Cogent Communications Holdings Inc
NASDAQ:CCOI
Cogent Communications Holdings Inc
In the bustling world of telecommunications, Cogent Communications Holdings Inc. stands out as a robust player, weaving its narrative through the ever-growing demand for bandwidth and connectivity. Founded in 1999, this Washington, D.C.-based company has carved a niche for itself by specializing in internet access, data transport, and colocation services. Cogent's success largely hinges on its extensive fiber-optic network, which spans North America, Europe, and parts of Asia. The company prides itself on providing simple, fast, and dependable internet service without the complex pricing structures often associated with larger telecom giants. By focusing on wholesale bandwidth and Internet access, Cogent became an industry staple by enabling a streamlined user experience, which it markets directly to businesses, internet service providers, and telecommunications carriers.
Revenue generation for Cogent is a straightforward affair, reflecting the company’s dedication to simplicity and efficiency. The business model revolves around selling data transport and connectivity solutions to customers who are increasingly reliant on seamless internet services. Cogent effectively monetizes its extensive network by leasing fiber-optic lines and operating data centers, offering competitively priced packages that appeal to a wide array of customers, from small enterprises to large corporations. The company’s pricing advantage comes from its investment in a modern and expansive network and a strategy of acquiring dark fiber to expand its reach. As more activities migrate online, Cogent's capability to offer high-capacity, low-latency connections secures its place in the expanding digital landscape, ensuring a reliable revenue stream that aligns with the rapid pace of technological advancement.
Earnings Calls
Cogent Communications achieved robust revenue growth in 2024, reaching $1 billion, up from $940.9 million in 2023. Key drivers included a 240% surge in wavelength revenue and a 24.5% increase in IPv4 leasing revenue. Adjusted EBITDA was $348.4 million, with margins expanding by 100 basis points annually, targeting continued 5-7% growth. Cost savings from Sprint integration are projected to surpass the initial $220 million goal. With a solid cash position of $227.9 million, the company plans a dividend increase of 4.1%. Management expects to maintain stable EBITDA and revenue, bolstered by a backlog of 2,700 wavelength opportunities.
You don't have any saved screeners yet
Good morning and welcome to the Cogent Communications Holdings Fourth Quarter and Full Year 2024 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 at Cogent's website when it becomes available. Cogent's summary of financial and operational results attached to its press release can be downloaded from the Cogent's website.
I would now like to turn the call over to Mr. Dave Schaeffer, Chairman and Chief Executive Officer of Cogent Communications Holdings.
Thank you, and good morning to everyone. Welcome to our fourth quarter 2024 earnings call and summary of our full year 2024 results. I'm Dave Schaeffer, Cogent's Chief Executive Officer. And with me on this morning's call is Tad Weed, our Chief Financial Officer.
Hopefully you've had a chance to review our earnings press release. Our press release includes a number of consistently reported historical metrics, which hopefully are helpful in understanding trends in our business.
Our combined Cogent business had a good quarter and a good year. Our total revenue for the quarter was $252.3 million, and $1 billion for full year 2024 compared to $940,900,000 for full year 2023. Our EBITDA, as adjusted, was $66.9 million for the quarter, and $348.4 million for the full year 2024 compared to full year 2023 of $352.5 million. Our EBITDA, as adjusted, for the quarter increased sequentially by $6 million. And our EBITDA, as adjusted, margin increased sequentially by 280 basis points to 26.5%.
Our wavelength revenues for the quarter grew sequentially at 31.8%, or $7 million, an increase of 124% over the previous year. Our wavelength revenue was $19.2 million for full year 2024, a 240% increase over 2023.
Our IPv4 leasing revenue for the quarter increased sequentially by 11.8% to $12.6 million, and that represents a 27.2% increase year-over-year. Our IPv4 leasing revenue was $44.9 million for full year 2024, an increase of 24.5% over the previous year.
Our network traffic in the quarter was essentially flat, and up 11% year-over-year on a quarterly basis. Our network traffic for the full year 2024 grew by 16% over 2023.
We are in the process of continuing to realize significant cost savings from the integration of the Sprint assets. We have realized over 90% of our targeted $220 million in annual savings. Our projected savings are expected to continue to be achieved through 2026 and will be greater than the $220 million initially targeted.
Our SG&A decreased by $4.5 million quarter -- for the quarter or by another 7.5% from the previous quarter, and it decreased by $19.2 million or 25.6% from Q4 of 2023. Our SG&A as a percentage of revenues decreased 22.2% -- to 22.2% this quarter from 23.7% in Q3 of 2024.
Our cost of goods sold decreased by $6.4 million in the quarter or 4% from the third quarter of 2024 and decreased by $19.5 million or 11.2% from the fourth quarter of 2023.
We ended the year with $227.9 million of cash and cash equivalents on our balance sheet.
Our sales force rep productivity was 4 units per rep per month in the previous quarter and 3.5 units per rep per month in the fourth quarter. Our sales reps headcount and productivity in connection with the Sprint acquisition, we hired a total of 942 employees in May of 2023. At year-end, 624 of these employees remain employed with Cogent. This represents a 34% reduction in the former Sprint employee base.
Now, for a comment on our wavelength and optical transport business, in connection with our acquisition of the Sprint network assets, we expanded our product offering on our fiber-optic network to include wavelength services or optical transport services to both existing customers, as well as new customers. At the end of the quarter, we had wavelength sales capability and connectivity to 808 locations throughout North America. This exceeded our target of 800 sites by year-end with provisioning times of approximately 30 days. We ultimately will be able to reduce that provisioning time to 2 weeks. But at this point, it is in fact at around 30 days. We have actually sold waves and installed them in 280 locations. And as we go through our existing sales funnel and backlog, some of those orders have dropped out, and at quarter-end, we still had 2,700 orders in our wavelength funnel.
Our Sprint network acquisition materially expanded our data center footprint. At year-end, we have reconfigured 115 acquired Sprint facilities and added those to our 1,646 carrier-neutral data center footprint. There are 104 Cogent data centers, which have 177 megawatts of power. We have also converted some of the smaller Sprint facilities. 55 of these former facilities are now Cogent edge data centers. These edge data centers are smaller in their footprint and can typically support about 40 racks, and in aggregate, have an additional 20 megawatts of power. So in total, we have a combined Cogent footprint of 159 data centers with 197 megawatts available for customers. We are in the process of decommissioning some legacy Cogent data centers and leased facilities, where they are redundant with our [ honestly ] simple acquired Sprint facilities.
Our Board of Directors, which reflects on our growth in cash flow and our continued strong cash flow generating capabilities, as well as our investment opportunities, once again approved the increase in our quarterly dividend by another $0.01 per share sequentially for the quarter, raising our quarterly dividend from $0.995 a share to $1.005 per share. This represents the 50th consecutive sequential increase in our quarterly dividend and a dividend growth rate of about 4.1%.
Now that the Sprint business is combined with our legacy Cogent business, we anticipate annual growth of 5% to 7% and EBITDA, as adjusted, margins expanding by about 100 basis points per year. Our revenue and EBITDA guidance are designed to be multi-year targets and are not designed to be specific quarterly or annual guidance. Our EBITDA, as adjusted, and our leverage ratio are impacted by the timing and sequencing of payments received under our $700 million IP agreement with T-Mobile as part of the purchase of these assets.
Now, I'd like to turn it over to Tad to read our safe harbor language, provide some additional detail on our operating performance, and then I'll follow up with some summary comments, and then we'll open the floor for questions and answers.
Thank you, Dave, and good morning, 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.
If 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.
We analyze our revenues based upon network connection type, which is on-net, off-net, wavelength services and non-core. And we also analyze our revenues based upon customer type, and we have 3 customer types: NetCentric, corporate and enterprise customers.
Our corporate business represented 44.8% of our revenues for the quarter, and that was $113.1 million. Our quarterly corporate revenue decreased by 10.7% year-over-year and sequentially by 2.7%. These decreases in our corporate revenue are primarily due to the continued grooming of low-margin off-net customer connections and the elimination of non-core products. We had 46,371 corporate customer connections on our network at year-end.
Our NetCentric business continues to benefit from the continued growth in video traffic, activity related to artificial intelligence, streaming and wavelength sales and was $93.6 million for the quarter. Our NetCentric business represented 37.1% of our revenues for the quarter, which was an increase of 0.5% year-over-year and a sequential increase of 1.9%. Our quarterly NetCentric revenue under our commercial service agreement with T-Mobile declined sequentially by $2.6 million, and declined by $7.1 million year-over-year, negatively impacting our NetCentric revenue results. We had 62,236 NetCentric customer connections on our network at year-end.
Our enterprise business represented 18.1% of our revenues for the quarter and was $45.6 million. We had 14,776 enterprise customer connections at the end of the year on our network. Our quarterly enterprise revenue decreased by 12.8% year-over-year and sequentially by $3.5 million or 7.1%, primarily due to a reduction in non-core and low-margin enterprise revenues.
On revenues by network type, on-net revenue: we serve our on-net customers in our 3,453 total on-net buildings. We continue to succeed in selling larger 100 gigabit connections and 400 gigabit connections in carrier-neutral data centers and selling 10 gigabit connections in select multi-tenant office buildings. Our on-net revenue was $128.8 million for the quarter, a year-over-year decrease of 6.7% and sequentially 5.7%.
Our on-net revenue results were negatively impacted by 3 items: the $2.6 million sequential decline in the commercial services agreement, the on-net revenue component of that with T-Mobile; a $1 million of negative FX sequentially; and revenue from a low-margin resale customer we acquired in the Sprint acquisition that we mentioned last quarter and [ essentially ] terminated. That was $1.7 million of the decline. Our on-net customer connections were 87,500 at year-end.
Our off-net revenue was $113.2 million for the quarter, a year-over-year decrease of 8.5% and a sequential increase of 1.7%. The sequential increase is primarily due to increase in our margins on former Sprint off-net services. Our off-net revenue results are also impacted by our migration of certain off-net customers to on-net and the continued grooming and termination of low-margin off-net contracts. Our off-net customer connections were 28,963 at the end of the year.
Our wavelength revenue was $7 million for the quarter, a sequential increase of 31.8%, and year-over-year, 124%. Our wavelength customer connections were 1,118 at the end of the year. Our IPv4 leasing business had another good quarter and a very productive quarter. We were leasing $13 million of our IPv4 addresses at the end of the year, and our IPv4 revenue increased by 11.8% from last quarter and 27.2% year-over-year to $12.6 million for the quarter.
Lastly, non-core revenue was $3.4 million for the quarter. That was a $0.8 million decrease, and a year-over-year decrease of $3.9 million, again, due to our decision to end of life these non-core products.
Some comments on pricing. Our average price per megabit for our installed base decreased sequentially by 9% to $0.21 and decreased by 25% year-over-year, consistent with historical trends. Our average price per megabit for our new customer contracts was $0.11, which was actually an increase of 31% sequentially and 11% increase year-over-year.
Our ARPU -- our on-net ARPU decreased sequentially by 5.7% from $520 to $490. Our off-net ARPU increased sequentially by 8% from $1,138 to $1,229, demonstrating our increased margins. Our wavelength ARPU increased sequentially by 9.5% and was $2,151 this quarter compared to $1,964 last quarter.
Our average revenue per IPv4 address sold was $0.44 per address this quarter. That is a 47% increase from the average base of $0.30 per address at the beginning of the year. So prices have been increasing.
Our churn rates are relatively stable. Our on-net unit monthly churn rate was 1.3% this quarter compared to 1.2% last quarter. And our off-net unit monthly churn rate was 2.6% this quarter, the same as last quarter.
Comments on EBITDA, EBITDA classic: we reconcile our EBITDA to our cash flow from operations in each of our quarterly earnings press releases. Our EBITDA for the quarter increased sequentially by $6 million, and our EBITDA margin increased by 270 basis points to 16.6%. For the full year, our EBITDA was $122.8 million and our EBITDA margin was 11.9%. EBITDA, as adjusted: our EBITDA, as adjusted, as a reminder, is adjusted for Sprint acquisition costs if we incurred any during the period and payments under the IP Transit agreement with T-Mobile. There were no incurred and costs classified as Sprint acquisition costs after the second quarter of 2024, when our 1-year purchase accounting adjustment period ended for the Sprint acquisition. Our EBITDA, as adjusted, for the quarter was $66.9 million, and our margin was 26.5%. That's a sequential increase of 280 basis points. Our EBITDA, as adjusted, for the year was $348.4 million, and the margin was 33.6%.
In accordance with our IP Transit Services Agreement, we received the 3 monthly payments totaling $25 million this quarter, the same as last quarter. All payments have been paid in full and on time. For full year 2023, we received total payments of $204.2 million, and coincidentally, the same amount for full year 2024, $204.2 million, but the payment streams were different. The total was the same, but the streams were different. The total payments for 2023 included 7 payments of $29.2 million each, and the total monthly payments for 2024 included 5 payments of $29.2 million and 7 payments of $8.3 million. An additional 35 monthly payments of $8.3 million each will continue until November 2027. There are additional payments we will receive under the purchase agreement with T-Mobile. There are further payments related to lease obligations that we assumed at closing, and these payments will total at least $28 million and to be paid over 4 equal payments in months 55 to 58, so from November '27 to February 2028.
Foreign currency comments: our revenue earned outside of the United States is relatively consistent with last quarter and was about 18% of our revenues. 11% of that was in Europe and 7% related to Canada, Mexico, our Oceanic, South American and African operations. Our average euro to USD rate so far this quarter is $1.04, and the average Canadian dollar exchange rate is $0.70. If those rates continue at the current level, the FX conversion impact on sequential quarterly revenues would be negative and about $1 million, and year-over-year for quarterly revenues would also be negative and about $1.7 million.
Our revenue and customer base is not highly concentrated. Our top 25 customers was about 18% of our revenues this quarter.
On CapEx, our CapEx was $46.1 million this quarter. That was down 22.2% from last quarter, and our CapEx was $195 million for the full year. We are continuing our network integration of the former Sprint network and legacy Cogent network into one unified network and converting former Sprint switch sites into Cogent data centers. We have accelerated and expanded our data center conversion program due to the high level of demand for our power availability. This program will require capital spending for the first half of 2025 that will be similar to the last half of 2024 and then tail off.
Comments on our IRU lease obligations: our finance IRU obligations are for long-term dark fiber leases. Our IRU finance lease obligations were $538.4 million at the end of the year. We have a very diverse set of suppliers and contracts with 369 different suppliers of dark fiber.
At year-end, our cash and cash equivalents and restricted cash was $227.9 million. $29.4 million of cash is restricted, and $22.3 million of that was tied to our swap and $7.1 million was tied to the requirements under our IPv4 restricted notes.
Some comments on debt and debt ratios. Our total gross debt at par, including our finance leases, was $2 billion at year-end, and net debt was $1.8 billion. Our total gross debt to the last 12 months EBITDA, as adjusted, ratio was 5.72, and net debt was 5.07. Our leverage ratio as calculated under our note indentures was 5.81, and our secured ratio was 3.38. And our fixed coverage ratio under the note indentures was 2.76. Our fourth quarter 2024 annualized EBITDA, as adjusted, increased sequentially by $24 million or 9.8% from $243.4 million to $267.4 million.
Some comments on the swap. As a reminder, we are a party to an interest rate swap that modifies our fixed interest rate obligation with our $500 million 2026 notes to a variable interest rate obligation based on the SOFR rate for the remaining term. The fair value of our swap agreement decreased by $7.6 million from last quarter to $22.3 million. Changes in the fair value of the swap agreement are required to be classified with interest expense under U.S. GAAP.
Lastly, some comments on bad debt and days sales outstanding. Our days sales significantly improved at year-end and was 29 days versus 32 days at the end of the third quarter. And our bad debt expense also substantially improved and was only $6.6 million, only $0.6 million -- and only [ 0.2% ] of our revenues for the quarter and was 1% of our revenues for full year 2024, which is consistent with historical results.
As always, I want to thank and recognize our worldwide billing and collections team members for doing a fantastic job in serving our Cogent customers.
And with that, I will turn the call back to Dave.
Thanks, Tad. I'd like to highlight a couple of strengths in our network, our customer base and our sales force. We continue to experience significant year-over-year traffic growth in our legacy NetCentric business. We are direct beneficiaries of over-the-top video, artificial intelligence activities and streaming traffic. At year-end, we had 1,646 third-party carrier-neutral data centers, 104 of our own data centers and an additional 55 of the smaller edge data centers connected to our network with Cogent's available power capacity of 197 megawatts. The breadth of this coverage enables us to serve our NetCentric customers better, help them optimize their networks and reduce latency.
We continue to expect to widen our lead in this market as we project adding over 100 carrier-neutral data centers to our network this year and for the next several years. At year-end, we were able to sell our wavelength service in 808 of these carrier-neutral facilities with reduced provisioning times. This surpassed our year-end target of 800 carrier-neutral data centers being wave-enabled. At year-end, we were directly connected to 8,250 different networks or autonomous systems, with 23 of these networks representing peers, and 8,227 of these customers were, in fact, purchasing transit from Cogent for their networks.
We remain focused on our sales force productivity, and we are committed to managing out underperforming reps. Our sales force turnover in the quarter was 5% per month. That is down from the peak of 8.7% at the height of the pandemic. It's still below our average rep turnover of 5.6% per month. At year-end, we had a sales force of 288 professionals focused on the NetCentric market. These individuals sell both transit and wavelength services, 347 professionals focused on the corporate market, and 15 individuals focused on our enterprise customer base and the new market opportunities in that segment.
In summary, we're very optimistic about our position in being able to serve small and midsized customers with IP services located in the central business districts of major cities. We had 1,871 on-net multi-tenant office buildings comprising over 1 billion square feet of rentable office space.
We are continuing to provide profitable on-net and off-net services to enterprise customers geographically around the world. We're enthusiastic about the progress and continued opportunity in our optical transport business and the wavelength services that we have been able to provision.
We have a significant backlog and funnel of over 2,700 opportunities. While this is down slightly from the previous quarter, it's been due to grooming that funnel. Many of those opportunities have been in the funnel for over a year as we could not provision them. But now with the ubiquitous coverage and the more rapid provisioning cycles, we believe that we will continue to accelerate our wavelength business. We do have reduced provisioning cycles, and we hope to convert many of these opportunities and funnel into installed orders.
We're digitally -- we are working diligently on continuing cost reductions as a result of the integration of the Sprint network and the various customer bases we acquired. We look forward to the ability to monetize additional parts of our unleased IPv4 address inventory, our dark fiber footprint and our excess data center space on a wholesale basis, either with direct sales or long-term leases over the next year or so. We're currently in discussion with multiple counterparties for multiple sites. And hopefully, we'll be in a position to announce something over the next several months.
Now I'd like to open the floor for questions.
[Operator Instructions] And your first question comes from the line of Michael Rollins with Citi.
Dave, curious -- just to start us off, if you look at the customer verticals that you're now reporting in terms of corporate, enterprise and NetCentric, can you give us an update just holistically in terms of where each of those segments are in terms of getting through some of these legacy acquired revenues? And what's the opportunity for each of these segments to grow over the coming years, whether it's 1-year, 2-year or maybe a 3-year view?
Yes. Sure, Mike. I think those are fair questions. So when we acquired Sprint, we acquired both corporate and enterprise customers, only a handful of NetCentric customers. Our NetCentric business continues to grow, and that growth rate will actually accelerate since virtually all of the wavelength sales will be NetCentric and approximately 85% of our IPv4 leasing is to NetCentric. So within the NetCentric segment, we are growing today, and we anticipate that growth rate to accelerate meaningfully as both the underlying IP business performs and we are able to add wavelength revenue to those customers.
For our corporate segment, we acquired a number of corporate customers from Sprint, who were purchasing non-core services, were purchasing very low bandwidth, [ non-fiber-delivered ] off-net services. And we have been in the process of grooming those less desirable services. It is why our margins have exceeded our guidance in terms of growth rate, and we expect that to continue. I think we're probably still just a couple of quarters away from getting through that grooming exercise and then seeing positive growth in our corporate segment. Those corporate customers will predominantly purchase DIA and VPN services based on VPLS. It is a segment that has been heavily exposed to central business district office occupancy, and we are continuing to see improvements in that. Many markets that had been negative net absorption, meaning increases in vacancy rates since the pandemic, have actually turned positive now, and we're seeing net positive absorption across most markets across the country. So, as a result of the grooming, we will see probably a couple more quarters of negative corporate growth, stable and then beginning to grow.
On the enterprise segment, the majority of these locations are off-net. Most of these customers take global services from Cogent, and they take 2 primary products, Internet access and VPNs based on MPLS. Our decision to support MPLS going forward for at least a decade has actually allowed us to actually win a handful of brand-new enterprise accounts that today only have a handful of connections with Cogent but have indicated they will increase that. So we anticipate that the reduction in enterprise revenue will probably last into early 2026, or longer than corporate. And the main reason for that is, we are grooming some international operations and bifurcating the service to the customer, where we will continue to sell the port in a country where Cogent is licensed, but we are requiring the customer to enter into direct contract with the local loop provider as we are not licensed to resell those. Sprint had previously worked through a resale structure, where they were reselling and then effectively sub-reselling, and we were uncomfortable with that legal construct. So, as a result, as these off-net circuits in non-licensed countries -- to remind investors, we're licensed in 56 countries, but we provide service in over 170 countries around the world. In the differential, we require the customer to purchase their own loop. Cogent had always used that structure, and we are migrating the acquired Sprint enterprise customers to that. So probably about a year of continued enterprise decline, then flat to very modest growth probably. In that segment, we do not anticipate more than kind of a 1% annual top line growth.
And your next question comes from the line of Sebastiano Petti with JPMorgan.
Just in regards to the waves, Dave, can you remind us how should we think about perhaps additional investments needed to perhaps unlock the opportunity there, whether it be from a sales force investment perspective, maybe connection costs related to just enabling that from a CapEx perspective? And then, if we think about just the pace of installs, I think part of that is probably predicated on the backlog, but I think you said 500 or so monthly installs in waves is probably a near-term target, but perhaps accelerating from there. And so, how should we think about the cadence or the trend there? That would be helpful as well.
And then, I guess, one area of housekeeping related to waves. On the provisioning times, I think you're down to 30 days now. Can you remind us where was that last quarter? And what's the time line to get to the 2 weeks?
Yes. Fair enough. I'm going to take those in reverse order, Sebastiano. So last quarter, we were at -- so in Q3, we were probably at 120 days for the waves that we can provision. By Q4, for the ones that we provisioned, we got that down to 90. In this quarter, we are down to a 30-day provisioning window. And we are continuing to refine some of the processes and field deployment mechanisms and feel comfortable that we can get down to the same 2-week average that we have in IP services, which we've been able to do now for over 20 years. We tested all of these provisioning systems and configurations, but there are always corner cases in the real world that you learn from. But we feel very comfortable that the 30 days, which is far beyond what anyone else in the industry can do, is great, but we have the ability to effectively cut that in half over the next several months.
With regard to the pacing of installs, we are ramping up to our full install capabilities. And we have both the field resources and service delivery coordinators to support the 500 a month. The hard part now has been working through the backlog that we have and flushing out which orders are still available for install [ and in ] which the customers had to go somewhere else. And some of these orders have been in the funnel for over a year. And that is a time-consuming process. And we have groomed the funnel. It did shrink, even though there were some new orders coming into the funnel. And now, we expect the funnel to actually accelerate in new order intake as we have demonstrated our ability to install. I think we're still probably a month or 2 away from getting to kind of full cadence of 500 installs a month as a result of this grooming exercise because it does take an iterative back and forth with the customer.
Now, to the investment part of the wavelength question, and I'll answer that in 2 different areas, both capital in the network and sales force investment. I'll take the easiest one first, which is CapEx. We are substantially complete in our footprint. There have been additional data centers that have come online over the past 18 months since we announced that 800 target, and we are rapidly catching up. So we have ubiquitous wavelength capability across the footprint. In fact, the 808 that we ended the year with, 6 or 7 weeks later, is now close to 880. There are still probably about 60 or 70 data centers that have IP connectivity from Cogent in North America that don't have waves that have come online in the past 18 months that were rapidly adding wave capability. And on a going-forward basis, in new data center additions, they will automatically have both services. So it won't be a 2-part process.
In terms of the capital intensity of that, we think our kind of run rate stabilized CapEx will be in that roughly $100 million a year range. That is somewhat dependent on the rate at which new data centers come online. But the amount of incremental capital to bring wavelengths to the data center versus just IP is fairly trivial, probably less than $15,000 per incremental data center. So we don't see a material uptick in CapEx. In fact, we see our CapEx declining after midyear. And the elevated CapEx that we are spending now is being focused heavily on the data center conversion program, where we are trying to unlock 100 megawatts and 1 million square feet of the data center footprint to be marketed on a wholesale basis. We targeted 23 of the largest facilities to do that. Three have been fully converted. 20 are in the process of completing their conversion. Their conversion has been underway. It is capital intensive. And we anticipate, by midyear, by the end of June, all 23 of those facilities will be fully ready to be able to allow a counterparty to take control of them, whether they purchase them or lease them. We have been marketing these facilities in kind of collecting offers and letters of intent, as well as doing incremental tours. We believe that probably in the next 6 to 8 weeks, we will be far enough along that it will make sense for us to do some calls for offers and kind of flush out how many of the LOIs are in fact serious and are ready to transact and at what economics.
Now for the sales force piece, we have invested in our CRM to automate the provisioning and quoting of wavelengths. We actually automatically produce a route map with 1 meter accuracy with each wave order. That is something that no other provider can do. And it's in large part due to the fact that the Sprint network was based on extremely accurate GIS data since it was not a roll-up or a series of acquisitions, but it was organically built by one company. We also have a number of sales force training programs that have been ongoing and are now accelerating due to the ability to actually provision in a reasonable window in these additional sites. And the cost of that is not an extraordinary expense. It's just a redirection of our sales training resources. We have a total of 13 dedicated sales trainers. They train salespeople on all products. But over the past several months, their primary focus has been on wave training and certification. So we do not anticipate any extraordinary sales force investment.
And your next question comes from the line of Greg Williams with TD Cowen.
Maybe dovetailing off the wave answer there, Dave, you mentioned your competitive advantage is obviously getting to 2 weeks of provisioning, and then you just mentioned 1 meter of accuracy for your sales force. We just got back from a fiber conference and Lumen specifically calling out investing heavily in waves. Zayo is talking about 400 gig waves across the country. Just trying to understand the competitive concerns that they ramp up their wave product. And how much of your future growth in your $500 million target is share [ scaling ] versus the market expansion opportunity? Is there enough to go around for all of you?
Second question is just on the IP version 4 price hikes. It sounds like it's $0.44 per address. You mentioned you're going to hike your prices on about 10 million addresses. Where is that process and how is that going along? And what's the math there? If I think maybe $0.10 of price hike on 10 million addresses, that's like $12 million in revenue opportunity. Just trying to understand if I'm thinking of it correctly.
Yes. Thanks for the questions, Greg. And again, I'll take these in reverse order. On the IPv4 pricing, we initially raised the price of new sales, and we began cycling through the base that was not covered by contracts. We still do have some customers at lower rates that we will be increasing prices on over the next several months. In addition to that, we will evaluate whether we can further raise prices for new sales. We do have inventory of addresses that are unleased, but we also know that there is strong demand. So I do think we can continue to have outpaced growth in our IPv4 leasing. It has now become much more of a focus for Cogent. It represents almost 5% of our total revenues, and it will be the combination of incremental leasing activity, which you can see the cadence and progress of in our press release, as well as raising the average price of the base by both raising new incremental and going back to customers out of contract and raising that.
Now, to the wavelength question. We should never be naive. We should always believe we have competitors, and we have to win business on a wave-by-wave basis. The companies you mentioned have access to capital and are established players with substantial footprint. However, each of those companies has a much smaller number of data centers where they at least tell the market or advertise that they could deliver wavelength services, too. We were at 808 at year-end and 880 today. We'll be over 900 in the next few months, as we close out those remaining new data centers and bring them on to the wave network. That additional ubiquity is important.
The second key point is the architecture we deploy. It is radically different than that of Zayo and Lumen, and it allows us to provision any to any in this very rapid period with minimum number of field dispatches and accurate fiber inventory. The 2 companies you mentioned had acquired a number of companies. They operate disparate unintegrated networks. And for that reason, their typical wave install does take 90 to 120 days, typically requires 6 or more field visits and custom engineering in which they are unable even to produce an accurate route map to the level that we have. So we think we have unique right of way. We've got faster provisioning, more endpoints. But at the end of the day, we also have a much lower cost basis than they do in their assets. So we will be aggressive. I think we're going to grow our business by capturing new opportunities, but we will also displace current providers by offering a better solution. The switching costs are trivial in a data center because all you do is issue an LOA to the data center operator to migrate the cross-connect in that data center from the previous provider to the new provider. It's a very easy process.
And then finally, on the interface speeds that we can support, 100% of our sites can support 10 gig, 100 gig and 400 gig waves. Again, that's very different than our competitors. We can support all 3 speeds at 100% of the sites. The caveat I'm going to put on that is, in every vector, so every direction out of a site, we can support 10 gig and 100 gig with 100% coverage. And today, with 400 gig, we're at about 82% of the vectors that we can provide. So we do have a little bit of work to take some sites that are 400 gig enabled and make sure every direction out of that site can support 400 gig. That work should also be complete this quarter, in the first quarter of '25.
And your next question comes from the line of Walter Piecyk with LightShed.
Dave, Tad highlighted the sequential growth in IPv4 revenue. But I was looking at last quarter's revenue, it was higher. So, that shows that there was a sequential decline. Was there some type of a restatement in IPv4 lease revenue? And then, in terms of you saying focusing on IPv4 in terms of growth, again, the net adds, meaning new lines, IPv4 addresses that you're leasing, I think, only increased by like 90,000 in the quarter, and it used to increase by 500,000, 600,000, 700,000. Was that based on some restatement on Q3? Or is 90,000 kind of the run rate that you're delivering in terms of focusing on this area?
So the IPv4 revenue, the entity that has the secured notes is going to be audited. So we have begun that process in auditing the revenue streams. The data that has been disclosed have been based upon MRCs and totaling MRCs, as opposed to U.S. GAAP revenue per quarter. So you are correct in observing that we adjusted some of the amounts for prior periods to get to the U.S. GAAP numbers that will be audited coming up in the next few weeks here.
Right. So that's on the revenue side. And on the unit side, as we raise prices, there was some initial sticker shock. We did add just under 100,000 units in the quarter. But we also expect to be able to continue the average growth rate of nearly 0.5 million addresses per quarter over a multi-quarter period. If you go back and look at the data that we provided, there is volatility in every quarter in terms of the number of net new addresses added.
Okay. So it's going to bounce back to 500,000, even though we haven't seen that since Q2? And even though you just increased price and are contemplating new price increases, and this is just because you're focusing, I guess, on that business more. And then the revenue restatement was just because it wasn't audited before? So you actually were generating less revenue in IPv4?
No, the Disclosure was based upon adding up MRCs, as opposed to prorating the proper revenue amounts. If a customer installed, for example, on the 20th of the month, we were adding up MRCs, right?
So end-of-quarter MRCs, right?
That's right. So, you had to go...
It might have been helpful to actually call that out in the press release or the prepared remarks [ or in a side ]...
So, to be honest with you...
Rather than having to have it come up in Q&A. But can we move on to the wavelengths?
The total revenue...
When you talk about sequential growth, and it's actually down based on what you reported, yes, I think maybe it's not material in dollars, but whatever.
Wavelengths, I guess, a similar type of question. The backlog declining 20%, was this basically like you'd had orders from some time ago and then you checked in with these customers and they had gone elsewhere? And I'm just curious, like, if we look at what remains after this 20% decline in your wavelength orders sequentially, when is the last time you checked in with the customers that have existing orders in your backlog? And is there any chance that we'll see a similar reduction next quarter because they've gone elsewhere for their wavelengths?
Well, as we have stated on multiple previous calls, as we were building an order funnel and backlog, we knew that we could not provision those orders. Some of those orders were provisioned. So some of the decline came from the incremental orders that were provisioned. Some of the decline came from going to customers who had previously put in an order maybe 15 months ago and had gone somewhere else. Now that we have established a commitment to deliver today with a contractual penalty if we fail of 30 days in 808 sites at year-end and 880 sites today, we are cycling through that. There will be some breakage, but there are also new order intakes at an accelerating rate. I think the way you characterized it as a 20% decline in waves is not correct. What it is...
Well, [ 3,400 to 2,700 ], that's basic math, Dave. It's math.
It is a grooming of a funnel. The funnel reduced but not the waves. The number of waves installed actually grew.
Okay. Can you also just -- just lastly, my favorite topic is just the, I guess, aside from leverage, but I'll leave that for the next call, which is now north of 5, your corporate revenue, which was declining 2.7%. I think in the prepared comments, again, you talked about off-net and shifting off-net to on-net. But like what -- can you just kind of cut to the root of it what's going on in corporate? Is there any organic growth opportunity in that business? Or is this forever just in decline? I'm talking about the traditional Cogent corporate that really drove your dividend growth historically.
Well, our dividend growth has been driven by growth in cash flow and that growth in cash flow...
Not cash flow, debt leverage. Debt leverage, not cash flow. [ There's no ] cash flow. It's leverage.
I let you ask the question. I'll finish answering it. So I already answered the question for Mike Rollins earlier, which is that corporate growth will decline for another quarter or 2 as we continue to groom. As we groom both non-core and negative-margin off-net services that cannot be migrated to on-net, that results in a top line decline but an expansion in EBITDA. We think this is the appropriate strategy. The underlying on-net corporate footprint continues to grow.
And your next question comes from the line of Tim Horan with Oppenheimer.
Back on the wavelength; in your conversations, how important is the redundant route and new architecture and what type of price discount do you think you need to get to capture share? And can you give us an update on maybe how much revenue you can get in wavelengths now and what you're kind of still expecting in 5 years?
Yes. So thanks for the questions, Tim. First of all, I would actually say route redundancy is actually the most important factor to potential customers. If we rank their concerns; it's route redundancy, data center coverage, provisioning time and pricing; in that order. So price is probably not the most important factor, but it will help us close some business. In terms of redundancy, even for companies that have chosen to buy or build dark fiber, initially you would take that demand out of the market. But because the services on either dark fiber or waves are an unprotected service, they actually value the redundancy even more. So it actually generates incremental demand for the same city pair on a unique route.
We feel very comfortable about the target that we laid out at the closing of the deal in May of '23 and said within 5 years, we would be at a run rate of $500 million annually in wavelength sales. We think that the demand that we have received both in terms of order backlog and continued customer interest gives us a great deal of confidence we're going to make that number, particularly as we prove to the market our ability to actually deliver. The fact that there have been some customers in the funnel for over a year that are still taking waves is to me a little surprising, but some of them had enough confidence in us and we have delivered. For others, I think we have to build credibility and the way we're going to do that is install these waves. But we're very comfortable with our $500 million by mid-'28 target for wavelength revenue.
So can you just characterize how the sales effort's going right now? I mean you're basically there now on provisioning and will be, a couple of months from now you're there on coverage, all the things we're talking about. Have, I guess, the pitches ramped dramatically?
They have. So as we turned the calendar on 2025, we told customers in 808 sites we would do 30 days. The sales force was empowered to go out and reach into the funnel of people that had already signed and convey that message as well as reach out to new customers. And the cadence of getting new orders into the funnel has actually begun to accelerate. I think that cadence will accelerate even further as we end up with an ability to show that we can provision.
Lastly, Dave, can you just give us some -- I know you're reluctant to do this and maybe you can't. Some color on what we should expect for consolidated revenue and maybe consolidated EBITDA in total for this year. Obviously there's a lot of moving parts.
There are. And at the beginning of '24, we said that our EBITDA would be relatively equivalent to '23 and I think most sell-side analysts did not believe that. And then we ended the year at $348 million versus $352 million so about $4 million or 1% delta. I think going into the year, the kind of consensus was for about a 20% reduction in EBITDA due to the step down in payments from T-Mobile. We have demonstrated our ability to take out costs. So I think as we look at '25, we should deliver EBITDA effectively equivalent to '24 or '23 so around $350 million. You're right, we don't like to give specific annual guidance. We also kind of addressed the various customer segments and the declines and when we expect each of those segments to have flushed through the grooming of the noncore products. I would suspect that we will be delivering revenue that is slightly up for the full year due to wavelength sales.
And your next question comes from the line of Jim Schneider with Goldman Sachs.
Maybe I wanted to kind of get the corporate revenue question in a different way, if we could, Dave. Maybe just kind of think about it in terms of either corporate connections or on-net connections and core customer trends. I mean what's going to be required to get those sort of connections growing in a more positive direction over the next few quarters? And can you specifically talk to what's required in terms of sales force headcount and sales force productivity and how much productivity that needs to increase in order to get there?
Yes. So in terms of corporate connections, we need to groom those noncore products that are sold to on-net customers. So there is going to be some connection grooming that goes on and we are also migrating customers from off-net to on-net. We've done a significant amount of that, but there are other cases where there were off-net customers and on-net buildings that were trapped in a loop contract that we acquired from Sprint that we only migrate that customer once that contract has ended so we can recognize that savings. So I think on a connection basis, we've got a couple more quarters of headwinds and then positive. On a revenue base, it will be very similar as we think the ARPUs will be relatively flat. There is some price decline that is typically offset by migrations to larger connections.
To address the sales force piece of it, roughly 390 -- 387 of our salespeople focus on the corporate segment. That is the portion of the sales force that has the highest turnover. That is where we focus the majority of our normal training activities. As I mentioned earlier, we've diverted some of those to wave training to the NetCentric sales force, which is very stable, but we will be able to divert back training to the corporate team. And we do need the average number of connections sold per rep per month to go up. We're at 3.5 today. The long-term average is just under 5. We were at 4 in the previous quarter. There is some volatility around that, but I do believe that we have the right number of salespeople, the right turnover. I'd always like turnover to be lower. And I think we will continue to see corporate on-net growth.
Now the return to office environment is the backdrop we're working against. It has been tough. It's lasted much longer than we expected and we're only seeing positive net office absorption turning in a dozen or so markets this quarter, but that is at least a sign that we have reached bottom and we are starting up. So it's a combination of companies making a decision and companies occupying the space. But as I said to Mike's question, we feel comfortable that we're probably 1 to 2 quarters away from positive corporate revenue growth.
And maybe just a quick confirmation. I want to make sure that I understood what you said -- heard what you said before properly, which is the sustaining CapEx in the business should be $100 million annualized run rate all-in after all the noise is flushed out of the numbers.
That's correct. And we should be at a run rate of roughly $100 million by Q3 and Q4 of this year. We are at an elevated rate for the first half of the year and most of that incremental capital is being spent on the data center enablement program.
And your next question comes from the line of Christopher Schoell with UBS.
Maybe just 1 on the undermonetized assets. Dave, would you say there's a preference for leasing versus selling the data center and IPv4 assets at this stage? And as you think about maximizing value, do you believe this most likely will be achieved through a few large deals or more so through a series of transactions?
So on the IPv4s, at this point we are committed to leasing them. While they are liquid and could be sold since Microsoft and Amazon are today not acquiring at least in scale, we think that we would be leaving money on the table. The average price per address peaked at the end of '23 at about $60. Today, blocks are trading on exchanges between $48 and $52. So that is about a 15% reduction. If prices trade back up and that is I think likely as you look at the long-term appreciation, the first trades occurred in 2011 at $4 an address and today 14 years later, they're at about $50. I think the long-term trend is up, but we would consider being a seller of some of that address space.
In terms of the data centers, we really are agnostic to whether it is a cash sale or a lease. We're allowing the market to tell us both which data centers they want and which of the 2 models they would prefer to use. I would say if I looked at the pool of letters of intent today; they're about 60% lease, about 40% purchase. So while there's a slight preference for leasing, it's not overwhelming and that's out of after several hundred tours by different counterparties. There's almost 50 counterparties still actively working on these assets. We have been reluctant to accept any offers because the assets just are not ripe for monetization. That will happen over the next few months and that was to my point about calling for offers.
If we decide to take a lease deal because the overall economics are better for Cogent, we would probably look to securitize that to raise cash off of those leases. So therefore, the credit quality of the underlying tenant would be critical in our decision-making process. And then the final part of your question, is it going to be a holistic deal or one-off? While we have several offers for the entire portfolio, they do not maximize value on specific assets within the portfolio. Again until we do a call for offers and can line everything up, I'm just not in a position to answer that. I want to be absolutely clear with investors. This is something we haven't done before. We've got strong indications that our pricing is reasonable and there's strong demand, but we really want the market to answer the question for us.
That's helpful, Dave. And then maybe just 1 more. You pointed to stable EBITDA in '25. Can you just remind us what you're contemplating with the step down in payments and your comfort level with the dividend policy as you await the uplift from growth initiatives?
Yes. We're very comfortable with our dividend policy. As we mentioned on the previous call, our LQA EBITDA troughed in Q3. On an LTM basis, that is going to trough in Q3 of '25. We are 5.07 today. We will get into the high 5s, 5.7, 5.8 in part due to the expenditures that we're making on the data centers and that drain on capital. We are comfortable with that because we also know that with the growth in EBITDA, we see a rapid delevering occurring starting in Q4 of this year and probably by sometime in mid-'27, we will be actually below the midpoint of our target.
And your next question comes from the line of Nick Del Deo with MoffettNathanson.
Just 2 quick ones given the time. I think you said traffic growth was 11% year-over-year, which I think is the lowest you've ever had maybe aside from the quarter when you lapped Megaupload back in 2012. I guess what's going on there and what does it mean for the NetCentric outlook? And then on the data center front, obviously we're mostly focused on the facilities you're trying to spruce up for wholesale or for sale. But I guess as you think about your retail assets, I think they've historically had utilization rates in the 30% range or so. I guess should we expect that to go up in light of the broader supply-demand dynamic in the data center space?
Yes. So first of all on the traffic growth number, traffic was essentially sequentially flat. On a quarterly basis, it was up 11% year-over-year. But for the full year, it was up 16% so calendar year '24 versus full calendar year '23. In terms of what's going on, we are reaching a level of maturity in streaming in the developed world with roughly 53%, 54% of all video now being streamed and that growth rate in streaming versus the hockey stick we saw post pandemic is moderating. I do think that we are going to see a new leg up in growth as AI matures and as inference models become more ubiquitous and more commonly used. I think that will drive a new leg of growth. And if we go back and look at Cogent's reported traffic growth over a 20-year history as a public company, there have been ups and downs.
Yes, Megaupload was a particularly abrupt [ shot ], but we've also seen other slowdowns when YouTube's casual video business started to mature, that slowed down our traffic growth. When Napster's file sharing business matured. Now these were off of much smaller bases. I think for our NetCentric IP business, I think we will continue to see growth at similar historic rates. That's kind of in the 9%, 10% rate. And there are ebbs and flows in quarterly traffic. I know that several of our CDN customers did have a pretty tough quarter in terms of their traffic volumes and they are heavily dependent on us. But I don't think the Internet stopped growing. I think in aggregate, we're going to grow as fast or faster than the Internet, gaining share and it's got a long way to go before it's matured.
To your data center question, you are correct in the premise that our primary focus today is getting that excess space marketable so we can monetize it. But we did add a lot of these smaller facilities, which are more focused on our retail base. The 55 edge data centers that we added average about 40 racks of capacity and about 350 kilowatts so 1/3 of a megawatt in existing power. With this added footprint, our utilization rate for the retail side of our business has declined fairly materially. We brought a lot of supply on and haven't sold it. I think we will be in a position to increase occupancy in the retail portion. We've been a little reluctant to put retail into some of the data centers that will also have a significant amount of wholesale and the reason for that is someone may want to take the whole facility and we may exit our retail footprint.
Our retail footprint in those facilities typically is about 10,000 square feet and about 1 megawatt. And it didn't make sense to put someone in that 10,000 feet that was taking 2 or 3 cabinets and that could be an impediment 2 or 3 months from now when someone comes on and says I want to take the whole facility. But I think in general, we are probably a couple of years away from getting back to 30% occupancy. Now hopefully, the inference phase and the geographic ubiquity of our footprint changes that and accelerates that. But today, we don't have a lot of data around AI edge cases as a significant driver of colocation. Basically, our retail colo business has been aimed mostly at small corporates and I think that's at least where we're focused today. That may change as some of the AI inference models evolve.
And your final question comes from the line of Brandon Nispel with KeyBanc Capital Markets.
Just housekeeping questions I suppose given time. Dave, how many waves specifically did you groom out of the funnel?
We groomed about 1,500 out of the funnel.
Okay. 1,500. Were there any onetime network expenses? I think last quarter you called out some data center costs?
So the answer is yes. We are spending elevated capital on the data centers and there was about $5 million of COGS associated with some of the clearing of those facilities that is not capitalized. I mentioned that in the Q3 earnings call, we spent that in the third quarter and we said we would spend that also in the fourth quarter, but that demolition component of spending would go away and that is a noncapitalizable expense. So yes, there was some direct costs for that effort.
Perfect. And then lastly, is there any way you can quantify just with 1 number on enterprise, corporate and NetCentric; what the grooming activities related to so purposeful disconnects and/or provide a churn rate that includes those disconnects and a churn rate excluding sort of the purposeful disconnects?
It is hard for me to give you a single number because there are really 3 things going on at the same time -- well, 4 things. You're moving some people from off-net to on-net, you're keeping the customer. But there are other cases where you're turning off a low capacity corporate circuit that there's no fiber alternative for in North America. You're turning off enterprise circuits in international footprints where we are not licensed. And then finally, for both of those segments, corporate and enterprise, we're end of [ lifing ] these products and we had about $1 million headwind sequentially this quarter, $900,000, and we'll continue to turn that off. So there's not a single number. You really need to look at each of the subcomponents to kind of know what's going on. A kind of 1 headline number would be misleading.
Thanks, everyone. I want to appreciate everyone listening. We did cut our prepared remarks. We got done in less than 1.5 hours. And thanks, everyone. Take care and we'll talk soon. Bye bye.
This concludes today's conference call. Thank you all for joining. You may now disconnect.