CCOI - NASDAQ

Good morning, and welcome to the Cogent Communications Holding Third Quarter 2023 Earnings Conference Call. As a reminder, this conference call is being recorded, and it will be available for replay at www.cogentco.com. A transcript of this conference call will be posted on 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. I would now like to turn the call over to Mr. Dave Schaeffer, Chairman and Chief Executive Officer of Cogent Communications Holdings. Please go ahead.

Thank you, and good morning, everyone. Welcome to our third quarter 2023 earnings call. I’m Dave Schaeffer, Cogent’s CEO. 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 historical quarterly metrics, which we present in a consistent manner each quarter. Now, for a few comments on our results. We closed our acquisition of the Sprint business on May 1, 2023. This transaction significantly expanded our network, our customer base, and materially increased the scope and scale of our business. Our annualized revenue run rate now exceeds $1 billion. We acquired a number of large enterprise customers, many of which are Fortune 500 companies. Customers that are typically larger than our typical Cogent corporate customer base. We acquired significant owned fiber optic roots and facilities. We acquired numerous right-of-way in relationships with the underlying landowners, which represent over tens of thousands of route miles of dark fiber. These assets and relationships would be virtually impossible for us to assemble on our own. We hired many valuable experience Sprint business employees. Many of these sprint business employees had an average tenure with the company of 22 years prior to the acquisition. We acquired a network with an appraised value of over $1 billion for $1. We will receive a total of $700 million from T-Mobile to offset operating losses for serving enterprise customers and for providing T-Mobile IP Transit Services. $350 million in the first year for $29.2 million in monthly installments, and then $350 million over the next 42 months for a monthly installment of $8.3 million per month. We are very optimistic about the cash flow generating capabilities of the combined operation. Our recent results show we achieved immediate and substantial savings in many areas, many of which exceeded our initial expectations. We anticipate additional substantial cost savings from our current run rates in many areas. As we had mentioned in our last earnings call, we have combined the Cogent classic legacy business and the Sprint business operations. As a result, we will no longer be reporting separate metrics. The combined Cogent business had a very good quarter. Our total revenues were $275.4 million. Our operations from the quarter include a full quarter of the Sprint business versus 2 months, as reported in Q2. Our EBITDA as adjusted was $131.4 million, an increase by $77.4 million from Q2 of 2023. Our EBITDA as adjusted margin was a record at 47.7%, a significant increase from the EBITDA as adjusted margin last quarter of 25.2%. We received three payments totaling $87.5 million under the Transit Services Agreement from T-Mobile in the quarter versus only one IP Transit Service payment of $29.2 million in Q2. Our gross debt to trailing 12 months EBITDA as adjusted and our net debt ratio both significantly improved in the quarter. Our total gross debt to trailing 12 months EBITDA as adjusted ratio was 4.56 at the end of the quarter and our net debt ratio at the end of the quarter was 4.24. This is compared to a gross debt of 5.63 times in Q2 and a net debt of 4.79. We anticipate further improvements in in these leverage levels over the next several quarters. Our network traffic increased sequentially by 6%. It was all 26% year-over-year. This traffic growth acceleration was better than the 4% sequential growth rate we had seen in Q2, and the 21% year-over-year traffic growth. We’ve begun to realize synergies and over the next 3 years, we will continue to anticipate achieving annual savings of $180 million annually from the Sprint North American network, $25 million from the Sprint International wireline network, and a $15 million reduction in O&M expenses for Cogent’s North American network. We anticipate achieving additional SG&A savings and other cost and revenue synergies over the next several years. Our recent progress in achieving these savings is very encouraging and in fact exceeded our initial targets on savings. Our total revenue for the quarter increased sequentially about 14.9% to $275.4 million and increased by 83.6% on a year-over-year basis. Our rep productivity at 9.2 last quarter and 3.6 this quarter for full-time equivalents. This number included the full-time equivalent productivity because of the 9,000 commercial services orders sold T-Mobile under our commercial services contract. This commercial service contract with T-Mobile is in addition to our $700 million IP Transit contract. Our rep productivity results also included the impact of the enterprise customer sales reps that we had acquired from Sprint, which are now counted as full time equivalents and are continuing to receive training on Cogent’s products and are not yet fully productive. In connection with the Sprint acquisition, we hired a total of 942 employees. During the quarter, our total sales reps actually decreased by 10 or 1.5%. We ended the quarter with 637 sales reps and 621 full-time equivalents. A 9.5% increase in full-time equivalent reps, primarily due to the reps that were hired from the Sprint business now being counted as full-time equivalents. Now, for a comment on our optical transport services or wavelength businesses. In connection with our acquisition of the Sprint business, we’ve expanded our offerings of optical wavelength services and optical transport network to utilize the Sprint network. We’re selling these wavelength services to existing customers, acquired customers from Sprint, and to new customers. The customers require dedicated optical transport without the capital and ongoing expenses of owning their own infrastructure and network. Our wavelength revenue for the quarter was $3 million and there were 449 wavelength customer connections at quarter end. We have sold wavelengths in a total of 50 locations with shorter provisioning cycles. We have connectivity and wavelength sales capabilities in over 250 locations, but with longer provisioning cycles. In approximately 14 months, we expect to be able to offer wavelengths in over 800 carrier neutral data center locations in the U.S. with more rapid provisioning cycles. Our Sprint acquisition material expanded our network footprint. We added 18,905 route miles of intercity-owned fiber; 12,570 route miles of metropolitan-owned fiber network. We added approximately 11,400 route miles of intercity IRU fiber and approximately 4,500 metro route miles of IRU fiber to the Cogent network. Most of this is redundant and will be eliminated as part of our cost saving measures. We eliminated approximately 430 redundant fiber route miles that were released in the quarter. We will reconfigure 45 of the acquired Sprint facilities into data centers, and add those to the 1,528 carrier neutral data centers that we operate in and the 60 proprietary Cogent data centers. To date, we have converted four of these acquired Sprint facilities into Cogent data centers. Now, for a comment on our dividend program. During a quarter, we returned $45.1 million to our shareholders for our regular dividend. Our Board of Directors which reflects on our strong cash flow generating capability and investment opportunities, including the Sprint acquisition decided to increase our quarterly dividend yet again by an additional $0.01 per share sequentially. Raising our dividend quarterly from $0.945 per share to $0.955 per share. This represents the 45th consecutive sequential increase in our regular quarterly dividend, and a 4.4% annual growth rate in our dividend. Now, for a comment on our future guidance and expectations. Now that we have combined the Sprint business with the Cogent business, we anticipate long-term average revenue growth of between 5% and 7% per year, and EBITDA margin expansion of approximately 100 basis points per year. Our revenue and EBITDA guidance were intended to be multi-year goals and are not intended to be used as specific quarterly or annual guidance. Our EBITDA as adjusted and our leverage ratios were impacted by the $700 million IP Transit Services Agreement that we are entered into with T-Mobile. Beginning in May of 2024, these payments will be reduced from $29.2 million per month to $8.3 million. That reduction will impact future EBITDA as adjusted in our leverage ratios beginning in the second quarter of 2024. However, these metrics are measured on a trailing 12-month basis. Now, I’d like to ask Tad to read our safe harbor language and provide some additional operating performance metrics for the quarter. Following our remarks, we will open the floor for questions and answers.

Thank you, Dave, and good morning to everyone. This earnings conference call includes forward-looking statements. These forward-looking statements are based upon our current intent, belief, and expectations. These forward-looking statements and all other statements that may be made on this call that are not historical facts, are subject to a number of risks and uncertainties, and actual results may differ materially. Please refer to our SEC filings for more information on the factors that could cause actual results to differ. Cogent undertakes no obligation to update or revise our forward-looking statements. 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 services. And we also analyze our revenues based upon customer type. We classify all of our customers into three types, net-centric customers, corporate customers, and enterprise customers. Our corporate customers buy bandwidth from us in large multi-tenant office buildings or in carrier neutral data centers. These corporate customers are typically professional services firms, financial services firms, and educational institutions located in multitenant office buildings, or connecting to our network through our carrier neutral data center footprint. Our net-centric customers buy significant amounts of bandwidth from us and carrier neutral data centers, and include streaming companies and content distribution service providers, as well as access networks who serve consumer and business customers. Our corporate and enterprise customers generally purchase our services based upon a price-per-location, and our net-centric customers purchase our services based upon price-per-megabit. Comments on the corporate business. Our corporate business continues to be influenced by real estate activity in central business districts. Two key statistics including the level of cards, rights and buildings, and leasing activity indicate that year-to-date the real estate market and leasing activity in central business districts, where we operate has seen some continuing improvement in certain areas of the country, but has not returned to pre-pandemic levels in most geographic regions. We continue to remain cautious in our outlook for our corporate revenues given the uncertain economic environment and other challenges from the lingering effects of the pandemic. Our corporate business represented 43.7% of our revenues for the quarter, and our quarterly corporate revenue increased year-over-year by 40.9% to $120.5 million from last year and increased sequentially by 8.5%. We had 55,045 corporate customer connections on our network at quarter end. This was a sequential decrease of 10.2% and a year-over-year increase of $21.8. The sequential net decrease in corporate customer connections primarily resulted from the elimination of a non-core product for corporate customers. During the quarter, we eliminated 8,486 Session Initiation Protocol, or SIP, non-core customer connections of which 5,006 of these 8,400 were non-core corporate customer connections. Excluding the impact of the SIP corporate customer connections that reached end of life, our corporate customer connections decreased by 1,200 connections or by 2.2% from last quarter and that was also due to some other non-core products being end of life. For the quarter, the sequential impact of USF on our corporate revenues was a positive $3.5 million and a positive year-over-year impact of $10.4 million. Some comments on the net-centric business. Our net-centric business continues to benefit from continued growth in video traffic and streaming. For the quarter, our network traffic growth accelerated and was up by 6% sequentially and 26% year-over-year. Our net-centric business represented 34.5% of our revenues for the quarter and grew sequentially by 8.4% to $94.9 million, and grew by 47.2% year-over-year. We had 62,291 net-centric customer connections on our network at quarter end. That was a sequential decrease of 6.6% and a year-over-year increase of 21.8%. Explaining the sequential decrease included in our net-centric customer connections at the end of last quarter for 8,028 net-centric customer connections under the commercial services agreement with T-Mobile that Dave mentioned earlier and 1,088 of the SIP customer connections that product that reached end of life. At the end of the quarter, there were 4,661 net-centric customer connections under the commercial services agreement with T-Mobile. If you exclude the impact of both these net-centric customer connections under the T-Mobile commercial services agreement and the SIP product that reached end of life in both periods. Our net-centric customer connections increased sequentially by 35 connections. Our enterprise business was 21.8% of our revenues for the quarter. We had 20,689 enterprise customer connections at the end of Q3. There were 2,392 SIP enterprise connections at the end of last quarter that reached their end of life during this quarter. Again, all the SIP connections were canceled during the quarter, whether they were corporate, net-centric, or enterprise, and all of that product was non-core. Revenue and customer connections by network type, on-net revenues. Our on-net revenue was $130 million for the quarter. That was a sequential increase of 1.9% and year-over-year 14.9%. Our on-net customer connections were 89,623 at the end of the quarter. We serve our on-net customers in 3,257 total on-net multi-tenant office in carrier neutral data center 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. Selling these larger connections has the impact increasing our year-over-year on-net ARPU. Off-net revenue. Our off-net revenue was $131 million for the quarter. Sequential increase of 28.4% and year-over-year increase of 257.7%. Including our new off-net locations from the Sprint business, we now serve off-net customers in over 27,800 off-net buildings. These off-net buildings are primarily located in North America. Wavelength revenue. Our wavelength revenue increased by 88.8% sequentially and was $3 million for the quarter. Our wavelength customer connections were 449 at quarter end. Non-core revenue. Our non-core revenue was $11.4 million for the quarter. Non-core customer connections were 11,187 at quarter end. At the end of last quarter, total non-core customer connections, again included the 8,486 SIP customer connections that reached our end of life this quarter. Comments on pricing. Our average price per megabit for our installed base increased sequentially by 7.7% to $0.30 and also increased year-over-year by 9.6%. Our average price per megabit for our new customer contracts was $0.17 and that was almost 64% increase over $0.10 from last quarter and a year-over-year price increase of 8.8%. So increases all around. Comments on ARPU. Our on-net and off-net ARPUs for the quarter decreased sequentially primarily from the impact of the Sprint business ARPUs. However, our year-over-year on-net and off-net ARPUs increased. Our on-net ARPU decreased sequentially by 1.7% from $483 to $475. And year-over-year our on-net ARPU increased by 3.8% and it was $458 last year for the third quarter. Our off-net ARPU decreased sequentially by 10.7% from $1,294 to $1,156. And year-over-year our off-net ARPU increased by 25.6% and it was $920 last year for the third quarter. Churn rates. Our sequential churn rate for our on-net connections for the combined business did increase from the impact of the Sprint business this quarter. Our on-net unit churn rate was 1.8% for the quarter, up from 1.4% last quarter. Our off-net unit churn rate was 1.5% for the quarter which was a slight decrease from 1.6% last quarter. These on-net and off-net churn rates do not include large number of non-core churn customer connections. On EBITDA. We reconcile our EBITDA to our cash flow from operations in each of our quarterly press releases. We incurred $0.4 million of Sprint non-capital acquisition cost this quarter compared to $0.7 million last quarter. Our EBITDA for the quarter increased sequentially by $19.4 million and decreased by $14.3 million year-over-year. Our EBITDA margin increased to 15.8% from 10.1% last quarter. Now, on EBITDA as adjusted. Our EBITDA as adjusted, which includes adjustments for Sprint acquisition costs and the cash payments received under our $700 million IP Transit Services Agreement with T-Mobile. We billed and collected $87.5 million under the IP Transit Services Agreement this quarter. Last quarter, we billed $58.4 million and collected $29.2 million under the agreement. All amounts billed under the IP Transit Services Agreement have been paid on time, and as of this call, we have received two additional payments. Our EBITDA as adjusted for Sprint acquisition costs and cash payments received under the $700 million IP Transit Services Agreement, was $131.4 million for the quarter. That was a 47.7% EBITDA as adjusted margin. This EBITDA as adjusted margin is a company record and a substantial increase from 22.5% last quarter. And the increase was from both additional payments under the IP Transit Services Agreement and cost reduction, and an increase in revenue. Foreign currency impact. Our revenue earned outside of the United States is reported in U.S. dollars and was approximately 16% of our revenues this quarter. About 10% of our revenues this quarter were based in Europe and 6% of our revenues related to our Canadian, Mexican, Oceanic, South American and African operations. The average euro to U.S. dollar rate so far this quarter is about $1.6, and the average Canadian dollar exchange rate is about CAD 0.73. If these average foreign exchange rates remain at their current levels for the quarter, we estimate that the FX conversion impact on our sequential revenues would be a negative about $1 million and the FX conversion impact year-over-year would be a positive of about $0.8 million. Customer concentration. We believe that our revenue and customer base is not very highly concentrated, although it has increased with the Sprint acquisition. Including the impact of the customers acquired in the Sprint business, our top 24 customers represented about 19% of our revenues this quarter. We acquired a number of larger enterprise customers with the Sprint business, and we are providing services to T-Mobile under the commercial services agreement. Our quarterly CapEx materially decreased and was $25.4 million this quarter compared to $37.4 million last quarter. On finance leases and finance lease payments, also known as capital leases. Our finance lease IRU obligations are for long-term dark fiber leases and typically have initial terms of 15 to 20 years or longer, and often include multiple renewal options after the initial term. Our IRU finance lease obligations totaled $483.2 million at quarter end. We have a very diverse set of IRU suppliers and IRU contracts with over 315 different dark fiber suppliers. We acquired relationships with several new suppliers of dark fiber with the Sprint business. During the quarter, we recorded a purchase accounting measurement period adjustment to reclassify a lease agreement from right-of-use operating lease assets acquired from T-Mobile to finance lease assets. This adjustment under U.S. GAAP Accounting Standard 842 accounting for leases resulted in a reclass of almost of about $161 million from our acquired operating lease liabilities to a finance lease liability. The corresponding adjustment also reduced our cost of goods sold run rate by $12.6 million per quarter and increased our depreciation expense by about $11 million in the quarter. Some comments on cash and operating cash flow. At quarter end, our cash and cash equivalents and restricted cash totaled $166.1 million. Our $56.4 million of restricted cash is tied to the estimated fair value of our interest rate swap agreement. And as of November 6, so just recently, the swap valuation reduced from $56.4 million to $43.8 million. Our operating cash flow results are materially impacted by the timing and amount of our payments under our Transition Services Agreement with T-Mobile and the presentation of the payments under our $700 million IP Transit Agreement. Payments under the IP Transit Agreement under U.S. GAAP, are considered cash receipts from investing activities and are not classified as operating activities. Our operating cash flow was a use of $52.4 million for the quarter compared to a positive $82.6 million last quarter, mostly from the impact of the Transition Services Agreement. Last quarter, the TSA agreement provided $118.8 million of operating cash flow, since no payments were due or made during the quarter. This quarter, we paid $153.1 million under the TSA, under their terms and when the payments were due. Combined with amounts billed under the TSA for the quarter, net cash provided from the TSA was $9.5 million this quarter and that change from last quarter under this one-line item is $109.3 million. Most of the amounts paid under the TSA are for direct reimbursement of Sprint business vendors paid by T-Mobile on our behalf. We have transitioned a significant majority of these payments to our payable systems and expect to transition the remaining vendors by the end of the year. IP Transit payments under the IP Transit Agreement to $700 million. Our payments received under the IP Transit Agreement are recorded as cash provided by investing activities, and were $29.2 million last quarter compared to $87.5 million this quarter. Total net cash used in investing activities was $22.3 million last quarter, and cash provided by investing activities was $62.1 million this quarter. That was a sequential quarterly increase of cash of $84.4 million for this line item investing activity. As Dave mentioned debt and debt ratios. Our total gross debt at par, including our finance lease obligations, was $1.4 billion at quarter end, and our net debt was $1.3 billion. Our total gross debt to last 12 months EBITDA as adjusted, and our net debt ratio both significantly improved this quarter. Our total gross debt to last 12 months EBITDA as adjusted ratio was 4.79, and net debt ratio was 4.24. This was a material improvement and compared to gross debt to trailing ratio as adjusted of 5.63 last quarter and net debt ratio of 4.56. Our consolidated leverage ratio, as calculated under the note indentures reduced to 5.09 from 5.30 last quarter. Our secured leverage ratio as calculated under our note indentures increased slightly up to 3.5 from 3.45. Additional comments on the swap agreement. We are party to an interest rate swap agreement that modifies our fixed interest rate obligation associated with our $500 million 2026 notes to a variable interest rate obligation based upon the secured overnight financing rate for the remaining term of our 2026 notes. We record the estimated fair value of the swap agreement at each recording period and we occur corresponding non-cash gains or losses due to the changes in market interest rates. The fair value of our swap agreement increased by $4.8 million from last quarter at quarter end to a liability of $56.4 million. We are required to maintain a restricted cash balance with the counterparty equal to the liability. And as I mentioned previously, as of November 6, our swap valuation reduced to $43.8 million. Lastly, some comments on bad debt and DSO. Our day sales outstanding or DSO remain stable. Our DSO for worldwide accounts receivable was 24 days, the same as last quarter. In the fourth quarter, we will be converting the billing of the Sprint business customers to the Cogent billing platform, and in fact, we just completed that process. Our bad debt expense was only $0.8 million and only 3% of our revenues for the quarter, outstanding results. Again, we want to thank and recognize our worldwide billing and collections team members, including our new billing and collections employees from the Sprint business. We’re doing a fantastic job in serving our legacy Cogent customers and our new Sprint customers, and collecting for them and converting the Sprint billing to the Cogent billing system. I will now turn the call back over to Dave.

Hey, thanks, Ted. I’d like to highlight a couple of strengths of our network, our customer base and sales force. We continue to experience significant revenue and traffic growth in our legacy, net-centric business. We are direct beneficiaries of increased streaming over-the-top video and other streaming applications, particularly in international markets. At quarter’s end, we were on-net in 1,528 carrier neutral data centers and 60 of Cogent-owned data centers for a grand total of 1,588 data centers, more than any other carrier as measured by independent third-party research. The breadth of our coverage enables our net-centric customers to better optimize their network and reduce legacy. We expect that we will continue to widen this lead in the market and project to add approximately an additional 100 carrier neutrals per year to our network footprint over the next several years. We expect to convert an additional 41 of the former Sprint technical facilities to new Cogent data centers. To date, we have converted four of these facilities. As of today, we are selling wavelengths in 50 carrier neutral data centers with reasonable provisioning windows, and we are selling wavelengths in 250 additional locations with a prolonged ability to install those services. In the next 14 months, we expect to be able to sell wavelength services in 800 U.S. carrier neutral data centers with substantially reduced provisioning times. Our network traffic growth accelerated in a quarter. Our traffic growth increased by 6% sequentially and 26% year-over-year. We expanded our footprint and have additional IRUs and owned fiber and right-of-way agreements. At quarter’s end, we are the most interconnected network in the world with 7,971 networks directly connected to Cogent. This collection of ISPs, telephone companies, cable companies, mobile operators, and other carriers allow us to reach directly the vast majority of the world’s broadband subscribers and mobile phone customers. At quarter’s end, we had 257 sales professionals solely focused on the net-centric market. This group of professionals is one of the largest, most successful sales teams in the industry. This team will be primarily responsible for growing our wavelength market. Now for a couple of comments on our corporate business. We are seeing positive trends in our corporate business. Corporate customers are increasingly integrating new applications and part of their working environment includes the regular use of streaming video conferencing. This requires high capacity connections both inside and outside of their premise. Our aggressive push to lower bandwidth costs and provide greater coverage has begun to boost corporate demand for bidirectional symmetric 1-gig and 10-gig ports. Corporate customers are increasingly buying connections in carrier neutral data centers for redundancy for the ad hoc VPNs that support a hybrid work environment. Our enterprise customers continue to focus on our dedicated Internet access and VPN services. We are continuing to terminate non-core products and the significant reduction that you saw in the last quarter was the elimination of our voice services based on session initiation protocol. Now for a comment on our sales force productivity, we remain focused on training and improving our sales force efficacy. We do manage out under performers. On a sequential basis, our total headcount and sales decreased by 10 to 637 reps. Our sales force turnover rate was stable at a 5.7% per month for the quarter, down from a peak of 8.7% during the pandemic and consistent with our long-term average rate of sales force turnover of 5.7%. We remain very optimistic about our unique position to serve small- and medium-sized businesses located in the 1,860 multi-tenant office buildings representing over 1 billion square feet of rentable space in central business districts. We’re excited about the addition of a large enterprise customer base and our ability to sell optical transport networking to our net-centric as well as enterprise customers. We also remain encouraged by demand in our data center footprint and our ability to expand that footprint. We have a significant backlog and funnel of wavelength orders of approximately 1,000 unique wavelengths. However, due to longer provisioning cycles that we are temporarily experiencing, we are unsure if all of these opportunities will convert into installed orders. Currently, we see key indicators of office activity, work re-entry, and leasing activity all improving as tenants continue to require employees to return to their offices and commercial office vacancy rates have continued to decline. Certain corporations have decreased the amount of square footage they’re taking per location, which will ultimately allow us to grow our addressable market for unique corporate customers. Under our indenture, including our $250 million general basket, we have a cumulative amount of cash available for dividends and buybacks that actually exceeds the amount of cash that we have available. So we have effectively the ability to use all cash for shareholder friendly activities, whether it be dividends and/or buybacks. We are diligently working on integrating the Sprint business. We’re excited and optimistic about the cash flow generating capabilities. We’ll continue to achieve annual savings of about $180 million on the North American network, $25 million internationally, and $15 million in reduction in Cogent expenses. We also anticipate additional SG&A savings and other cost and revenue synergies over the next several years. Now, I’d like to open the floor for questions.

[Operator Instructions] Our first question comes from Greg Williams from TD Cowen. Your line is now open.

Great. Thanks for taking my questions. Dave, I was hoping you can unpack the EBITDA beat by a little bit. I know you’ve integrated the Sprint business with Cogent, but the way so the investors are looking at it is in the core EBITDA for Cogent, typically around $60 million. You’re going to add another $87 million to the T-Mobile payment. And then we’re subtracting the EBITDA burn from the Sprint GMG business. And it sounds like you cut more from Sprint GMG than we appreciate it. So I guess the question is, typically you said you’re at a run rate when you close the business at $180 million burn. Where’s that run rate today? And the second question then is, if you are cutting more than we expected, do you still anticipate breaking even by year two – or end of year two? Or could that be accelerated to breakeven faster? Thanks.

Yeah. So as Tad mentioned, we’re no longer segregating the customer. So the acquired Sprint customers are now fully integrated in the Cogent, and are part of our single accounting and single billing platform. We are ahead of schedule in terms of reducing the expenses of that business, probably the most significant savings comes from the immediate elimination of gross margin negative products. The nearly 9,000 SIP services that we were able to eliminate in the quarter allowed us to improve margins. That termination did not occur until the end of the quarter. So it was a September 30th termination. These customers were given notice of end-of-life of these products a year-ago, when the initial transaction was signed with T-Mobile in early September of 2022. As part of that agreement, T-Mobile agreed to end-of-life a number of products. The SIP product was the largest of those products. The 24 products that we identified for elimination carried negative gross margins. So there are either direct costs or direct employee involvement in those products that we’ve been able to eliminate. We are not in a position today to modify or breakeven 2 years post-closing, but we do feel right now we are running ahead of schedule on those cost synergies and should be able to continue to achieve better cost reductions. There are other non-core products beyond SIP that are also being eliminated. You saw that in the 1,233 products that were eliminated for corporate customers in the quarter that were non-SIP products. Not all of that reduction was non-core, but the majority of it was. So what we’re trying to do is as quickly as possible manage customers away from these unprofitable products. But also, as part of our agreement with T-Mobile, we are going to honor each customer’s contract. And customers have various contractual provisions that require us to provide some of these services through the end of 2026. So it is why we were so definitive in our ability to schedule out the reduction in cost. Now in some cases, we’ve been able to convince customers to allow us to terminate these products early, and that is an added benefit to us in helping us reduce the cash burn.

Got it. Thank you.

Hey, thanks, Greg.

Our next question comes from David Barden from Bank of America. Your line is now open.

Hi, good morning. You got Alex [ph] on for Dave. Dave, maybe just first question here, just in terms of wavelengths, I think it came in a little bit lighter than our expectations. And can you just kind of frame the opportunity and where you think you could be in wavelength revenues at the end of 2024? And then secondly, on SG&A, sorry if I missed it on the call, but can you talk a little bit more about what drove the decrease quarter-over-quarter? And then what kind of the run rate we should expect here for SG&A heading into the next year? Thanks.

Yeah. So I’ll take those in order. On wavelengths, as we commented on the previous earnings call, we had a very narrow set of locations where we can provision within a 60-day window and a larger, but still not adequate set of locations where we can provision within a 120-day window. We have increased the number of locations, where we can provision within that shorter window. We’ve also partially reduced that window down from 60 to about 50 days, still not the 17-day SLA target that we give our transit services. We have over a 1,000 unique wavelength orders either in provisioning or in our sales funnel. That would meet all of our growth objectives, and there will be additional orders being added to that funnel on a daily basis. I want to caution though that some of these orders will have 120- to 130-day provisioning, and as a result some of those orders may not ultimately get installed, customers may be frustrated. While we are trying to manage those expectations, we have a number of foundational steps that we’re taking to streamline our provisioning of wavelength services. We’ve already provisioned this quarter about 150 additional wavelengths just since the close of last quarter. So basically one-third of what we had in the base just got provisioned in the past 6 weeks. We expect that pace to continue to accelerate. We also believe that we will be on a revenue run rate of close to $100 million in wavelength sales post-closing. So in May of 2024, the monthly run rate for wavelength sales should be in the order of about $8 million a month. And our sales funnel, I think supports that and the continued interest that we’re seeing due to the uniqueness of our routes and the ubiquity of locations we wish to serve, I think, are all indications that we’ll do better. On SG&A, some of it is headcount reduction, some of it is facilities consolidation. Those are the two main areas that we’ve been able to achieve SG&A improvements.

There was also a substantial improvement in bad debt expense quarter-to-quarter. That was a $4 million reduction. So we were less than 1% of revenues this quarter. Our typical bogie is about 1% of revenues. So we outperformed on that and bad debt expense was abnormally high last quarter, as we had to reestablish some reserves.

I think we’re at three-tenths of a percent this quarter, which I think is the lowest bad debt expense we’ve had in the company’s history.

Thank you, both.

Our next question comes from Frank Louthan from Raymond James. Your line is now open.

Great. Thank you. I want to talk about the outlook for the business if return to office doesn’t really improve. So how much of that long-term guide that you’ve given out is kind of dependent on an improvement in the return to office environment and what level of occupancy do we kind of need to hit over the long-term to achieve that? And then what are your options if that doesn’t happen to still kind of hit that long-term guidance goal? Thanks.

Yeah, so our long-term growth targets of 5% to 7% are predicated on office occupancy and corporate performance being similar to the current levels that they are at today. Again, we have diversified our total revenue base, now having three discrete segments, being less exposed to pure corporate growth. Our enterprise customers tend to be much more global in their footprint. Just to remind you, 44% of total revenues are corporate, 34% are net-centric, 22% are enterprise. Our net-centric business has actually continued to outperform long-term averages and our ability to have a total revenue growth rate in that 5% to 7% range is possible with vacancy rates remaining elevated at about 15% in central business districts. While we believe that vacancy number will come down, we can achieve our growth rate and our guidance targets at these elevated levels. And our guidance is also predicated on our net-centric business moderating, which it continues to accelerate, as you saw in the traffic stats that we provided sequential growth growing from 4% sequentially last quarter to 6% year-over-year growth growing from 21% to 26%. So material improvements in that business continuing longer. And then, finally, we are very optimistic about our wavelength opportunity based on the breadth of our sales backlog.

And you mentioned on the waves at an $8 million run rate for sales, what is sort of a quarterly run rate of conversion to that from sales to kind of what you’ll be able to recognize in the quarter?

So obviously, the growth rate we achieved in this quarter on waves of 88% sequentially is not sustainable. That growth rate will moderate. I would say that for the, I guess, second quarter of 2024, wavelength sales will probably be in the $20 million range for the quarter, but building throughout the quarter.

All right. Great. Thank you, Dave.

Our next question comes from Walter Piecyk from LightShed. Your line is now open.

Hey, Dave. Just some questions on corporate. I know there’s a lot of moving parts now that you have that Sprint T-Mobile business in there, but it looks like on a pro forma basis, it was down sequentially. I’m just curious when you expect that to grow on a sequential basis.

Actually, I would disagree with it being down. I think the majority of what was down in corporate was the elimination of the SIP product and other non-core products. We did acquire corporate customers from Sprint as well as enterprise customers. If we looked at our MTOB footprint, we actually saw the number of connections grow. So I would kind of disagree with the premise of your statement.

I’d be great if you actually reported, I know, Tad in his prepared remarks gave pro forma sub-growth type numbers, but maybe providing pro forma revenue and EBITDA would have been more helpful. So if we look at the fourth quarter then, is there going to be a similar type of excuse in terms of the shutdown at the end of the fourth quarter? Or should there be actual sequential growth in the fourth quarter?

Well, as we have said, there are still non-core products that we are trying to eliminate as quickly as we possibly can. These products carry negative gross margin. They were a significant contributor to the losses at T-Mobile, and SIP was the largest of these products. But, the run rate...

So what is the baseline that for corporate of non – like, can you give us some type of comparable? So we don’t always have the excuse of like, oh, we just churned off non-core stuff like what is the baseline in corporate revenue of stuff that’s generating gross margin?

So we report the non-core products separately, and we had a run rate of those non-core products of about $11 million. That will go to zero, or it’s as close to zero as possible.

So is that $11 million in the $120 million from the quarter for corporate revenue, that was reported? Any of that $11 million?

$11 million is non-core revenue.

But non-core counts corporate, yes. So the majority of that is in corporate.

So your baseline in corporate is effectively basically one, whatever it is, $120 million minus $11 million, and we’ll just have to get an update on the $11 million every quarter, and then we’ll figure out where your true organic growth rate is? I would suggest maybe you actually just put that in the press release and provide that as a pro forma number to give better transparency to what’s going on in the business. I also have a question on the lease expense. So you moved, I think it was like $12.5 million out of OpEx, so you boost your EBITDA by $12.5 million into CapEx. So if I look at that lease number on CapEx, I think that was like $40 million.

It’s not in CapEx, it’s in principle payments, because it is a lease payment. It is not a CapEx payment.

Got it. Now I understand. So whenever I do this for the telcos, telcos always like to try and exclude that from CapEx, I consider that CapEx. So that’s fine. But bottom line as you moved it, you helped EBITDA and you moved it onto the cash flow statement, fine. Is that a recurring $12.5 million? Is that like how do we look at that number, because obviously free cash flow is ultimately everything that matters?

Right. Yeah, that’s the change in the accounting. So it will continue until the lease expires.

And as we indicated, in the appraisal for the acquired assets, there was approximately $150 million of uneconomic lease obligations that reduced the appraised value to get to the $1 billion value for the network. And the majority of that was associated with this lease. And this lease, in fact, met all of the criteria to be treated as a financing or a capital lease as opposed to an operating lease. This was just a correction, but it will continue until that lease expires.

But the capital lease principal payment, I guess, is how you’re calling it, was $41 million. So of that $41 million, only $12 million was moved out of OpEx helping your EBITDA. Is that, do I have that right?

Yeah, that about right.

That’s correct.

Okay. So what was the other, because last quarter was $8 million. So what are the other $30 million or $25 million payments that are capital leased? And is that an ongoing payment that’s going to pinch your free cash flow?

So we have a total of 315 different suppliers and over 3,000 unique IRUs. Those IRUs fluctuate, some are paid annually, some are paid quarterly, some are paid monthly, some are paid upfront. And you can go back and look at the principal payment on capital lease run rate that we’ve had and see a fair amount of sequential volatility based on the existing [ph] quarter or annual, but you can take the annualized rate and extrapolate that and add this additional $12 million a quarter and that would get you to a...

So what is the annual rate? Because legacy obviously didn’t include Sprint. So if I’m adding $12 million, what is the legacy? Because I look at 2022 and you had $45 million. So you’re saying that it’s basically going to be $50 million.

This is the only finance lease with the acquired business.

This is the only incremental lease.

Got it.

Only finance.

So we’re talking like $60 million on a go forward basis for what you call principal payments, what I call CapEx, but whatever you want to call, it impacts the free cash flow.

That’s correct.

Got it. Awesome. Thank you, Dave.

Yeah.

Our next question comes from Tim Horan from Oppenheimer. Your line is now open.

Thanks a lot, guys. Dave, can you just level set where you think EBITDA margins will be 4 or 5 years from now? Also maybe where you think the wavelength revenue run rate will be at that time? And then just a clarification on your $20 million of sales of wavelength in the second quarter next year. Is that bookings or is that recognized revenue? I mean, will it take another quarter or two to recognize that revenue? Are you talking about actually sales bookings or you actually, when you use the word sales, [that revenue will cost?] [ph] Thanks.

We’re actually talking about revenue run rate recognized installed revenue. And, again, to be clear, we anticipate this backlog that we have as well as additional sales to begin to install with shorter and shorter provisioning windows. But we expect to be doing about $20 million of reported, not bookings, but reported. And, the idea of presenting a backlog number is something, and a bookings number is something that we historically don’t do and probably will do for the next couple of quarters until investors see a consistent cadence in our growth and wavelength revenue. And at that point, we will only be reporting actual revenue, not pipeline and funnels and provisioning cues. But I think in the short-term, that’s necessary. And then to go to answer your question around a 5-year target, and I’m going to use 5 years from closing, so May of 2023 to May of 2028. We anticipate being on a run rate for wavelength sales of approximately $500 million, and a total run rate for the combined business in excess of $1.5 billion, up from the [$1.1 billion of $1.1 billion, $51.2 million] [ph] that we’re running at right now, and EBITDA margins in the mid-30s. And that will be, well, actually, it’ll probably be a little bit above that, because we will still be getting a payment stream from T-Mobile, which will be counted, but will be going away probably by year 6.

Very helpful. And just the way to big market, can you give us just a little bit more color, what’s going on, do you think with the overall growth in that market and volumes and pricing and the competitive dynamics, now that you’ve had more time to be in that market?

Yeah, I mean, we did a fair amount of customer outreach during the period between signing and closing, and since closing, we’re actually taking orders, and the backlog I’ve mentioned is a good indication of that. The demand set seems robust. The list of data centers that people want connectivity to is long. It is fortunate that we’re already in those facilities but are not yet wave-enabled in those facilities. So there’s kind of two steps. One, can we even sell a wave in the facility? And then two, can we provision it in a much more expeditious way? We are working on both fronts. In terms of the demand set, we’re seeing it from some of our larger content companies, hyper-scale type customers. We’re seeing it from a broader set of content players, and a number of regional and international access network operators. And we are seeing a small, but growing set of AI-centric businesses that had traditionally not been wavelength buyers approaching us to buy wavelengths in some of these data centers. So we’re seeing four discrete buckets to three net-centric buckets I’ve just spoken about, and then finally some enterprise customers, who historically had bought waves from Sprint and other enterprise customers who are constructing private networks that are independent of the Internet. A wavelength is more expensive than transit on a per-bit utilized basis. However, it has the three positive attributes of being able to support large file transfers, having predefined latency, and complete diversity from the Internet for greater security. Those are the reasons why companies will pay a premium per bit mile for a wavelength.

I can do.

Our next question comes from Nick Del Deo from MoffettNathanson. Your line is now open.

Hi, good morning, guys. I didn’t catch all the Tad’s comments regarding the TSA change, or at least how that’s recorded. It looks like the balance sheet value in terms of what you owe T-Mobile went down quite a bit. Just to review, is that just a function of getting off their platforms faster? And I think you noted the drop their billing system effective now. What’s left to do on that front?

Yeah, the TSA is entirely associated with paying vendors. So initially they were paying 100% of the vendors we assumed with the Wireline Business on our behalf, and that has been winding down. In the second quarter, we were billed for May and June, however, those payments were not due yet. So we made no payments in the second quarter. That resulted in those charges being cash flow from operating activities of about $118 million. In this quarter, we have made three monthly payments, similar to the IP Transit, but on the cost side instead of the revenue side. So that’s why you see the large swing in that line item if you look at the cash flow statement. At the end of the corner, we still owed $69 million under the TSA agreement, which is essentially 2 months of activity. As we sit here now, we have migrated most of the vendors to our accounts payable systems, and we anticipate having all of the vendors migrated by the end of the year. The long pole and that 10 [ph] has no one would be surprised are some of the circuit vendors, which just take longer based upon their nature to get those migrated.

Right. And then complex to that T-Mobile had consolidated its circuit spends for its wireless business in many instances with the Wireline Business. So not only do we have to migrate that vendor, we have to segregate, the portion of the bill that’s attributable to the business we acquired. And T-Mobile’s been very cooperative in helping us do that, but it is an arduous task with hundreds of vendors. I mean, there are a couple of dozen that really matter in terms of scale. But, I’m actually very pleased that we’re as far along in transitioning our accounts payable into our own systems and being able to pretty confidently say, we’ll have virtually all of them within Cogent systems by year end, and equally impressive, I think, is the fact that we have migrated the billing platform to our platform and will be billing with Cogent bills going forward, actually shutting down the acquired platform that Sprint is used. Just as a point of reference, we still actually get off-net circuit bills from Verizon that say MCI on them. And it’s been probably over 15 years since Verizon acquired MCI. The bill does not say Verizon. It says MCI.

That’s fine. But, again, good work. I’m going to get your system grew over that quickly. Sort of separate question, on the data center front, you converted a few more facilities this quarter, because are you finding that the broader supply pinch in certain data center markets is increasing interest in your data center facilities, whether it’s old Cogent ones or the ones that you’re converting from Sprint? Or is there size and locations and other attributes, power densities, et cetera., I mean they might not necessarily be benefiting from that new supply dynamic?

Yeah, so I’m going to segregate the two groups of data centers, because they are substantially different. The Cogent data centers are all in leased facilities and tend to be smaller both in size and power. Today, we have in the Cogent footprint, including the facilities that we’ve converted, 109 megawatts of power available in those facilities. We still have 41 facilities to convert and we anticipate about another 100 megawatts coming from those facilities. So we concentrated on the facilities that had the biggest footprint of rack space and power. The second challenge we’ve had is that these facilities were occupied by unused telecom equipment. As I had mentioned on our previous call, there were over 22,500 bays of equipment that are not in service, but physically sitting on the floor. We’re removing those right now at the pace of about 400 a week across the footprint. I mean, that’s going to take a year. We’re trying to accelerate that. We have a footprint today that will support a little over 40,000 bays of equipment. So the challenge has been not the interest in our data centers, but really not disappointing customers in making promises that we can’t keep in terms of when these facilities will have vacant space and power that customers can use. The demand set has been strong. There is, at least for now, a short-term crunch in power availability. I think the locations of our facilities are of interest to companies, who are trying to have a more decentralized component to their data center model. So for the largest consumers of data centers space, they currently have kind of a two-tier hierarchy. They’re around proprietary, very large-purpose built facilities, and then a footprint and carrier neutral facilities for connectivity to the greater internet. Most of those customers are looking to add a third-tier, which would be something between those two extremes, and our data centers are very interesting to them. We are in discussions around wholesale capacity in our centers, but again, we’re just not in a position to sell at this point.

Okay. Great. Thanks, Dave.

Our next question comes from Brandon Nispel from KeyBanc Capital Markets. Your line is now open.

Great. Thanks for taking the questions. Did you just go through the corporate, net-centric and enterprise connection that adds this quarter, excluding the SIP impact? That I was hoping you could give us what the revenue impact of the SIP shutdown could be, because it did sound like was a September 30th shutdown. And then can you help us sort of bridge from 3Q to 4Q from a revenue standpoint across the three customer segments? And we talked already about the 11,000 non-core connections that you still have. How much of that will be end-of-life at the end of next quarter would be helpful to get the trajectory of revenue, right? Thanks.

Yeah, so I’ll start by taking those in reverse order, Brandon. The first one is the 11,000 non-core connections that we have are across 23 product categories. They will go away much slower. SIP was the largest of these categories. It was the one that also had the most notice requirement, because it was regulated. At least one customer protested to the FCC that the 1-year notice that they received was inadequate and wanted an extension, which was not granted. But, out of the 19,000 non-core connections, we went after the biggest group first, the SIP, and then for the others, products are much more heterogeneous and the path will be much longer. I would suspect we’ll see a reduction every quarter, but it’s probably going to take until the end of 26 till these non-core products are completely eliminated. It is a major component of our cost savings.

I can repeat the customer connections for you. So the SIP connections were 8,486 at the end of last quarter. Those are all non-core connections. When you look at corporate, net-centric, and enterprise of the 8,486 connections, 5,006 were corporate, 1,088 were net-centric, and 2,392 were enterprise. And that’s the numbers at June 30, 2023, and obviously at 930 those numbers are zero.

Yes, so for the remaining 11,000 non-core connections, they are spread across all three customer types. I will say that there’s less net-centric even than there was for SIP. So, it’s probably roughly 80% of it is in the enterprise base and corporate base, and less than 20% of it would be in net-centric.

Did I just follow-up? Did you guys say that these products were end-of-life at the end of September, and so there’s a revenue impact in the fourth quarter? I think that’s what I’m trying to get at, because obviously this quarter, I mean trends were below expectations and it doesn’t make sense that it was the SIP in fact that that was shut down at the end of the quarter. So just trying to understand sort of where we should be going and looking for in terms of total revenue next quarter? So there isn’t sort of a big headline [in this stage.] [ph]

Yeah, I might disagree with that statement, but we ended up pushing these customers on almost a weekly basis reminding them this product was going away. So the stragglers were shut off at the end of the quarter. So the product is completely gone. It is end of life. But the revenue was declining ever since we acquired the business, actually it was declining under T-Mobile’s ownership, and it was a big contributor to why the revenue run rate at signing was $560 million and was down to about $485 million at closing 9 months later, a lot of it was runoff in these non-core products and in SIP, in particular. We, I think, pushed even more aggressively. So while the unit count was down materially, the revenue impact from these end-to-life products in the third quarter was not very material.

Got it. Can I just do one more of the thousand in wavelength in terms of backlogs? What percentage of those were signed during the quarter and then what’s the average sort of provisioning timeline on the thousand that you have in backlog? Thanks.

So I would say the majority of them were signed in the quarter. Some were actually signed in the previous quarter on Q2 right after closing. And an average is very misleading on provisioning times, because it’s dependent on the two data centers that need to be connected being wave enabled. It is why I tempered, my backlog comment on saying some of these may never install because people may get frustrated. They are not going to wait 14 months till we have all of the data centers wave enabled. We increased the number of wave enabled facilities by about 25% in the quarter. I think, we’ll see that pace accelerate. If you looked at an average, it’s probably more than 150 days, because you’ve got a subset that are in the kind of 50-day provisioning window, both ends are insights we can do today. Some are more like 120 days, meaning one site is in the shorter window and one site is in the longer window. And then we’ve got waves to sites that are not yet wave enabled. Now, we’ve told customers we’re working as quickly as we can. They’ve still wanted to sign orders, but some of those may not install, because I’m not convinced the customer’s going to wait 6, 7 months to get a wave. And there could be a subset of those that take that all.

All right. We don’t hear any responses from Brandon. So for now we’re going to move on to Phil Cusick from JPMorgan & Chase. Your line is now open.

Hey, Dave, this is Jerome [ph] on for Phil. A couple of follow-ups if I could. You mentioned that corporate grew this quarter. Could you quantify that for us? Has there really been any change in trajectory? Could you talk about customer conversations and corporate? Or are the weaker markets starting to come along? Now should we think about the potential for growth to pick up? Second, could you just talk about the level of SG&A in 3Q and how that should look in 4Q? Given some of the cost cutting that’s going on, how should we think about overall margins heading into the fourth quarter and into 2024? Thank you.

Yeah, so on corporate growth, I would say it was very similar to the growth rates that we’ve had in Q2. That’s kind of an underlying kind of same-store growth rate of about 1% year-over-year far less than the kind of 10% to 11% that we had long-term average. We are seeing slow, but consistent improvement in corporate buying cycles and expect that corporate growth to continue to improve. The non-core services in answering Brandon’s question that are heavily weighted towards corporate and enterprise will continue to decline, but probably not nearly as precipitously as they did this quarter. So you may see a low customer connection count, but you will see, I think, revenue growth probably continued to be positive. From this point forward since SIP was the largest of these non-core products. And then in terms of SG&A, as Tad mentioned, we had record low bad debt expense. I think we’re probably expecting that to revert back to historical norms. And then, we are continuing to continue to grow in headcount and expect to see some underlying improvement. Tad?

Yeah, so the Q3 run rate is about reflective of our current run rate as we exit at the quarter.

Okay. Thank you.

Our next question comes from Michael Rollins from Citi. Your line is now open.

Thanks, and good morning. Just a couple of questions going back to some of the comments from earlier in the call, and then just one on the business. So, first, in the same way you just a couple questions ago, recapped the breakdown of what happened with the SIP product? Could you do the same with the T-Mobile commercial services agreement in terms of just recapping and summarizing in total? What happened, which customer verticals that volume came out of and what’s left going forward for that? And then secondly, with respect to the reclass of the operating leases to the capital lease. If I’m doing the math right, it looks like it’s about 3 years in terms of the change in the cost of the lease versus how much you increase the balance sheets accounts by? And just curious if that’s something that once it expires, it goes away, or is this something that needs to get renewed? Like, how should we be thinking about what needs to happen for this lease after you get through the next few years of the balance that you’ve increased? And then I have an operating question I’ll follow up with.

That’s actually a very good question, Mike, and you did your arithmetic quickly. That lease ends at the end of 2026. It will not be renewed. It does not need to be replaced. It is completely uneconomic. And it is for an IRU that is not even fully in use and is totally redundant to fiber that we have today is something that Sprint signed almost 35 years ago and had CPIs. The end of the lease is our first – what we have the ability to exit it in about 3 years at the end of 2026, and we’ll exit it as quickly as possible. The lease is way out of market, as I indicated, and we indicated on the previous earnings call, there’s about $150 million of uneconomic value, and that means it’s about a lot of market, and it was a reduction in the gain in our purchase of the assets. It’s just something we had to take, and I’m sure, as T-Mobile would say, that’s why we’re paying them $700 million. It took some bad stuff, and that’s probably the single worst item. It clearly meets the test to be a capital lease. It is for fiber. And it’s something that will not be replaced.

I’ll take the CSA. So under the commercial services agreement with T-Mobile, and this is just like the regular customer, not under the IP Transit Services Agreement. So the revenue was $7.3 million in the second quarter and $8 million in the third quarter. The connections, this is all net-centric revenue, the connections were $8,028 at the end of the second quarter and $4,661 at the end of the third quarter. So the revenue was about the same, and the connections dropped about 50%.

And Mike, to give you a little more granularity, there are really two primary services that are not covered by the transit agreement. The first is co-location, these are T-Mobile bays or racks that are located in our facilities that we are removing at the request of T-Mobile. But there’s a long tail on those. And then the second are the VPN services, the Ethernet point-to-point services that are providing backhaul for T-Mobile through our network. And they are grooming those circuits as well. So we would expect the unit count to continue to come down. We also do expect the revenue to eventually come down. But I think what has been going on in the short-term has been grooming units more than a revenue focus.

And so, even though the volume fell significantly quarter-over-quarter, would you expect as you’re describing a more measured roll-down of this over the next few years?

I don’t have visibility that far out. That would really be a question you need to ask T-Mobile, because they can cancel these with 30-day notice. I do have visibility this quarter and it looks very similar to Q3. Q4 should be similar to Q3 in terms of revenue, but with fewer units.

And just moving to the operating side of the business and thanks for all that detail. You made a comment earlier in the discussion about how customers on the net-centric side are really pushing to much higher levels of switching? I think you mentioned 100-gig, 400-gig, and as customers are moving up to these higher port speeds. What does that mean in terms of volume being a lesser indicator of revenue, because customers are pushing more volume, they’re just doing it through fewer connections. And is this a short-term blip, where this type of grooming or optimization happens quickly on the volume side? Or do you see for whatever the reason that there might be an ongoing difference between the way the revenue in net-centric performs and the way volume performs, because more customers across more ports adopt these higher speeds?

Yeah. So we’ve been through two similar grooming exercises in the past 20 years. The primary driver of these exercises is to reduce cross-connect costs. So when you say volume, there are really two different volumes. Volume of connections may actually go down as people consolidate 100-gig ports into 400-gig ports, and the remnant of the 10-gig to 100-gig consolidation continues. And then the second volume measure is the number of bits flowing. So the average price per bit is declined for the 23 years, almost 23.5 years at Cogent, since we’ve been in business and will continue to decline. But the big volume growth has outpaced that, allowing us to achieve an average of about 9% net centric revenue growth in a flat addressable market. I think what will continue to happen over the next couple of years is we will see an acceleration in the 100- to 400-gig conversion, because it cuts your cross-connect costs by 75%. And the cost of those 400-gig portable optics has come down meaningfully in the past year and, today, I would say only a couple of percent of net-centric ports are at 400-gig. So we have a long way to go in this grooming cycle that will depress the number of connections or port sold, but will not depress revenue. We literally had the same exact phenomena, when we went from multiple 10-gigs to a lot fewer 100-gig interfaces, and I think that’s happening. We’re also seeing on the wavelength side, customers looking to take 400-gig wavelengths. Our network is equipped to be able to support that. We will be selling those. We have some of those in the funnel that I had described. And, I think customers increasingly are sensitive to cross-connect expense.

Thanks very much.

That concludes our question-and-answer session. I’d now like to hand back over to Mr. Dave Schaeffer, for any closing remarks.

Well, thank you all very much. We’ll be chatting as anyone has any follow-up questions. And we’ll talk soon. Thank you all very much. Take care. Bye-bye.