through those undesirable products and locations.
When does Sprint kind of reach stability, and can it ever grow again?
When we acquired Sprint, 80% of the revenues were enterprise, roughly 17% of revenues were corporate, and about less than 3% were net-centric. That business was a declining business and probably will not return to material top line growth. I think those customers that we acquired from Sprint have been stabilized, but it will be a flattish business. In the corporate segment, we're actively selling those services with over 300 corporate reps selling. That business, in its entirety, with the remaining Sprint legacy customers, will end up growing in that kind of 5%-ish type range that Cogent's business had grown. The true enterprise customers, of which Cogent only had three prior to the acquisition, now we have about 250. They account for about 30% of our revenues and are flat. They are roughly 88% off net. They tend to be smaller locations and global in nature.
We think we can get that business to flat to up 1%, which compared to other wireline enterprise companies is a good result. A large percentage of those customer locations will never be on net for Cogent. We view that as a maintained business, not a growth business. The net-centric portion of our business without wavelengths is a very desirable business. It's a business that's roughly 90% on net and only 10% off net. It's a business growing at 7%, 8%. That business will continue to perform. It tends to be a little bit lumpy. Everybody got excited at the beginning of the pandemic when net-centric revenue growth accelerated to 26% year- over- year from 3%. It's since moderated to that kind of 7%, 8% range that you mentioned, Tim.
We layer on wavelengths, which are almost exclusively a net-centric product, and we get to a business that is growing in mid to low double digits. Because all of the wavelengths and roughly 90% of the IP-based services in net-centric are on net, the contribution margins are much higher. That's how we get to the EBITDA number with the revenue growth.
Dave, we have eight minutes. I have 30 questions from investors. I have another 30 myself, so maybe we can go in a quick order here. I hate to ask this, but can you talk about, I know you've had a bunch of stock sales for a bunch of different reasons I don't want to go into, but can you talk about how many stocks you still own and do you have more to sell?
Oh, I have nothing left to sell. I worked at Cogent for 25 years, never drew a salary, and received 100% of my compensation in stock. To pay the taxes on that stock, I had borrowed against it. I had a gross tax basis in those awards of $185 million. I subtract the roughly $30 million in return of capital payments over that time and had a net basis of $155 million. I had borrowed about $80 million against 2.7 million shares with two banks. When the stock sold off, I was literally flushed out. The only stock that I have left are the shares that are held by the company that are restricted. Those restrictions will lift. I will be able to use that stock, but I have no longer any debt on it.
If there's a silver lining in a very dark cloud, I now have a loss carry forward on my Cogent stock. As that stock vests, I won't have to pay any taxes.
Got it. Okay. Sorry about that.
That's not your fault. It's my fault for taking my company stock and paying the taxes.
Why don't you just cut the dividend and buy back stock with the stock so cheap? Why are you still paying a dividend that no one believes you can support?
Cogent has returned $2 billion of capital to shareholders. We have bought back 10.9 million shares or about 23% of our float. We have grown our dividend for 52 sequential consecutive quarters. We have done that by growing EBITDA and levering that EBITDA up. We've returned more than cash, and our leverage has peaked. It will naturally come down along with the growth in return of capital. Not just even at the current levels, but returning more each quarter, we can naturally delever based on the extrapolation of the EBITDA growth that we have achieved over the past eight quarters, just extrapolating going forward. It is true that to be able to achieve $5 million plus of EBITDA incrementally each quarter, for a quarter or two more, we have some cost savings, but then it has to come from top line growth. We will delever.
We were very clear saying our leverage peaked in the second quarter of 2025 at 6.61x net. It will come down. Whether or not we do it exclusively through a dividend or we pivot more to buybacks is really dependent on the stock price. With such a profound dislocation in the stock, the IRR on the buyback is very high, much higher than our net cost of capital. It may make us pivot more to buybacks. I also am cognizant of a number of shareholders who buy Cogent and have supported us because of our dividend consistency. The message I want to leave everyone with is we are going to return an increasing amount, not even a constant amount, but an increasing amount of free cash flow and monies to shareholders each quarter.
Very helpful. Lastly, any more update on the data center or wavelength, or asset sales in any form?
The wavelength we've touched on, the data center conversion and optimization is complete at this point. You had the opportunity to tour one of those early in the conversion process. We have been working diligently, investing capital, converting the DC plant to AC. We are talking to a number of parties. We have six letters of intent. I actually had just between my last one of your meetings, they gave me a break and I responded to two more new entrants who expressed interest. I put them in touch with the right parties at Cogent. We are very motivated to sell non-core assets. We cannot fill up these data centers. We are looking to either sell them or lease them or joint venture them, but we know that they are not strategic.
We also know that we have excess IPv4 addresses beyond the roughly 13 million addresses that are securitized and the 15 million that are currently leased. We are aggressively continuing to lease them. We've ramped the revenue run rate on that product from $8 million a year to $60 million a year in three years. It is 100% margin. We will continue to do that, but we will also try to see if we could sell some of those. The enterprise value creation at Cogent doesn't come from selling assets. It comes from growing recurring revenues. That is our primary focus.
Dave, we're out of time. I really, really, really appreciate all the help here. Good luck and thank you.
Hey, thanks everyone. Thanks for hosting me, Tim. Take care. Bye-bye.