Cogent (CCOI) Q3 2026 Earnings Call Transcript

DATE

Nov. 5, 2026

CALL PARTICIPANTS

• Chairman and Chief Executive Officer - Dave Schaeffer
• Chief Financial Officer - Tad Weed

TAKEAWAYS

• Service Revenue -- $231.2 million, representing a 1.8% sequential decrease and a 4.2% constant currency decline year over year.
• Adjusted EBITDA -- $73.9 million, growing 3.9% from $71.1 million in the previous quarter.
• Adjusted EBITDA Margin -- 32.0%, increasing from 30.2% in the second quarter.
• Wavelength Revenue -- $16.0 million, representing 57% year-over-year growth and 7.8% growth from the previous quarter.
• Wavelength Customer Connections -- 2,722, adding a net of 277 connections in the quarter as compared to 182 in the previous quarter.
• On-Net Revenue -- $153.9 million, increasing 2.4% sequentially and comprising 67% of total revenue.
• Off-Net Revenue -- $76.6 million, decreasing 9.3% from the second quarter and 19.5% year over year.
• IPv4 Leasing Revenue -- $18.4 million, increasing 1.8% sequentially and 5.4% year over year.
• Total Headcount -- 1,603 employees, a reduction of 79 from the previous quarter and 279 individuals since the third quarter of 2025.
• Capital Expenditures -- $36.3 million, declining 5.8% from $38.5 million in the second quarter.
• Finance Lease Payments -- $7.1 million, a sequential decrease of 26.6% from the second quarter.
• Total Net Debt -- $1.8 billion, which includes $137.5 million of discounted value for amounts due from T-Mobile.
• Gross Leverage -- 6.94 times, reduced from 7.5 times in the previous quarter as adjusted for T-Mobile receivables.
• 2032 Secured Note Repurchases -- $118.4 million par value repurchased at an average price of 90.07 during the third quarter.
• 2027 Unsecured Note Repurchases -- $50 million par value repurchased at an average price of 99.25 in the third quarter.
• Total Customer Connections -- 114,678, representing a 1% decline from the previous quarter.
• IP Network Traffic -- Increased 4% sequentially and 14% year over year.
• Sales Rep Productivity -- 4.6 units per full-time equivalent, an increase from 4.5 in the previous quarter.
• On-Net Buildings -- 3,665, an increase of 38 buildings from the end of the previous quarter.
• Wavelength Enabled Locations -- 1,151 data centers, compared to 1,137 at the end of the previous quarter.
• IPv4 Address Inventory -- 15.3 million addresses leased out of a total title to 37.8 million addresses.
• New Contract Pricing -- $0.05 per megabit for new customer contracts, compared to $0.06 in the previous quarter.
• NetCentric Revenue -- $107.2 million, a year-over-year increase of 6.9%.
• Corporate Revenue -- $97.1 million, representing a 1.6% sequential decrease.
• Enterprise Revenue -- $26.9 million, a decline of 8.7% from the previous quarter and 26.1% year over year.

Need a quote from a Motley Fool analyst? Email pr@fool.com

RISKS

• Schaeffer stated, "we continue to experience certain equipment price increases from vendors due to supply chain constraints," which served as a headwind to capital expenditures during the quarter.
• Schaeffer indicated that customer constraints in power availability and equipment deliveries are delaying service activations for wavelength products.
• Schaeffer noted, "sales rep turnover was elevated in the quarter at 7.7% per month, which is above our historical average of 5.7 reps per month."

SUMMARY

Cogent Communications Holdings, Inc. (NASDAQ:CCOI) reported a third quarter focused on shifting its revenue mix toward higher-margin on-net services as it integrates former Sprint wireline assets. Management announced the suspension of the dividend program to prioritize the refinancing of $700 million in unsecured notes maturing in June 2027. The company is advancing its data center monetization strategy, with 14 of 24 targeted facilities now sold or under nonbinding letters of intent. Operational strategies include headcount reductions and the expansion of wavelength services, which management expects will eventually capture 25% of the North American long-haul market.

• Management reported that 81% of third quarter sales were for on-net services. CEO Schaeffer stated, "the improvement in our margins and the growth in our EBITDA during the third quarter was primarily driven by our focus of selling our more profitable on-net services."
• The company engaged Evercore and Latham & Watkins to address the 2027 note maturity. Schaeffer stated, "we have received a number of proposals, and we are in the process of evaluating each of these offers with the assistance of Evercore and Latham & Watkins."
• Cogent has entered into two nonbinding letters of intent to sell four data centers for approximately $100 million. Schaeffer noted, "we anticipate to be able to close these transactions early next year."
• Wavelength market share reached approximately 4% during the quarter. Schaeffer indicated, "we continue to expect to capture 25% of the North American long-haul wavelength market."
• Management is upgrading select data centers to support high power density requirements. Schaeffer stated, "some of our Cogent data centers have been further upgraded to accept high power density cabinets capable of up to 20 kW per cabinet."
• The Cogent Classic revenue base grew to $201 million, while acquired Sprint revenue continued to decline. Schaeffer noted, "the Cogent Classic revenue base has increased by 29%, from $155 million to $201 million" since the closing of the Sprint acquisition.

INDUSTRY GLOSSARY

• On-Net: Services provided to customers in buildings physically connected to the Cogent network.
• Off-Net: Services provided to customers in buildings connected to Cogent's network using third-party carrier circuits for the last mile.
• Wavelength: A dedicated, high-capacity fiber-optic connection using specific colors of light to transmit data across a network.
• NetCentric: Customers who are high-volume users of bandwidth, such as internet service providers, streaming services, and content delivery networks.
• IPv4: The fourth version of the Internet Protocol, used to identify devices on a network through a 32-bit addressing scheme.
• ARPU: Average Revenue Per User, measuring the average monthly revenue generated by each customer connection.
• DSO: Days Sales Outstanding, representing the average time taken to collect payments from customers after a sale.
• Cogent Classic: The legacy core business and network infrastructure of Cogent prior to its acquisition of the Sprint wireline business.
• IP Transit Services Agreement: A contract with T-Mobile USA under which Cogent provides transit services in exchange for $700 million in total payments over a multiyear period.

Full Conference Call Transcript

Operator: Good morning, and welcome to the Cogent Communications Holding Third Quarter 2026 Earnings Call. As a reminder, this call is being recorded and will be available for replay at www.cogentco.com. A transcript of this conference call will be posted on 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 website. After today's prepared remarks, we will host a question-and-answer session. If you would like to ask a question, please press star one to raise your hand. To withdraw your question, press star one again. I would now like to hand the call over to Mr.

Dave Schaeffer, Chairman and Chief Executive Officer at Cogent Communications Holding. Please go ahead.

Dave Schaeffer: Hey, thanks, and good morning to everyone. Welcome to our Third Quarter 2026 Earnings Conference Call. I'm Dave Schaeffer, Cogent's Chief Executive Officer, and with me on today's call is Tad Weed, our Chief Financial Officer. I'd like to touch on a few key events and significant matters in the quarter. During the quarter, third quarter in 2026, we continue to make progress in advancing the Cogent Classic business and making the company operate more efficiently. I want to begin by recognizing some significant accomplishments we achieved, as well as provide an update on several important matters. First, ongoing rotation continues into our more profitable on-net services and away from less profitable off-net services.

Cost reduction, margin expansion, and completion of many integration projects are continuing. Third, our wavelength business continues to make progress and continues to grow. A reduction in our capital expenditures and capital lease payments. Continued progress on our data center monetization. We continue to reduce leverage and our debt, and we are addressing our 2027 unsecured notes. First, the rotation to on-net products. The improvement in our margins and the growth in our EBITDA during the third quarter was primarily driven by our focus of selling our more profitable on-net services. We continue to serve on-net customers in 3,665 on-net buildings.

Our total on-net revenues, including on-net wavelength revenues, were $153.9 million in the quarter, a year-over-year increase of 5.8% and a sequential increase of 2.4%. 81% of our sales in the third quarter of 2026 were for our on-net services, inclusive of wavelengths, and 67% of our Q3 2026 revenues, including wavelengths, were on-net services as compared to 64% last quarter. Now for a word on our cost reduction and additional margin expansion. Diligent cost reductions and the completion of many of our integration projects have contributed to increasing our gross margin and EBITDA margins this quarter. We continue to seek out efficiencies and reduce our costs.

Our cost of network services was reduced by 3.1% sequentially and 7.7% on a year-over-year basis. As a result, our gross margin percentage increased on a year-over-year basis by 180 basis points and sequentially increased by 60 basis points to 47.6%. Our SG&A expenses were reduced by 5.1% sequentially and reduced 1.3% on a year-over-year basis. As a result, our adjusted EBITDA increased by $2.8 million sequentially to $73.9 million, and our adjusted EBITDA margins increased sequentially by 150 basis points to 32%. We continue to evaluate the optimal size of all of our departments as the integration of the former Sprint assets is now being completed.

We reduced our total headcount in the quarter to 1,603 employees at quarter's end, a reduction of 79 from the headcount at the end of last quarter and a reduction of 279 individuals from Q3 of 2025. This reduction represents approximately 5% of our workforce from the previous quarter. The expenses associated with these reductions were recognized in the third quarter of 2026. Wavelength sales. Our wavelength business continues to grow. At quarter end, we offered wavelength services at 1,151 locations, with the ability to deliver 10-gigabit, 100-gigabit, and 400-gigabit services. Our provisioning intervals are approximately 30 days. Our wavelength revenues in the quarter were $16 million.

Our wavelength revenue increased by 57% on a year-over-year basis and increased sequentially by 7.8%. Our wavelength customers increased on a year-over-year basis by 55.5% and increased sequentially by 11.3% to 2,722 wavelength customer connections. We added a net of 277 wavelength connections in the quarter, as compared to 182 in the previous quarter and 199 in the first quarter of the year. We are encouraged by this acceleration. At the end of the quarter, we have sold waves in 660 unique locations to 578 unique customers. We continue to expect to capture 25% of the North American long-haul wavelength market. Today, our market share has increased to approximately 4%. For a word on capital expenditures.

Our CapEx declined by 5.8% sequentially to $36.3 million in the quarter. As somewhat of a headwind to CapEx, we continue to experience certain equipment price increases from vendors due to supply chain constraints. Also, principal payments on our capital leases declined sequentially by 26.6% and were $7.1 million in the quarter. Our combined CapEx and capital lease principal payments declined sequentially by 10% to $43.4 million from $48.2 million the previous quarter. Now for a word on our continued data center monetization. We continue to work to monetize the 24 former Sprint central offices that were converted to data centers that we have earmarked for outright sale or leasing on a wholesale basis.

We have sold and have agreements to sell a combined 14 of these facilities out of the 24 that are targeted for sale. Last quarter, we closed the sale of 10 former Sprint facilities for $225 million in cash. We have entered into two non-binding letters of intent to sell four more of these former Sprint facilities that have been converted into data centers for a combined cash price of approximately $100 million. We anticipate to be able to close these transactions early next year. We continue to have multiple parties interested in many of our former Sprint facilities that we have converted into data centers. Now for a word on our leverage and debt reduction.

We have repurchased a total of $138.8 million at par of our 2032 notes at an average price of 90.348, resulting in a cash cost of $125.4 million and a gain of approximately $12 million. These purchases have reduced our annual interest expense on our 2032 notes from $39 million annually to $30 million, and reduced our gross leverage. We also purchased $50 million of our 2027 notes for an average price of 99.25. We have reduced our capital lease balances from $630.2 million at the end of last quarter to $622.1 million this quarter. As a result, our gross leverage, as adjusted for the amounts due from T-Mobile, was reduced from 7.5 times last quarter to 6.94 times.

These transactions reduced our total gross debt by $197 million in the third quarter. Now to address the maturity of our $700 million remaining in our 2027 unsecured notes. During the quarter, we reduced the principal amount outstanding from $750 million to $700 million. This transaction, combined with the $138.8 million of our 2032 notes, satisfied the requirement for $168 million of these debt repurchases using the funds from the proceeds of the 10 data centers as required in our supplemental indenture to our 2032 notes. Our $700 million unsecured notes mature on June 15, 2027.

We have engaged Evercore and Latham & Watkins to assist us in addressing these obligations under the terms that are most beneficial to Cogent and its stakeholders. We have received a number of proposals, and we are in the process of evaluating each of these offers with the assistance of Evercore and Latham & Watkins. We continue to solicit additional proposals and will evaluate all options available to Cogent. While we are going through this process, our board felt it was prudent to suspend our dividend program, which was modest, at least until we have settled on a path forward. Now for a word on expectations against our guidance.

Remember, our guidance is designed to be multi-year and not specific to a given quarter. While our growth rate for the third quarter of 2026 was negative, we believe that the decline in revenues from the acquired Sprint customer base continues to moderate. We anticipate long-term growth in the 6%-8% range, with EBITDA margins expanding at about 200 basis points per year. Our revenue and EBITDA guidance, again, are multi-year and not meant to be specific quarterly or annual guidance. Now I'd like to turn it over to Tad to read our safe harbor language, provide some additional details on our operating performance in the quarter. I will then summarize, and we will open the call for questions and answers.

Tad Weed: Thank you, Dave. 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 the call, you will find these reconciled to the corresponding GAAP measurement in our earnings releases that are posted on our website at www.cogentco.com.

Some comments on the revenue results. Despite our revenue decreases on the top line in the aggregate, we have been able to increase our gross margin and EBITDA margin. The increases in our gross margin and our EBITDA margin have been driven by cost reduction and a rotation to our more profitable on-net products. Comparing our revenue by connection type from the third quarter of 2023, the first full quarter when we were combined with Sprint Wireline, to this quarter illustrates the material changes to the composition of our revenues and the strength of our underlying Cogent Classic business. Our on-net revenues were 47% of our total revenues in the third quarter of 2023.

Our total on-net revenues, including wavelength revenues, increased to 66.6% this quarter. That is compared to 63.8% last quarter and 60.1% in the third quarter of 2025. Our off-net revenues were 48% of our total revenues in the third quarter of 2023. Off-net revenues have decreased to 33.1% this quarter, and that is compared to 35.9% last quarter and 39.3% of our total revenues in the third quarter of 2025. Non-core revenues were 5% of our total revenues in the third quarter of 2023, and they have now decreased to less than $1 million and about 0.3% of our total revenues this quarter. Our total revenue for the quarter was $231.2 million.

That was a decline sequentially of $4.4 million, or 1.8%. USF tax revenues had a negligible impact on sequential revenue results and a negative year-to-year impact of $0.3 million. FX had a negative impact of $0.4 million on sequential revenue and negative $0.5 million on year-over-year revenue results. We analyze and classify our revenues into four network connection types and three customer types. Our four network connection types are on-net, off-net, wavelength, and non-core. Our three customer types are NetCentric customers, corporate customers, and enterprise customers. For the quarter, sequentially, our on-net revenues, including wave revenues, which are on-net predominantly, increased by $3.7 million. Our less profitable off-net revenues declined by $7.9 million.

Our non-core revenues decreased by $0.1 million, and our wavelength revenues, which again is almost entirely on-net, increased by $1.2 million. On IP lease revenue, our IPv4 leasing revenue increased sequentially by 1.8% to $18.4 million, and by 5.4% year over year. Our lease price is very stable at $0.40 per month per address. We have title to 37.8 million IPv4 addresses, and we have leased about 15.3 million of those addresses. The substantial changes in the acquired Sprint Wireline revenue base have masked the underlying performance of our Cogent Classic business. Since we closed the Sprint acquisition from T-Mobile, the Cogent Classic revenue base has increased by 29%, from $155 million to $201 million.

On the other hand, the acquired Sprint Wireline corporate and enterprise and non-core and off-net revenues declined to $29 million for this quarter. Corporate business. Our total corporate business represented 42% of our revenues this quarter, and that was a decrease of 7.7% year over year and sequentially by 1.6%. NetCentric. Our total NetCentric business continues to increase and benefit from the growth in video traffic, activity related to artificial intelligence, streaming, IPv4 leasing, and wavelength sales. Our NetCentric business was 46.3% of our revenues this quarter. That was a 6.9% year-over-year increase. Sequentially, it was flat. Enterprise business. Our total enterprise business represents 11.7% of our revenues for the quarter.

Our quarterly enterprise revenue decreased by 26.1% year over year and sequentially by 8.7%, primarily due to a reduction in the acquired Sprint Wireline enterprise off-net revenues. Revenue and customer connections by network type. We serve our on-net customers in 3,665 total on-net buildings at the end of the quarter. Our total on-net revenue, including wavelength, was $153.9 million for the quarter, a year-over-year increase of 5.8% and a sequential increase of 2.4%. Our off-net revenue was $76.6 million for the quarter. That was a year-over-year decrease of 19.5% and a sequential decrease of 9.3%. Our off-net revenue results are impacted by the continued grooming and termination of low-margin off-net contracts. Some comments on pricing.

Our average price per megabit for our installed base decreased slightly sequentially to $0.10 from $0.11 last quarter, and from $0.16 for the third quarter of last year. Our average price per megabit for our new customer contracts for the quarter was $0.05, compared to $0.06 last quarter and $0.07 from the third quarter of last year. Our ARPUs. Our ARPU for the quarter were as follows: Our on-net ARPU was $524. Our off-net IP ARPU was $1,123. Our wavelength ARPU was $2,062. Our IPv4 ARPU was $0.40 per address. Churn. Our on-net unit monthly churn rate was 1.1%, an improvement from 1.3% last quarter.

Our off-net churn is primarily driven by the reduction in our acquired Sprint customer base, and our off-net unit churn monthly rate was 2%, an improvement from 2.3% last quarter. Our wavelength monthly churn rate has been very stable and is about 0.5%. Traffic. Our IP network traffic growth continued for the quarter. Our IP network traffic for the quarter increased sequentially by 4% and year over year by 14%. Sales rep productivity. Our sales rep productivity improved sequentially and was 4.6 for the quarter. That was up from 4.5 last quarter. Our long-term average rep productivity is 4.8. Foreign currency.

Our revenue earned outside of the U.S. continues to be about 21% of our revenues, and it was the same this quarter. Based upon the average euro and Canadian conversion to USD rates so far this quarter, so for the fourth quarter, we estimate that the FX conversion impact on our sequentially revenues for the fourth quarter will be negative $0.8 million, and year over year, the quarterly revenues negative impact will be $1.2 million. Customer concentration. Our revenues and customer base is not highly concentrated. Our top 25 customers represent about 16% of our revenues. Debt and debt ratios. Our total gross debt at par was $2.2 billion at quarter end.

This includes our reduced $62.1 million of finance IRU lease obligations, down from $630.2 million last quarter, and our reduced principal amount on our 2032 secured notes, reduced to $461.2 million, and our reduced principal amount of our 2027 notes of $700 million. Our net debt, total debt net of our cash, and our $137.5 million of the discounted value of the amounts due from T-Mobile was $1.8 billion. Our leverage ratio, as calculated under our more restrictive unsecured $700 million 2027 notes indenture- was 5.35, and our secured leverage ratio was 3.25. Our fixed coverage ratio was 2.32.

The definition of a consolidated cash flow under our 2032 notes indenture includes cash payments under our IP Transit Services Agreement with T-Mobile in the determination of consolidated cash flow. As a result, our leverage ratio was calculated under the 2032 notes indenture was 4.11, secured leverage was 2.5, and fixed coverage was 3.02. Cash and restricted cash. $168 million of the net proceeds from the sale of the 10 data centers was considered restricted cash at closing, since that amount was reserved for purchases of our debt obligations at a discount. We have purchased $138.8 million par value of our 2032 notes and $50 million par value of our 2027 notes for total par value purchases of $188.8 million.

As a result, there was no restricted cash related to the data center proceeds at quarter end, and we were also able to reduce our restricted cash by $20 million from the results of our leasing revenue under our IPv4 leasing. Bad debt and DSO. Our DSO was 30 days at quarter end compared to 29 days last quarter, and bad debt expense was 1% of our revenues for the quarter. With that, I will turn the call back over to Dave.

Dave Schaeffer: Hey, thanks, Tad. I'd like to highlight a couple of strengths of our network, our customer base, and our sales force. We remain direct beneficiaries of key trends in our industry, whether it be streaming video or artificial intelligence. At quarter end, we are able to sell wavelength services in 1,151 data centers across North America with a provisioning window of 30 days or less. We are selling waves to 578 unique customers as of quarter end in 660 of these locations. At quarter end, we are selling IP services in 1,992 data centers globally in 58 countries and 309 markets around the world. At quarter's end, our IP network directly connected to 7,531 networks.

22 of these networks are settlement-free peers, and 7,509 of these networks are Cogent paying transit customers. We remain continually focused on improving our sales force productivity and efficacy and managing out underperforming sales reps. Our sales rep turnover was elevated in the quarter at 7.7% per month, which is above our historical average of 5.7 reps per month, but it was lower than our all-time high point of 8.7. At quarter's end, we had 471 quota-bearing sales reps. Our sales force included 244 sales professionals focused on the NetCentric market, 216 professionals focused on the corporate market, and 11 sales professionals focused on the enterprise market. In summary, we've made progress in a number of key areas.

We are improving our margins, growing our EBITDA due to diligence and cost reductions, the completion of many of our integration projects, and the desire to sell on-net services over off-net. 81% of our sales in the quarter, inclusive of wavelengths, were on-net. 67% of our revenues in the third quarter, including wavelengths, were on-net as compared to 64% the previous quarter. We are actively working to further monetize the remaining acquired Sprint facilities that we have converted to data centers and to accelerate the growth in our EBITDA, which will allow us to further delever. We are very disciplined around our capital allocation decisions. We remain focused and optimistic about our wavelength business.

Our wavelength services are differentiated by quality, breadth of footprint, uniqueness of routes, reliability, and efficiency in provisioning. Our on-net services truly offer unparalleled value to our customers. at quarter's end, we were providing IP services in 1,820 carrier-neutral data centers and 172 Cogent-owned data centers. This footprint represents a footprint of approximately 18 GW of installed power. Our Cogent data centers have a total of approximately 150 MW of power available and approximately 1.5 million square feet of floor space. Some of our Cogent data centers have been further upgraded to accept high power density cabinets capable of up to 20 kW per cabinet above the standard 6 kW cabinets that we have historically offered.

The proceeds from our data center sales have materially improved our liquidity and enabled us to reduce our leverage and continue to allow us to invest in our network, and we are continuing to market additional facilities for sales. We are working diligently to address the 2027 note maturity. Most importantly, Cogent offers superior products, unparalleled quality, a broad footprint in traffic-rich locations, expedited provisioning, and market disruptive pricing. In summary, we win on value. With that, I would like to open the floor for questions.

Operator: We will now begin the question and answer session. Please limit yourself to one question and one follow-up. If you would like to ask a question, please press star one to raise your hand. To withdraw your question, press star one again. We ask that you pick up your handset when asking a question to allow for optimal sound quality. If you are muted locally, please remember to unmute your device. Please stand by while we compile the Q&A roster. Your first question is from the line of Greg Williams from TD Cowen. Your line is now open. Please go ahead.

Greg Williams: Great. Thanks for taking my questions. First one is on the refi process. It sounds like you are still in the phase of soliciting offers, so a bit of a back and forth here. What is your desired pursuit here and outcome, whether it is exchanges or reducing principal, and where the impasse is, generally speaking? Bottom line is, what is taking so long, and what are your updated rate expectations and timing for deals here? Second question just on EBITDA and OPEX. EBITDA came in a little light here. Just wondering where are the savings that you have messaged before?

We would have thought the data center removals from the ones you sold to I Squared would have come through a little bit, the transition cost removals and the Sprint cost reductions. I understand there is a little bit of severance in the quarter that might have eaten up on some of the EBITDA. How much were the severance costs this quarter? Again, what are the puts and takes on EBITDA as you think about the fourth quarter setup? Thanks.

Dave Schaeffer: Yeah, sure. Thanks for both questions, Greg. First of all, on the refinancing process, because the company has received multiple proposals for its refinancing and addressing the maturities, we have engaged both Evercore and Latham & Watkins to help us evaluate the pros and cons of each of these different proposals. We are continuing to receive both revised proposals from parties who have already submitted proposals, as well as additional proposals. Because there are many different permutations, the evaluation process is more deliberate. I would suspect that we will get these done in short order, but we are not in a position to give an exact timeframe.

Now, with regard to the outcome, we are very focused on the total cash cost of our debt service. That can be addressed either through interest rates, additional collateral, converting unsecureds to secureds, or through principal reduction, or a combination of all of the above. Because there are many variables here, the analysis is more complex, and that's why we have the additional resources helping us. Now, with regard to your second question around EBITDA progression. Our EBITDA grew $2.8 million on a sequential basis, while our top line revenue continued to decline.

We have seen substantial EBITDA margin expansion, and we did recognize some of the savings associated both with the data center sales, and therefore the alleviation of some of that carry cost, along with some of the savings achieved through the reductions that we had implemented in the previous quarter. We again reduced total headcount by 5% in the quarter and did bear the costs associated with those further headcount reductions.

We generally don't want to break out our severance policies because they are different across different segments of the organization, but I feel comfortable that the savings that we described in our last earnings call, which would total approximately $50 million on an annualized basis, are starting to show up and will continue to flow through. The final point is the product rotation. We continue to focus on higher-margin services, with 81% of all sales in the quarter being on-net, carrying much higher contribution margins, and virtually all of the revenue decline being in the off-net products, which have much lower margin contribution. We have materially improved the mix. The quarter immediately prior to acquiring Sprint, Cogent was 76% on-net.

The quarter immediately after that deal closed, the combined company was only 47% on-net. We have improved that on-net percentage from 47% back up to 66%. We are still slightly below where we were pre-acquisition, but as we continue to focus selling on-net, we expect those on-net sales to flow through and help us grow our EBITDA, because we are very conscious of the fact that in February of 2028, the subsidy stream from T-Mobile will be complete.

Greg Williams: Great. Thank you.

Dave Schaeffer: Thanks, Greg.

Operator: Your next question from the line of Chris Schoell from UBS. Your line is now open. Please go ahead.

Chris Schoell: Great. Thank you. Dave, you have stated that you remain confident in the ultimate opportunity for waves, but scaling the business has taken longer. At this stage, can you just walk us through the constraints you are experiencing, whether it be customer readiness, sales execution, or competition, and what gives you comfort that these headwinds can start to abate as we move into 2027? Thank you.

Dave Schaeffer: Yeah, sure, Chris. First of all, we saw a material acceleration in the number of waves installed on a sequential basis. We went from basically 199 to 277, or about a 40% increase in the number of waves we installed. We continue to build a backlog. We see strong demand, but many of our customers are facing constraints around their ability to activate their services, whether it be data center readiness, whether it be power availability, or equipment deliveries. In each of these areas, there have been material supply chain constraints.

While there are nearly $7 trillion of announced AI investment, we also know that most of that investment is yet to be completed, with only about $1.5 trillion of that $7 trillion actually being completed. And it is these limitations that are delaying customers' acceptance. We have been fortunate that we have been able to avoid supply chain constraints in our network equipment, but this is an industry-wide issue. I feel the demand set remains strong, and our competitive differentiation allows us to capture share. I think that will continue going forward. Our progression from 3% of the market to 4% is slower than we would like, but still a positive sign.

And the extremely low churn rate that we are experiencing on waves, now having a couple of years of wave experience, is, I think, also validating the quality of our network. So, whether it is quality, price, location, speed to deliver, we win in all of these dimensions. And as customers are able to overcome their supply chain limitations, I think we will see some further acceleration in our wave business. It was a business that grew sequentially by nearly 8%, 7.8%, and grew by 57% on a year-over-year basis. While we would like it to grow even faster, those are the fastest growth rates of any of the products in Cogent's portfolio.

Chris Schoell: Thanks, Dave. If I can just follow up on the wavelength.

Dave Schaeffer: Yeah

Chris Schoell: ARPU has been relatively steady in recent quarters. How is the mix of 10, 100, and 400 gig evolving, and when can we start to see ARPU uplift from higher capacity circuits?

Dave Schaeffer: Last quarter, we actually did get a little bit of an uplift from that migration. This quarter, there was not a large number of existing waves that transitioned from 100 to 400, but we continue to see the percentage of new orders that are for 400 waves increase, and we do expect that the ARPUs will gradually increase. The ARPUs are governed by three inputs, the speed, the actual path length, and the contract term. We also have seen a slight shift towards longer-term contracts, which are offsetting the speed increases.

We finally are seeing a significant reduction in requests for quotes for 10-gig services, as we think many customers are sunsetting that product and moving to 100, and many 100s are moving to 400. I think this is a multi-year trend. There's talk of 800-gig interfaces and conversations with equipment vendors. I think commercially available equipment for 800 will probably be sometime in late 2028 or early 2029. The infrastructure that we built is capable of supporting not only 800-gig interfaces, but 1.6 terabit when they become available as well.

Chris Schoell: Great. Thanks, Dave.

Dave Schaeffer: Hey, thanks, Chris.

Operator: Your next question is from the line of Michael Rollins from Citi. Your line is now open. Please go ahead.

Michael Rollins: Hey, Dave. Thanks for the question, and good morning. Curious if we could approach revenue maybe in a different way. As you think about the path to significant year-over-year revenue growth, can you help us with the bridge, whether it's by product or by customer vertical, customer segment vertical, and just walking us through the pacing as you see it now and this path to improvement where the revenue can start to drive the EBITDA improvement for the company beyond the cost cutting that you've articulated? Thanks.

Dave Schaeffer: Yeah, sure, Mike. Part of the revenue decline has actually been helpful to EBITDA margin because of the product rotation. The premise of your question is correct, that aggregate top-line growth needs to improve to further accelerate the rate of margin improvement. While in the quarter, we outperformed our multi-year targets on margin improvement, over the long run, we expect that 200 basis points to be the more normal rate of annual margin improvement. In terms of customer segment, the enterprise segment is declining at the fastest rate, nearly 30% a year.

All of the customers acquired from Sprint are declining more rapidly, and in fact, the remaining revenue from those customers on an annualized basis is about $115 million, or $29 million in the most recent quarter, and continuing to decline at the fastest rate. It will be former Sprint customers going away. By network segment, it will be mostly off-net services, and in fact, you saw our on-net services grow on a blended basis at 5.4% year-over-year. There was a little bit of FX headwind in that, so it was closer to 6% without it.

It will be on-net, it will be organic Cogent sales, and it will be driven more by NetCentric sales, which is where the preponderance of our wavelength opportunity is. While the wavelength revenue is small, representing only about 7% of our total revenues, it did grow at 57%. We do anticipate wavelength being the fastest-growing product. IPv4 leasing, which grew at 5.4% year-over-year, continuing to grow at a healthy rate. We anticipate our on-net IP services growing, with transit being the larger portion of that growth for two reasons.

It's the bigger portion of the revenue, and two, it is benefiting from structural tailwinds where the corporate on-net business continues to face lower office occupancy and less days in the office, and has grown at a much slower rate than it had grown pre-pandemic. As we look at customer type, NetCentric, as we look at product, dominated by wavelength, and by network type, dominated by on-net. All of those things, coupled with the decline in the Sprint base, which is where the churn has been elevated from, will result in Cogent's total business going to top-line growth and a further acceleration in margin improvement.

Michael Rollins: Thanks very much.

Dave Schaeffer: Hey, thanks, Mike.

Operator: Your next question comes from the line of Michael Funk from Bank of America. Your line is now open. Please go ahead.

Michael Funk: Good morning, Dave. Thank you for the questions. I had two quick ones, Dave. First off, following up on the 2027 maturities, if you can't reach an agreement with the large credit investors, what are your options after that with those maturities creeping up on you in June 2026, or 2027, excuse me? Second one on the wavelength business, Dave, you mentioned some of the install being impacted by data center readiness or availability. How do you think about potential slowdown in data center build, AI spending, based on recent elections?

Dave Schaeffer: Yeah, sure. I'll take these in order. First of all, with regarding the 2027 notes, again, we have engaged both Evercore and Latham to help us. Two, some of the options we have are from totally new investors to Cogent that would provide sufficient financing to address that maturity. Whether or not that is the path forward is the reason why we engaged these two professionals to help us evaluate the options. The company continues to receive unsolicited offers for capital to address that maturity, and we have many paths forward in our organizational structure and our capital structure to provide us flexibility. All of these factors go into the analysis of what the optimal path forward is.

Again, we feel comfortable that we have chosen the correct advisors. They have ample options to evaluate, and there is a clear path forward for us. With regard to data center moratoriums, in an odd way, the slowdown in new data center construction could actually help, in the medium term, accelerate wavelength growth. The reason for that is, the diffusion of resources across more facilities has led to widespread delays on these facilities becoming fully operational. If those same resources are focused on facilities that are already permitted and under construction, maybe the slippage in their dates of in-service will stop, and therefore customers will be able to use wavelengths.

If you think about it, the wavelength is one of the last things that the customer needs. They need the building to be complete, they need the power, they need their equipment installed, and then they need a wavelength to connect that equipment to data repositories that are necessary for LLM or tokenization development. I think if all of the resources across the ecosystem are concentrated in fewer data centers, those data centers that are in process will come online more quickly.

Michael Funk: Thank you, Dave.

Dave Schaeffer: Hey, thanks, Michael.

Operator: Your next question comes from the line of Walter Piecyk from LightShed. Your line is now open. Please go ahead.

Walter Piecyk: Hey, Dave. Earlier in the call, you mentioned you upgraded the data centers to 20 kW from whatever the six. I think you said five or six. I think H100s need at least 30. Blackwells need 120. Are there plans to make further capital investments to get to those type of power? Can you, for these, I know it's a non-binding LOI, but for that, and you're doing it early 2027, so you can benefit from additional NOLs.

I assume that means same type of tax efficiency that you did in the last data center sale, and would that also apply in terms of tax efficiency to any future data center sales, or would they have to occur in 2028 or beyond?

Dave Schaeffer: Yeah, sure, Walt. I'm going to take those in reverse order. The reason why we have structured these two LOIs and the negotiations with the counterparties to close in early first quarter were designed to help us use NOLs that we have carried forward. We will have some additional capacity beyond these four data centers and there's approximately $100 million in proceeds that we can shelter. It is really hard for me to speculate on deals that have not yet even been negotiated, but we will look to monetize the remaining 10 data centers out of the 24 that we initially had targeted for sale.

We also have had some unsolicited offers for some facilities that were not part of that initial footprint, and we're evaluating those as well. Our belief was that these facilities were either too rural or too small to be of interest, but kind of tying into Mike's question previously, with many local moratoriums, the fact that these are preexisting data centers that have expandability as opposed to new greenfield may have some real value. It's kind of like a house. There are some neighborhoods where you can go and build a new house, but in many neighborhoods, you get permitting much more readily if you do renovations.

So you literally see people keeping one corner of the old house intact, building what is effectively a new house around it, calling it a renovation. We've had some parties approach us, and we're evaluating all of those options. With regard to the data center power availability, in the remaining 172 data centers that we own, we have approximately 150 MW of installed protected power. In many of those facilities, we have the ability to get incremental power, but only once we've used the power that is there. Within those facilities, they have typically started as telephone switches. All of the facilities that we inherited from Sprint were negative 48-volt DC power distribution.

We have converted those to AC 120, and we configured the drops to each cabinet at a standard power density of approximately 6 kW per cabinet. That is the norm in carrier-neutral data centers. We have had a number of requests for higher density cabinets. In a select subset, not all of the 172, we have adjusted the power drops to allow for 20 kW per cabinet. Now, there are needs in some cases for even greater density. These are repurposed older facilities and may not be suitable for every application, but there does appear to be substantial demand for that next tier up of power per cabinet. If we are successful-

Walter Piecyk: Got it. I understand, Dave, but basically none of those cabinets can handle an H100 is the bottom line.

Dave Schaeffer: That will be up to the customer to decide.

Walter Piecyk: I mean, they just can't. It's not how power to power. All right, so can we just go back also to liquidity? You said you substantially increased liquidity based on. By the way, H100 is legacy, right? The power requirements are even higher for the new compute. You said you substantially increased liquidity based on the sales, but when I think of liquidity, I think of actual cash on the balance sheet, and the cash on the balance sheet is obviously down because you used that to repay the debt.

When I look at your cash now, which is now 156.9, plus you've got some restricted stock, your revenue is still in decline, your free cash flow is still burning, you don't have a revolver. When you're looking at this refi, how is this negotiation working in terms of having enough cash on the balance sheet to withstand if perhaps maybe wavelengths don't grow as you expect and you can't grow revenue to actually generate free cash flow? Don't you need, or what is, I guess, at the root question here, Dave Schaeffer, what is the kind of bottom level of cash you need to have on your balance sheet to continue to operate smoothly or at all?

Dave Schaeffer: Right. Let me clarify a couple of your statements. We have $181 million of cash. We have no restricted stock.

Walter Piecyk: That's what I said.

Dave Schaeffer: We do have

Walter Piecyk: Well, I'm just looking at what's on the balance sheet, Dave. It's on the 10-Q.

Dave Schaeffer: No, restricted cash.

Walter Piecyk: 156.9 and 14.4. Yeah, restricted cash.

Dave Schaeffer: Right, restricted cash. You said-

Walter Piecyk: Literally on the balance sheet. Yep. Sorry.

Dave Schaeffer: I understand.

Walter Piecyk: Misspoke.

Dave Schaeffer: I am just clarifying what you said in your statement. We reduced gross leverage by $197 million in the quarter. We did not use all of the cash from the data center proceeds that we got from this transaction with I Squared Capital. We have more pending data center sales. There is no requirement to use that cash for gross debt reduction. We continue to reduce the amount of restricted cash, as Tad mentioned, it went from $168 million to $14 million, and that number may come down further. The company needs roughly $100 million of liquidity to have ample cash globally to function.

That is pretty typical of a company with a recurring revenue model where you need generally between four and six weeks of revenue to maintain optimal liquidity. We have grown our EBITDA consistently, we have reduced our capital intensity, and we are producing unlevered free cash flow and have additional liquidity coming in. In our discussions with the various options for refinancing, the options have many permutations, and for that reason, we have engaged advisors to help us and optimize it. That is all I can say at this point in time, but we feel comfortable with our liquidity position.

Walter Piecyk: Is the equity going to be untouched in that refi? Any additional potential warrants? You did not mention that as one of the options in the discussions that you were having.

Dave Schaeffer: I did not mention that, and I think it would be inappropriate for me to speak about any of the particular permutations. But we are very conscious of our equity value, and in fact, the company has worked diligently over the past 18 years to reduce its share count by 10.8 million shares.

Walter Piecyk: Got it.

Dave Schaeffer: Thanks, Walt.

Walter Piecyk: Just following up, one last question, Dave.

Dave Schaeffer: Sure.

Walter Piecyk: Because last quarter, you did mention, I forget who asked earlier about what you were anticipating in terms of a rate. I think you demurred on that, and I respect if you want to continue to demur, but last quarter, you did say that you thought maybe you would stick with single digits. But given kind of where the unsecured is now trading, or secured, it is implying a much higher rate. Obviously, there are ways, as you mentioned, three or four different ways to change that, but can you give us any other sense of where you think, on whatever the remaining principal balance, what the rate would look like?

Dave Schaeffer: As I mentioned earlier, we have multiple proposals that have different components that have trade-offs between rate and principal amounts. They are complex to evaluate, they have different collateral packages, and I think beyond that, it is inappropriate for me to comment while we have third parties advising us in these negotiations.

Walter Piecyk: And is there any timeline on when you expect to figure this all out?

Dave Schaeffer: Sometime between now and June 15th of 2027.

Walter Piecyk: Wow. Okay. Thank you. Appreciate it, Dave.

Dave Schaeffer: Hey, thanks, Walt.

Operator: Your next question is from the line of Nick Del Deo from MoffettNathanson. Your line is now open. Please go ahead.

Nick Del Deo: Hey. Morning, Dave. Thanks for taking my questions. First, with IPv4, what sort of strategies are you thinking about to accelerate leasing there that might be effective? Or is it just getting more challenging to push volumes into that market?

Dave Schaeffer: Nick, we took an approach initially of only leasing addresses to people that bought bandwidth. We've released that restriction, and the rate of growth materially accelerated. We then went to market with very low pricing, saw our unit volumes further accelerate, and then drove significant revenue growth by raising prices on customers because those addresses are a relatively small expense, and they are very sticky, meaning it's very difficult to renumber once they've been deployed. We are implementing a similar strategy going forward, where we are beginning to try to go to market with some introductory lower pricing and then maintaining the flexibility to then raise that pricing.

The analogy I would use is similar to cable operators or phone plans that offer you an introductory rate that's then supplanted by a more market-driven rate. The demand set is still very strong. I would say our biggest limitation is actually the strict adherence to abuse controls. We could probably almost instantaneously lease out the remaining inventory, but if we don't correctly vet the counterparties, we end up in a situation where those addresses quickly get polluted through inappropriate use and then become blocked by firewalls. It's a balancing act of us wanting to grow revenues, but also our need to ensure the quality of our customers to maintain the integrity of the asset.

But I do think we'll be able to continue to grow our IPv4 revenues.

Nick Del Deo: Okay. That's helpful. Then maybe one question on the data center monetization process. So, nice to see the LOIs being signed. What mechanical steps remain to getting those over the finish line? Then, with respect to the other 10 larger ones that you've been marketing, based on the discussions you've had and what you've heard from potential counterparties, what does your gut say about how many of those remaining 10 you'll ultimately be able to offload?

Dave Schaeffer: Yeah. So first of all, in terms of the sales process, there is earnest money deposit up that is refundable. We have done our diligence on the financial wherewithal of the counterparty, so we're confident they have the liquidity to be able to close. We have provided templates of a purchase and sale agreement that then will be negotiated. Both counterparties are completing their diligence. They've done several rounds of diligence. Probably the most significant of those are the commitments from the serving power utilities for power availability. Since while these sites have the power earmarked to them, they have not used that power for over a decade at full load.

So there just needs to be verification, usually in the form of a will-serve letter that the utility provides. Each utility deals with it in a slightly different way. We think that, based on the process we went through with I Squared, that these should close in the next several months, early in 2027. Then for the remaining 10, there's actually been activity, meaning there have been interested counterparties in all 10 of them. Many times, the counterparties fail our vetting process, and that's why we haven't proceeded with many of them. I do think that many of the 10, but probably all 10, will not be sold in the near future.

I do think, to the question that Mike asked earlier around the change in the political environment, there is an added value to these facilities that didn't exist maybe even earlier this week, in that because they are existing data centers that can just be expanded, that has a value above and beyond just the raw amount of power. Because if you can't get a permit for a new one, but you can get an expansion permit, that's a much easier path to operations.

In many states, there's a threshold of about 25 MW, and if you have an existing data center and want to upgrade it, and the upgrade stays below 25 MW at that single facility, it's a much more streamlined permitting process than building a new facility.

Nick Del Deo: All right. Thank you, Dave.

Dave Schaeffer: Hey, thanks, Nick.

Operator: We have reached the end of the question and answer session. I will now pass the call back over to Dave for closing remarks.

Dave Schaeffer: Hey, thanks. I want to thank you all for today's attendance. Hopefully, we've answered all the questions, and look forward to seeing you soon at various conferences. Take care, all. Thanks. Bye-bye.

Operator: This concludes today's call. Thank you for attending. You may now disconnect.

This article is a transcript of this conference call produced for The Motley Fool. While we strive for our Foolish Best, there may be errors, omissions, or inaccuracies in this transcript. Parts of this article were created using Large Language Models (LLMs) based on The Motley Fool's insights and investing approach. It has been reviewed by our AI quality control systems. Since LLMs cannot (currently) own stocks, it has no positions in any of the stocks mentioned. As with all our articles, The Motley Fool does not assume any responsibility for your use of this content, and we strongly encourage you to do your own research, including listening to the call yourself and reading the company's SEC filings. Please see our Terms and Conditions for additional details, including our Obligatory Capitalized Disclaimers of Liability.

The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.