Cogent Communications Holdings Inc (CCOI) 2022 Q2 法說會逐字稿

Welcome to the Cogent Communications Holdings Second Quarter 2022 Earnings Conference Call. As a reminder, this conference is being recorded and it will be available for replay at www. cogentco.com. A transcript of this conference call will be posted on the same website when it becomes available. Cogent's summary of financial and operational results attached to its press release can be downloaded from the Cogent website. (Operator Instructions)

And now I would like to turn the call over to Mr. Dave Schaeffer, Chairman and Chief Executive Officer of Cogent Communications Holdings.

Thank you very much, and good afternoon to everyone. Welcome to our second quarter 2022 earnings conference call. I'm Dave Schaeffer, Cogent's CEO. And with me on this afternoon'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 that we present in a consistent manner each quarter.

Now for a quick overview of our results. In June, we issued a $450 million unsecured 2027 note, and we used a portion of that proceeds to redeem our EUR 350 million unsecured notes that were due in 2024. We received net proceeds from this offering of $71 million, and we extended the average maturity of our debt by extinguishing our 2024 euro notes. We obtained an economic gain of $26 million from the difference between the euro to dollar rate at the date of issuance as compared to the euro to dollar rate at the date of settlement.

We also eliminated a restrictive covenant provision in our euro-denominated notes that was 4.25 to 1x EBITDA. This restrictive covenant limited our ability to transfer funds from our operating company to a holding company, and therefore, be available for both dividends and share buybacks.

As of June 30, our cash held at Cogent Holdings was $66.6 million, and we had an additional $219.9 million that can now be transferred to holdings for a total of $286.5 million of unrestricted cash that's available for either dividends and/or buybacks. Cash held at our operating companies was $283.3 million, and our total consolidated cash and restricted cash was $349.8 million at end of quarter. Our gross leverage ratio was 5.22, and our net leverage ratio was 3.70. Our consolidated leverage ratio as calculated under our notes indenture was slightly lower at 5.21 at quarter's end.

Our total revenues and our NetCentric revenues were materially impacted by the negative foreign exchange movement during the quarter and a significant strengthening of the U.S. dollar, in particular, against the euro. For the quarter, we had sequential negative foreign exchange translation of $1.4 million and a negative year-over-year of $3.4 million. The negative impact of foreign exchange on our revenues for the quarter -- third quarter is expected to be even more significant.

Our total revenues and corporate revenues are materially impacted negatively also by the change in USF or Universal Service Fund revenues in the quarter due to the reduction in the USF rate by the U.S. government. For the quarter, the negative impact of USF on our revenues was $0.3 million and a negative of $1.4 million on a year-over-year basis. The combined negative impact to our sequential revenues of foreign exchange and USF was $1.7 million sequentially and $4.8 million negative impact on a year-over-year basis.

Our corporate business continues to be influenced by real estate activities in the central business districts of North American cities. Two key statistics, including the level of card swipes in buildings and leasing activities indicate that during the first half of 2022, the real estate market and leasing activities in the central business district in which we operate have seen some improvement, but have not yet returned to their pre-pandemic levels. Leasing activities across major markets and workers return to offices continue to improve, albeit slowly.

On a U.S. GAAP basis, our corporate revenues declined by 1.1%, which was partially attributable to the change in USF rates. Adjusting for this negative USF impact, our corporate revenue decline for the quarter was 0.007% sequentially.

We continue to remain cautious in our outlook for our corporate revenues given the uncertainty of the economic environment and the continued lingering effects of the pandemic.

Our NetCentric business continues to benefit from robust growth in video traffic and streaming. For the quarter, our traffic was up 3% sequentially and increased on a year-over-year basis by 19%. On a U.S. GAAP basis, our NetCentric revenues grew by 0.3% and grew on a year-over-year basis by 10.2%. Adjusting for foreign exchange on a constant currency basis, our NetCentric revenues increased by 16.2% on a year-over-year basis and 2.5% on a sequential basis.

On a U.S. GAAP basis, our second quarter total revenues sequentially declined by 0.5% to $48.5 million and increased on a year-over-year basis by 0.4%. On a constant currency basis, our quarterly revenues increased sequentially 0.4%, an increase on a year-over-year basis by 2.7%. On a constant currency basis and also adjusting for the change and negative impact in USF revenues, our sequential quarterly revenues would have grown 0.6% sequentially and 3.6% year-over-year, approximately equivalent to the rate of growth in the last quarter.

Our EBITDA margins this quarter increased to 39.4%, which represents the highest EBITDA margin in the company's history. We're also encouraged by the fact that our sales force productivity numbers increased to 4.9 installed orders per full-time equivalent rep per month, up from 4.7 last quarter. This is actually the best sales force productivity that we've seen since the fourth quarter of 2018.

During the quarter, we returned $41.9 million to our shareholders through our regular quarterly dividend. We did not purchase any stock during the quarter and have a total of $30.4 million available for buybacks under our program, which our Board has authorized to continue through year-end. Our Board of Directors reflected on our strong cash flow generating capabilities as well as investment opportunities that the company has and decided again to increase our quarterly dividend sequentially by $0.025 per share in the quarter, therefore, raising our quarterly dividend from $0.88 per share to $0.905 per share. This increase represents the 40th consecutive sequential increase in our regular quarterly dividends, and our annual dividend growth rate today is 12.4%.

Now for a few expectations against our long-term guidance targets. Our targets are meant to be long term. Our EBITDA, annual margin expansion is anticipated to be approximately 200 basis points a year. Our targeted multiyear constant currency growth rate should be approximately 10%. Our revenue and EBITDA targets are meant to be multiyear targets rather than looking at a specific quarter. And these targets assume that businesses will continue to return to their offices, albeit in a hybrid work model. This is not intended to be quarterly or annual guidance.

Now I'd like Tad read our safe harbor language and provide some additional details on our operations in the quarter. Following that, we'll open the floor for questions and answers.

Thank you, Dave, and good morning, everyone. This earnings conference call includes forward-looking statements. These forward-looking statements are based upon our current intent, belief and expectations. These forward-looking statements and all other statements that may be made on this call that are not historical facts are subject to a number of risks and uncertainties, and the 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. And if we use non-GAAP financial measures during this call, you'll find these reconciled to the corresponding GAAP measurement in our earnings releases, which are posted on our website at cogentco.com.

Like many companies, we continue to be impacted by the COVID-19 pandemic and the accompanying responses by governments around the world. Our risk related to COVID-19 and other risks are described in more detail in our annual report on Form 10-K for 2021 and on our quarterly reports on Form 10-Q.

Some discussion on corporate and NetCentric revenue and customer connections. As a reminder, we analyze our revenues based upon network connection type, which is on-net, off-net and noncore, and we also analyze our revenues based upon customer type.

We classify all of our customers into 2 types: NetCentric and corporate customers. Our corporate customers buy bandwidth from us in large multi-tenant office buildings or in carrier neutral data centers, and these customers are typically professional service firms, financial service firms and educational institutions located in multi-tenant office buildings or they connect to our network through our carrier-neutral data center footprint.

Our NetCentric customers buy significant amounts of bandwidth from us in carrier-neutral data centers and include streaming companies and content distribution service providers as well as access networks who serve the consumers of content.

Our corporate business represented 57.4% of our revenues this quarter. And as Dave mentioned, our corporate revenue declined year-over-year by 5.9% and was $85.2 million and declined sequentially by 1.1%. A decrease in the USF tax rate, which only applies to our corporate VPN connections, had a $0.3 million negative impact on our sequential corporate revenue growth and a $1.4 million negative impact year-over-year.

The USF tax rate changes quarterly and we cannot predict the impact of future USF tax rates on our revenues. We had 45,103 corporate connections -- customer connections on our network at quarter end. That was a sequential decline of 0.6% and a year-over-year decline of 1.5%.

Our NetCentric business, which represented 42.6% of our revenues in the quarter, and despite very material FX headwinds, had another solid quarter and grew by 0.3% to $63.3 million sequentially and year-over-year by 10.2%. The volatility in foreign exchange rates primarily impacts our NetCentric revenue and the impact was materially negative both sequentially and year-over-year. On a constant currency basis, our NetCentric revenue increased year-over-year by 16.2% and sequentially by 2.5%. We had 50,674 NetCentric customer connections on our network at quarter end, a sequential increase of 2.4% and a 10% increase year-over-year.

Comments on revenue by customer network type. Our on-net revenue was $112 million for the quarter, which was a sequential decrease of 0.6% and a year-over-year increase of 0.8%. Our on-net customer connections increased by 0.8% sequentially to 82,277 and increased by 4% year-over-year. And we serve these on-net customer connections on our network in our 3,095 total on-net multi-tenant office and carrier-neutral data center buildings. Our off-net revenue was $36.3 million for the quarter. That was a sequential decrease of 0.3% and a year-over-year decrease of 1.1%.

Our off-net revenues are impacted by incorporating the cost savings we obtain from lower local loop prices into our pricing. The introduction of these customers into our off-net revenue base lowers our off-net ARPU and impacts our revenues -- off-net revenues. Our off-net customer connections increased sequentially by 1.8% to 13,160 and the year-over-year growth was 6.2%, which was an increase from last quarter. We ended the quarter serving our off-net customer connections in about 8,000 off-net buildings, and these off-net buildings are primarily located in North America.

Some comments on pricing. Consistent with our long-term historical trends, our average price per megabit of our installed base and our new customer contracts decreased for the quarter. The average price per megabit for our installed base declined sequentially by 6.2% to $0.29 and year-over-year by 18.7%, and that year-over-year rate was better than our historical rate of decline of 21.5%. The average price per megabit for our new customer contracts for the quarter decreased sequentially by 15% to $0.15 and year-over-year by 15.9%.

We continued to succeed in selling larger 100 gigabit connections and 400 gigabit connections in selected locations to our customers. Selling more of these larger connections results in a change to our connection mix and has the effect of lowering our average price per megabit at a greater rate than it impacts changes in our ARPU.

Regarding ARPU. Our on-net ARPU decreased primarily from the negative impact of foreign exchange and the negative impact of USF on our revenues. Our off-net ARPU continued to decline from the lower pricing we obtained from our off-net circuit vendors. And again, we passed that savings on to our off-net customers. Our on-net ARPU, which includes both corporate and NetCentric revenues, declined sequentially by 1.5% and was $455. Our off-net ARPU, which is predominantly corporate customers, declined sequentially by 2.2% to $927.

Comments on churn. Our sequential quarterly churn rates for on-net and off-net connections were relatively stable. Our on-net unit churn rate was 1% this quarter compared to 0.9% last quarter, and our off-net churn rate was 1.1% compared to 1% last quarter. In order to reduce our customer turnover, we employ a dedicated sales group that works to retain customers who have indicated that they are considering terminating their service with us. We may offer pricing discounts to these customers in order to induce them to reverse their termination decision or to purchase additional service from us and/or to extend the term of their contracts with us.

During the quarter, certain of our NetCentric customers took advantage of our volume and contract term discounts and entered into long-term contracts with us for over 2,460 customer connections, and that increased their revenue commitment to us by $23.3 million.

Some comments on EBITDA and EBITDA margin. We reconcile our EBITDA to our cash flow from operations in each of our quarterly press releases. Seasonal factors that typically impact our EBITDA and our SG&A expenses include the resetting of payroll taxes in the United States at the beginning of the year, annual cost of living or CPI increases, seasonal vacation periods, the timing and level of our audit and tax services and our annual benefit plan cost increases.

Our EBITDA increased sequentially by $1.3 million and also $1.3 million increase year-over-year. And the impact of foreign exchange negatively impacted our EBITDA dollar growth year-over-year by $1.3 million. So absent that, it would have been $2.6 million year-over-year. Our quarterly EBITDA margin increased sequentially by 110 basis points to 39.4%, an increase year-over-year by 70 basis points.

Some comments on earnings per share. Our basic and diluted income per share was $0.24 for the quarter. Foreign exchange gains and losses on the translation of our euro notes into dollars until we extinguished them on June 30 of this quarter. Losses on the extinguishment of debt and the noncash changes in the valuation of our interest rate swap agreement have been the primary contributors to the variability in our net income, and consequently, our per share results. Our foreign exchange gain on our euro notes before they were retired was $23.5 million this quarter. It was $8 million last quarter and was a loss of $5.3 million in the second quarter of last year.

We incurred a noncash charge of $7.5 million in this quarter related to the increase in the estimated fair value of our interest rate swap agreement, and that increase was $21.3 million last quarter. We incurred a loss on the extinguishment of our euro notes of $11.9 million in this quarter, and there was a loss of $10.8 million when we extinguished our '22 notes -- 2022 notes in the second quarter of last year.

Combining all these amounts, that increased our net income in this quarter by $4.2 million, decreased our net income last quarter by $13.3 million and decreased our net income in the second quarter of last year by $16.1 million.

Further comments on foreign exchange. Our revenue earned outside of the United States is reported in U.S. dollars and was approximately 25% of our total quarterly revenues. About 16% of our revenues in this quarter were based in Europe and 8% of our revenue is related to our Canadian, Mexican, Asia Pacific, South American and African operations.

As we experienced in this quarter, volatility in foreign currency exchange rates can materially impact our quarterly reported revenue and EBITDA results, impacts the valuation of our euro notes until we extinguish those notes in this quarter and also impacts our overall financial results as we translate foreign currency exchange amounts into U.S. dollars. The foreign exchange impact on our revenue in this quarter was materially negative and it's expected to be again materially negative for the third quarter.

The average euro to U.S. dollar rate so far in this quarter is $1.02 and the average Canadian dollar exchange rate is $0.77. If these average rates remain at their current levels for the remainder of our third quarter, we estimate that the FX conversion impact on our sequential quarterly revenues for the third quarter would be a negative $1.1 million and the year-over-year FX impact would be a negative $3.8 million.

We believe that our revenue and customer base is not highly concentrated, and our top 25 customers represented about 6% of our revenues this quarter, consistent with many quarters in the past.

Our quarterly capital expenditures decreased sequentially by 4.6% to $17.3 million. Supply chain uncertainties is causing us to shift our typical purchasing schedule for network equipment, and these anticipatory investments are designed to ensure that we have satisfactory inventory levels of network equipment to accommodate our growth plans and customer needs.

Some comments on our finance leases and IRU 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 after the initial term, include multiple renewal options. Our IRU finance lease obligations totaled $254.2 million at quarter end. We have a very diverse set of IRU suppliers, and we have IRU contracts with a total of 301 different dark fiber suppliers. Our finance lease principal payments decreased sequentially by 10.7% to $5.2 million.

Some comments on cash and operating cash flow. At quarter end, our cash and cash equivalents and restricted cash totaled $349.8 million, and our $37.8 million of restricted cash is tied to the estimated fair value of our interest rate swap agreement. Our cash flow from operations this quarter was $34.4 million, which was a decrease of $15 million from last quarter. But included in this quarter and part of that sequential decline was $16.8 million of semiannual interest payments on our note obligations that we made this quarter, with no corresponding amount last quarter.

Comments on debt and ratios. Our total gross debt at par, including our finance IRU lease obligations, reflecting the extinguishment of our euro notes and the issuance of our new $450 million 2027 notes was $1.2 billion at quarter end, and our net debt was $854.4 million. Our total gross debt to trailing last 12 months EBITDA as adjusted ratio was 5.22% and our net debt ratio was 3.7%. And the ratio, as calculated under our denture was slightly less, at 5.21.

Further comments on the swap. We're party to an interest rate swap agreement that modifies our fixed-rate obligation associated with our $500 million of 2022 notes to a variable interest rate obligation that's based on the secured overnight financing rate or SOFR. We record the estimated fair value of the swap agreement at each reporting period, and we occur corresponding noncash gains or losses due to the changes in market interest rates.

At quarter end, the fair value of the swap agreement increased by $7.5 million from last quarter to a net liability of $37.8 million. Under the swap, we're required to maintain a restricted cash balance with the counterparty equal to the net liability. The settlement payments under our swap agreement are made each November and May.

Under our initial settlement payment that was made in November of last year, we achieved a net interest savings of $0.6 million. That was the period from the swap inception date in August of last year to the end of October. Under the settlement payment made this May, May 2022, we achieved a net cash savings of $1.2 million for the period from November 1 through April 30. So we've achieved a total cash savings on our swap agreement since inception of $1.8 million.

Lastly, some comments on bad debt and DSOs. Our bad debt expense was only 0.4% of our revenues for the quarter compared to 0.2% last quarter and 0.5% in the second quarter of last year. Our days sales outstanding, or DSO, for worldwide accounts receivable was 22 days for the quarter. We want to again thank and recognize our worldwide billing and collection team members for continuing to do a fantastic job serving our customers and collecting from our customers.

And now I will turn the call back over to Dave.

Thanks, Tad. I'd like to highlight some of the strengths of our network, our customer base and our sales force. We achieved excellent revenue growth in our NetCentric business. On a year-over-year, NetCentric revenue growth was 10.2% and 16.2% on a constant currency basis. We continued to be a direct beneficiary of the increase in over-the-top video and streaming traffic, particularly in international markets.

At quarter's end, we were connected to 1,409 carrier-neutral data centers and 53 Cogent data centers, more than any other carrier as measured by third-party research. The breadth of our coverage enables our NetCentric customers to optimize our networks and reduce latency. We expect to continue to widen our lead in this market as we plan to connect an additional 100 carrier-neutral data centers to our network each year for the next several years.

At quarter's end, we were connected directly to 7,685 autonomous systems or unique networks that comprise the Internet. This collection of ISPs, telephone companies, cable companies, mobile operators and regional carriers provide us access to the vast majority of the world's broadband subscribers and mobile phone users.

At quarter's end, we had a dedicated sales force of 197 professionals focused solely on the NetCentric market. We believe this group of professionals is the largest group of salespeople focused on this market segment.

Now for a few comments on our corporate business. We're seeing positive trends in our corporate business. But in a work-from-home environment that is becoming established as part of people's normal work routine, we believe corporate customers will continue to upgrade their Internet infrastructure, particularly supporting larger connections to improve the user experience of their remote workers.

Our corporate customers are aggressively integrating some of these new applications to become part of their working environment, such as including video conferencing in all aspects of employees' jobs. This usage will require high bandwidth connections both inside and outside of the premise. Our aggressive push towards lowering the cost of bandwidth and providing greater coverage has begun to boost corporate demand for our robust bidirectional and symmetric 1 gig and now 10 gig connections.

Our corporate customers are also increasing their purchases of redundant connections in carrier-neutral data centers to help support the ad hoc virtual private networks that their remote workers need.

Now for a few comments on our sales force. We continue to experience improvement in our sales force productivity due to our training as well as managing out underperformers. On a sequential basis, our total pre head count did slightly decline from 479 reps to 477, and our full-time equivalents declined from 453 to 449. Both of these declines were more moderate than we've seen in other periods during the pandemic. On a year-over-year basis, our total sales rep head count decreased by 88 and our full-time equivalent rep in total decreased by 62.

Our sales force was working remotely for the majority of the first quarter and only returned to a Cogent sales office on March 1. Our sales force turnover numbers are improving. They were 5.9% per month. That is a significant reduction from the 6.9% per month in the previous quarter and a peak during the pandemic of a turnover rate of 8.7 reps percentage of the sales force per month at the peak. The factors that contribute to the increase in our sales force productivity have been better training and in-office attendance. This 4.9 is a material improvement from the 4.7 we experienced in Q1.

So in summary, we remain optimistic about our unique position in serving small and medium businesses in the central business districts of major cities. We have a total of 1,826 multi-tenant office buildings directly connected to our network, supporting approximately 1 billion square feet of rentable office space. We remain concerned about the slower pace of office leasing as well as tenants returning, but we're seeing a number of key indicators that are showing signs of improvement whether it be workplace reentry or declines in the rate of increase in office vacancy and a heightened level of new lease signatures.

While these remain below pre-pandemic levels, they are substantially improved over the past 18 months. We're encouraged that many tenants are indicating a return to office in 2022, and therefore, the commercial office leasing market should continue to improve.

We are optimistic with the combination of our sales teams being in office and the return to a more normalized environment for our customers. We believe this will benefit our sales organization and allow us to sell tenants who have deferred or delayed network upgrade decisions.

Many companies are establishing long-term network architectures that incorporate a hybrid work policy. This is a positive for Cogent. As certain corporations elect to downsize their office space requirements in multi-tenanted office buildings, this actually provides an increased addressable market opportunity as those spaces begin to be re-tenanted with new organizations. We believe the benefit from this opportunity will allow us to sell more services to more tenants as landlords aggressively strive to fill the vacant office space in U.S. and Canada.

Our targeted multiyear constant currency growth rate remains 10% and our long-term EBITDA margin expansion target is 200 basis points. This is in line with our long-term historical averages. But during the pandemic, we've been operating at about half of that level. We expect to see a continued but, unfortunately, a gradual improvement to these long-term targets.

We have extended the average maturity of our debt, eliminated restrictive covenants that were included in our euro notes. And as of June 30, we have substantially more cash available at holdings of a total of $286.5 million that is available to benefit our shareholders.

While we did not repurchase any stock in the quarter and have $30.4 million remaining, we felt that the decision the Board took to increase our dividend for yet a 40th consecutive quarter to $0.905 per share for the quarter, representing a 12.4% growth rate in our dividend is consistent with our belief that our business is continuing to grow. Our consistent dividend increases demonstrate the long-term optimism we have in our business.

Now I'd like to open the floor for questions.

(Operator Instructions) Our first question comes from the line of Gregory Williams with Cowen.

First one is just on corporate. Dave, you've noted in the past that you'd want to get to the 2%, 2.5% sequential growth target, the historical target in the next few quarters. Has that changed since you're now noting a little bit of concern about office leasing trends?

And the second question is just on your appetite for debt and increasing that. So I understand with this debt raise you did here, while it's nice to be paying in euros. And you now refi-ed it at a much higher rate. I understand that you're lifting some of those restrictive debt covenants. With those debt covenants removed or at least a lot more flexible, I was hoping you'd give us sort of a new range or a new target of how high you'd be comfortable taking your debt levels up now that you're at 2.7x?

Yes. So first of all, let me touch on the corporate growth rate. We have seen, I think, 3 factors impact corporate activity in central business district. At the beginning of the pandemic, we saw many businesses shutter and just go out of business. And that increased the vacancy rate in those buildings and had a material negative impact on our corporate growth rate.

The second thing that then occurred is for the remaining tenants in those buildings, they were unwilling to make architectural decisions around what their networks would look like post-pandemic in part because they had no visibility to the end of the pandemic. The third factor, which has been probably the most difficult to fully understand, has been the fact that some tenants and customers at the end of their leases are electing to continue to vacate properties. So while the increase in vacancy has abated, we have not yet seen an inversion and a reduction in vacancies.

I think for us to get back to a 2.5% sequential growth rate in corporate, we will have to see net absorption of central business district vacancy rates and tenants reoccupying sign those leases. We are much less focused on the dollar value of those leases or the duration. Those are very important factors for landlords.

While we have seen a slowdown in the rate of degradation, we have not yet seen that net absorption number turn positive. Because of that, while we were hoping that in the first -- the second half of this year we would get back to that positive rate, we don't have complete visibility to that. We are seeing signs of improvement. We appear to be kind of at a plateau with declines in corporate being similar to last quarter, but not yet improving. So I think at minimum, we're still several quarters away.

We are encouraged by the leading indicators, which are tours, card swipes and some leases, but there is still more work to be done to get the buildings that we serve back to their historical occupancy rate.

Pivoting over now to your debt question. We have a stated net leverage range that includes all that, inclusive of our capital leases of 2.5 to 3.5x EBITDA. Today, we're over that range at 3.7. We don't feel that, that is an alarming number at this point. It's a big part of the consideration that the Board had in making its decision to yet again raise the dividend by $0.025.

We also understand that over a long period, the growth in EBITDA has to match the growth in the dividend. We have historically grown EBITDA faster than we've grown the dividend. And since the pandemic, it has grown slower. We continue to monitor the situation. We are committed to our dividend growth policy. And we think while we'll be above that 3.5 for several more quarters, the Board is comfortable with the level we're at and has not yet changed an official range. And we do think we'll revert back into that 3.5 or below number over the next year or so.

Our next question is by Frank Louthan with Raymond James.

It's Rob on for Frank. So obviously, Dave, you just talked to the outlook for corporate with respect to the rate of vacancies and then re-absorption beyond that. Just pivoting to Internet traffic growth, what did traffic growth look like during the quarter? You might have addressed this earlier. And is there any reason that Internet traffic volumes might potentially be able to return to where they were like about a year or 2 ago? Like what's the outlook on traffic growth going forward?

We're actually very encouraged by the traffic growth that we saw. Traditionally, the summer months on a sequential basis are slower than the winter months. We grew sequentially in the second quarter by 3%, and our year-over-year growth actually accelerated to 19%. These are very positive signs.

We're seeing that growth come mostly from international markets. We saw our NetCentric revenues were a direct result of this growth, grew at 10.2% year-over-year on a stated basis and 16.2% on a constant currency basis. So far above the long-term average growth rate of 9%.

I know that many investors were concerned because some of our content delivery customers had called out slower traffic growth as part of the reason why their businesses were not performing as well as investors expected. We have over 200 content delivery networks and thousands of proprietary networks continuing to use our network. So we feel pretty good and have been actually pleasantly surprised to the upside for the past several quarters on how well our NetCentric business has performed and we actually expect this outperformance to continue at least for the foreseeable future.

Our next speaker -- our next question and answer is from Nic Del Deo with MoffettNathanson.

Dave, a couple of questions ago, you noted that you think you need to see vacancy rates decline to get corporate growth back to normal. If I think back to when your corporate unit was growing at much faster rates, vacancy rates bounced around and I don't think that was ever really thought of as a key driver of the business. I think the driver was improving your share of connections in the building you serve.

So the way you tied corporate growth to vacancy trends almost makes it sound like share gains aren't part of the story. I mean, shouldn't you be able to grow the corporate business even if vacancy rates stay flat via share gains? And maybe just help me understand that disconnect or if I'm interpreting something incorrectly?

So share gains are our primary way of growing our corporate business. You are absolutely correct. However, our addressable market shrank by over 10% when the average vacancy rate went from 6% to 18%, where it is today nationwide in CBDs. Some cities actually have CBD vacancy rates of nearly 40%. Those are probably the most extreme. Downtown San Francisco is probably the worst example.

But I think with that level of vacancy in the building, many tenants are just reluctant to make a switch. They don't know if they're going to stay or not, are they're going to move to a different building, are they going to downsize. And I think we just need to see a stable environment and one that's maybe a little more normal than what we've historically seen.

I do believe that most of Cogent's growth will come from capturing incremental customers in the building. But we've always stated that there are 3 windows in which we can win customers. That has not changed. When a customer moves into the building, when the customer's IT infrastructure is re-architected or when a customer experiences a failure from their incumbent provider. These 3 factors remain the key reasons why people will consider a new Internet provider.

And I think with a lack of new tenants coming in and these vacancy rates, a lot of tenants are just reluctant to make that architectural decision. Things have improved. They are definitely not where they were during the pandemic at its peak. But we're still not back to normal. And I'm not trying to make excuses, but the reality is our growth rate in corporate is materially below our long-term average. And I think it's because some of the existing tenants are also not willing to commit, figuring they may take advantage of this softer market, lower their rental rates either by renegotiating where they're at or migrating to another building.

The market for office space has dramatically changed during the pandemic, going from a seller's market to a buyer's market. And I think because of that, many tenants are reluctant to take a multiyear commitment for new bandwidth.

Okay. Okay. That's helpful. On the NetCentric side, obviously, you've talked about NetCentric returning to a more normal pace for some time. So it's not surprising that it's tapering off. If I look at the sequential constant currency growth rate, it did step down quite a bit from what we've seen over last 1.5 years. Obviously, that's volatile. It's hard to forecast. Should we think of this as a more normal sequential growth rate going forward? Or is this an anomaly and you think it could still kind of bounce up and down and take longer to taper to your target growth rate?

Well, I think it will definitely have some sequential volatility. And typically, we see a pickup in traffic, and therefore, corresponding NetCentric revenue in the latter part of the third quarter, throughout the fourth quarter and throughout the first quarter and then slowing down again in the second quarter. And that's been a historic trend both pre-pandemic and during the pandemic.

We've said that over time, we expect our NetCentric growth to converge to about a 9% annualized rate. We're obviously at least 50% above that today. It's a little hard for us to tell how much more of the internationalization of streaming will continue because many of the markets are still relatively nascent. We've seen, I think, a level of maturity in the U.S. and Canada for streaming. That doesn't mean it's not growing. It just means the growth rate has slowed materially.

That same trend has not happened in the developing world. And if it continues to follow the same trajectory as the U.S. and Canada did, we'd probably have 3 or 4 years of elevated growth. But again, we just don't have enough data to be comfortable to tell investors "expect 15%, not 9% growth for the next 3 to 4 years."

Our next question comes from the line of Timothy Horan with Oppenheimer.

So Dave, customers are kind of frozen. They're not really upgrading or changing their networks if they might move. But they're not really moving yet. So churn is relatively low. And new tenants aren't really coming in yet. Do you think that recession kind of extends out this frozen period, maybe this will last 2 or 3 years more or another 12, 18 months? Or what gives you the confidence with the dividend increase that this will reverse in a few more quarters?

Yes. So 2 different points. So the first one is, in previous recessions, with the exception of the financial crisis of '08, '09, we were not materially impacted in terms of our corporate growth rate. What we typically see is tenants use recessions to upgrade their offices at the same or lower rent by migrating from B and C buildings into A buildings. And we've experienced that in our customer base. If you look at third-party data, whether it's JLL, CoStar, Cushman & Wakefield, for previous recessions, that's always been the case.

Now if we have a recession, it is most likely not going to be as impactful as 2008, 2009. And for that reason, we don't believe that the rate of economic growth is really material to our corporate growth rate. It's really the aggressiveness of the landlords to lower the rents in the most expensive buildings to reduce the vacancy rates. And they typically have had more flexibility in doing that in previous recessions. We think that will continue.

We also think that at the lower end of the market, we're continuing to see supply be converted to residential. So I think part of the reason why, even though the world has migrated mostly to a hybrid work environment, that we're seeing the rate of vacancies plateau. That doesn't mean it's a V. It means it looks more like a U. But it has at least stabilized.

In terms of the dividend, we feel very comfortable in raising $0.025 today. We've got the outperformance of our NetCentric business. We have the expansion in margins and the best EBITDA performance in the company's history. The percentage of our business that is NetCentric has kicked up throughout the pandemic from roughly 40% of total revenues going into the pandemic to today being approximately 47% of revenues.

So as long as that increases as a share and it continues to increase at an outpaced rate, we feel comfortable in our ability to generate more cash. As I stated earlier, it's absolutely necessary that our dividend growth rate and our EBITDA growth rate eventually converge. We think that with the high percentage of on-net in our NetCentric business, we'll continue to see further margin expansion and acceleration in EBITDA growth even without a material reacceleration in corporate on-net. We do think corporate on-net and off-net for corporate customers will improve, but we also are realistic that the rate of improvement is not as aggressive as we would have thought even 2 or 3 quarters ago.

Very good. Just 2 other maybe brief. Are you seeing any change in pricing in the corporate side? Inflation has ticked up a lot and the prices have been down a lot. Have you seen any change in trajectory there?

So we live in the industry that inflation forgot. We are in the business of selling a deflationary service. As Tad mentioned, the rate of decline in the average price per megabit actually did moderate some. I think part of that is the fact that more of those megabits sold have tended to be in more expensive markets outside of the developed world, which I think has distorted that a little bit.

We are a huge beneficiary of substantial improvements in technology, both wave division multiplexing and optically interfaced routers. We think those trend lines will continue for the foreseeable future. Our equipment vendors support that prognosis. So we feel pretty good.

On the operation side, clearly, our expenses are rising just like all other businesses. We have been able to expand margins even though we have passed on some of those inflationary pressures and been able to increase our employee salaries, remain competitive in the market, and obviously, take advantage of market conditions for office rents, for other services that we buy.

We are also fortunate that most of our power purchases are fixed and are not susceptible to wild market fluctuations. We have had some increases in power, and we've been able to pass those on to our customers. So all in all, we think that inflation will not impact our ability to grow our EBITDA.

Our next question comes from the line of Walter Piecyk with LightShed Partners.

Dave, you mentioned in the prepared comments -- you mentioned in the prepared comments about some long-term deals on NetCentric. Typically, when those companies enter into long-term deals, they're asking for price reductions. But that also gives you some good visibility in terms of the growth of that line. Can you give us some kind of qualitative content -- or comments, excuse me, on the kind of puts and takes of those long-term deals?

Yes. Sure, Walt. And by the way, you can go back to transcripts for the past decade, and we have that same section in every one of them and we quantitatively give how many customers and what their contract value increase was.

So a NetCentric customer is accustomed to buying services on a metered basis. We offer discounts for fixed commitment and fixed term. So what a customer will typically do is buy a number of ports. They will typically commit to 50%, 60% total port capacity, and they will be running at a (technical difficulty) level. As time goes on, they may either add additional ports or they may start to run over their committed capacity, in which case they then pay a premium, a higher price per megabit.

Customers understand that market prices have declined at about 23% per year for like volumes. Because most NetCentric customers enter into a 3-year contract to get the lowest initial price, usually about halfway through that contract the customer is experiencing 2 things. One, they haven't gotten a reduction in price; and two, they're utilizing more than their committed capacity and are paying the penalty rate on that first traffic.

What they will typically do at that point is come back to us, say, "We'll sign a new 3-year contract," recast the 18 months that are left to 36 months, will take more ports, more total commitment and a lower price. We expect that to happen routinely with NetCentric customers, and we then factor that into their next 3-year contract commitment. That is the most common pattern for these repricings.

I would just add. For us to accept that new commitment, so a customer is pricing -- asking for a change in term, the total contract value for the new contract must be greater than the remainder on the original contract for us to accept that.

And there's a second governor, which is the salesperson only gets a commission if there is an increase in the monthly spend. So there's kind of both an internal control mechanism and a customer control mechanism.

I want to also go back to Nick Del Deo's question because I think your response to that was a bit more cautious than the kind of outlook that you provided really since this pandemic started. And then I want to kind of, I guess, put that in the context of long term, you're expecting EBITDA growth and -- well, not long term. You said at some point. I don't know what the time frame is, but I think you were non-specific about it. But you're going to converge the EBITDA growth rate with the dividend growth rate, which obviously is underpacing at the moment, as you already highlighted.

I think you've gotten some stuff wrong in the past. Obviously, everyone gets stuff wrong in terms of understanding when things don't go as we expect them to. But given how cautious you were to Nick, if we don't really know the time frame of when those can converge, then it seems like there's a possibility that you can't get leverage down below 3.5.

So let's put that aside. You're willing to stick this out. At what point do you not stick it out? Like is 4.0 the benchmark? Or the bogey is 4.5? At what point does the Board say, "You know what, we're tired of waiting for this convergence." And 4.0 or 4.2 or 4.5 or 5, like 7, like there's got to be a number. Are you willing to take leverage to 9x?

So 3 parts to your question, 3 answers. First of all, I do not have a crystal ball for return to normality. When the pandemic started, if you would ask me -- I thought everybody was going to go home for 3 or 4 weeks and we were going to come back and everything would be just like it was before the pandemic. Now 2.25 years later, we're still living in a different world, which has experienced supply chain issues, price distortions and a significant change in the employer-employee dynamic.

I think the pendulum swung from employers having all the power to employees having much more power as a result of the pandemic. And as a result of that swing, a hybrid work environment has become much more common place.

Now again, I can't predict is that a permanent change or a transitory change. What I do know is all of these have taken longer to revert back to historical norms than we had expected. As a result, I have been very clear: we expect our corporate business has bottomed out. I think we've proven that. I think our expectations were correct. On the previous earnings call, I was pretty clear, that I said I thought we were 2 to 3 quarters away from return to normality. While another quarter has passed, we've got additional data points -- and while we see things improving as leading indicators, the results are about the same.

So as a result of that, I think the pace of return to a normal work environment is slower than what I would have predicted. Because of that -- and really a tie in to next question -- the reality is we need customers to feel they can make architectural decisions. We absolutely win when we're in front of a prospect. We win 50% of our on-net proposals that we issue. But you got to get the customer to be willing to focus on it and know what their architecture is going to be.

And I could just go back to kind of the lineage of your firm and how you've transitioned through the pandemic. And maybe you'll transition back to an office, maybe you won't. And you play that out across literally tens of thousands, hundreds of thousands of tenants in our buildings, and we have seen a more cautious environment.

Now we are winning business. We are gaining customers and our sales efficacy is improving. These are all positive signs. We've also been fortunate that our NetCentric business has quite honestly outperformed the expectations we've had going into the pandemic and for a longer period of time. That business actually has the attribute of being more profitable on an incremental dollar contribution basis because the percentage that is on-net is much higher.

All of this put together says that we are going to see a growth in EBITDA. If we go back and look at Cogent's average since going public, we grew top line at 15% -- or 10% and EBITDA at roughly 15% throughout that period. Today, we're growing top line at 3.5%, 3.6% this quarter and EBITDA at about 5%. Those are materially slower. We do think it is improving, and that's why we have the confidence in growing our dividend.

As we think about the dividend going forward, we continue to monitor it. But we have substantial liquidity on the balance sheet, substantial incremental borrowing capacity, and we feel that we have a lot of confidence in our ability to continue to grow the dividend.

Now the pace of that growth may vary. It is not today. But again, we're going to monitor the situation as we go forward. But we feel very comfortable in this business producing more free cash flow each quarter.

The last part of that question, Dave, was is there a point where your leverage ratio gets to where it -- where you slow it substantially or just don't grow the dividend. Everything you said I understand is -- as you voice in your confidence and what you think the signposts are on the turn, but we know -- like we've got greater uncertainties than ever right now in terms of the economy, inflation, among a host of different things. I mean, I could argue also that the conversions of office to residential are going to be more than just small market, but I see what my kid has to pay for rentals and the cost of construction to put new residential in New York City and a lot of these cities.

So there's an uncertain future. And the core question is, at what point does the Board say 4x leverage is enough, 4.5 -- and there might be -- we can agree that 10x leverage is probably too high a leverage. So we can agree that there's a number. I'm just trying to get a sense of what is the number.

Okay. Well, I can definitely tell you that...

Is it 5? Is it 6?

Well, our indentures actually have limits. We cannot have more than 6x gross leverage and 4x secured leverage. Those are contractual terms that are built into our indentures. We are very far from both of those limits.

In terms of the business' ability to maintain elevated levels of leverage, it really is a question of the cost of rented capital of debt. Our debt capital costs remain substantially below our competitors and substantially below historic averages. So I don't think the Board has a hard number saying, "The day we hit a specific number, we're going to slow the dividend growth down." But it is something that we think about every single Board meeting, we consider it, and we look at the totality.

We also know, as Tad pointed out, there's $350 million on our balance sheet. And we've been pretty clear: we only need to keep $60 million or $70 million to run the business. So we have a lot of cash to eventually return. It's now unencumbered through the modification of our covenants. And for the foreseeable future, we continue to expect to return excess cash. That's as much -- I can't give you a hard number because one doesn't exist.

Our next question comes from David Barden with Bank of America.

Walts been giving me a lot of (expletive) on Twitter, but that was a good question. So I guess I've got a couple. Number one would be on the kind of sales force hiring and sales force productivity. I guess, Dave, are we looking at -- are we redlining productivity from what sales force we have left? And how is your success rate in recruitment? I remember when you guys went fully back into the office, you actually cited coming back into the office as being one of the problems that you had with retention. And so I'd like to kind of talk about that a little bit.

The second question would be, Lucent -- or not Lucent -- Lumen and AT&T have both kind of cited that wholesale prices have been rising for them, and it's been a problem for them. And people have been kind of hunting for who the potential beneficiary of that has been? And maybe that's Cogent. I don't know. But I would love to hear your view on that.

And then the last, if I could, would be -- at the risk of extending this call for a while, would be: there's a prominent short-seller/investor who's come out with a short thesis on the data center industry, which by the transit of property would have negative implications for Cogent and its connectivity prospects in the data center industry. I'm confident you're aware of what that is, and I'd love to hear your thoughts on the likelihood of the near-term disintermediation of the data center industry by the cloud providers.

Yes. And listen, I'm here to answer the questions. I really appreciate everyone staying late and listening to our answers. And I feel that it's important that we do answer them. So let me take the sales force one first.

The rate of decline in our sales force substantially moderated. So from Q4 to Q1, we reduced sales force by 12%. From Q2 to Q3 -- or Q1 to Q2, we only lost 2 net salespeople. We have actually since the beginning of the pandemic hired at record rates. We have no shortage of applicants. We have had peak hiring months of over 60 new salespeople and a 600 -- or 500-person quota-bearing sales force, 650 total organization.

Yes, there are individuals who don't like working in an office. There are individuals who don't like being vaccinated. And each of those represents some of the reason for higher turnover. And most of the reason is lower productivity. To me, it was extremely encouraging that we returned to nearly our 15-year average of productivity. So we have not done this well in sales force productivity since Q4 of 2018, so long before the beginning of the pandemic.

Now our turnover rate did tick up materially from an average of around 5% -- 5.2% of the sales force all the way up to 8.7%. It declined last quarter to 6.9% and declined again in this quarter. We're very much encouraged by that. We think the size of the sales force will grow from this point. It basically flattened and is growing. We're not seeing any issues in terms of sales force.

In terms of productivity, this is a combination of having the people in the office, learning from all the training that we experimented with in a remote environment and improving the environment. So we're pretty encouraged by our sales force efficacy and its ability to continue to drive incremental revenue growth without an increased subscriber acquisition cost.

Now to your question about wholesale pricing. I'll be honest. I heard that comment from one of the large carriers and I didn't quite believe it. We buy off-net circuits. We have almost 13,000 off-net customers. And we buy from 90 different suppliers. In almost every instance, we have at least 2 providers. We only sell off-net services that are delivered over fiber. And we have seen a 6% or 7% compounded rate of reduction in the price for those wholesale loops. And we think that will continue, particularly as there has been a host of fiber overbuild projects funded both by private equity as well as incumbent carriers.

So one, we're not raising our prices to the customers we sell in wholesale in NetCentric; and two, we're experiencing lower prices as we buy.

Now to your last question about Mr. Chanos' short thesis on data centers. While we serve 1,450 data centers operated by several hundred different carriers -- we're in more equinoxes than any other carrier in the world. We're in more interactions. We're in more core sites. We're in more sectors. Pretty much every multiple site data center operator, we're in the vast majority, if not all of their facilities.

We ourselves have a very limited data center footprint. It is typically associated with the sales office and our network equipment. There's a reason why we have chosen not to spend capital and build into that market. I know data center valuations have been very robust. But data centers have not, for the most part -- there are some limited exceptions -- demonstrated a consistent ROIC above their cost of capital, and they blame that on perpetual growth. There is absolutely hyperscale cloud substitution. There's technological obsolescence. There's oversupply and there's competition.

Now we're agnostic. What we have seen is many data centers have -- that we serve quarterly go out of business. We probably unwind 5 or 6 data centers every single quarter. What's ironic, however, is usually after we unwind it, the next quarter, we rewind it. We don't remove our fiber. And a new operator comes in and takes over that facility.

So I can't comment on the profitability long term of the data center operators. What I do know is there is fundamental industrial demand for space and power in condition facilities that have access directly to the Internet. That's what we sell. And whoever operates it, I don't really care. I care about the customers in that facility. And that part of our business has experienced growth beyond our normal trend line. It's our NetCentric business, and we feel comfortable. It's probably going to continue to grow longer than we had anticipated at the beginning of the pandemic.

Our next question comes from the line of Michael Rollins with Citi.

One follow-up and one question. So the follow-up would be, Dave, you did mention the goal to build 100 data centers annually for the next few years. What's the implications for CapEx on that type of a build plan? And then secondly, on the subject of tax. Where are you today in terms of the tax position and when you start paying material cash taxes? And does the proposed tax changes in D.C. have any influence on that timing or your expectations?

Yes. So let me take the data center one first. We have a pretty robust funnel of actual building addresses and facilities that are under construction, both by existing operators and new operators and feel that the 100 a year is very realistic. It's about the pace we did last year.

We are probably going to do less multi-tenant office buildings, not because the market is not recovering quickly enough, but because even before the pandemic, our MTOB expansion had materially slowed as we had reached the targeted buildings that we wanted to. Now there are several new buildings coming on in major markets, and we will go to them. But that MTOB footprint expansion has slowed over the past year from about 7% to about 2% a year as measured by square footage. The data center number has been pretty consistent.

With that, we think that our total CapEx will be in line with last year. It came down on a sequential basis, both in terms of straight CapEx and principal payments on capital leases. We think that in our CapEx guidance, the 100 data centers a year are very achievable without a change in CapEx.

Now to your tax question. Cogent still has slightly over $1 billion of NOLs. However, only about $40 million of those NOLs are in the U.S. The majority of those NOLs are in international markets with the largest concentration in our holding company in Luxembourg that holds many of our European subsidiaries.

We believe we will be subject to a 15% minimum tax due to the BD calculation independent of a new statutory 15% minimum. So over the next year or 2, we will become a cash taxpayer, but we believe we will be able to utilize our international NOLs not paying taxes outside of the U.S. but then being effectively caught by the foreign indirect tax capture at 15%. So I think as investors kind of model Cogent in out years, I think that's a pretty good outcome. It's not perfect, but it's surely better than paying at a 21% or a higher rate. Thank you. And now for our final question.

And now for our final question. Our final question comes from the line of Brandon Nispel from KBCM.

It's Evan on for Brandon. Do you think you could talk a little bit about your thoughts around the share buybacks and maybe what conditions would need to exist for you to start thinking about starting that?

Yes. Sure. So it first starts with our return to -- return of capital question. And I think that's the question that Walt was focused on. He phrased it as a dividend question. But it's really a question about are we committed to continuing to return increasing amounts of capital to shareholders. The answer is unequivocally yes. The rate of growth of that is something that we will evaluate, but we are absolutely committed to returning increasing amounts.

Now the mechanism of return today is mostly dividends, in part because we have not seen the level of volatility that told us suspending the dividend and using a buyback was more effective. We have been able to classify the majority of our dividend as a return of capital, therefore, getting a tax advantageous result for the dividend recipient. As a result of that, we have decided to use dividends more than buybacks.

We consider that every quarter. We've returned $1.1 billion in total. Roughly $230 million of that has been through buybacks and about $850 million of that has been through dividends. I think at this point, it will probably remain dividend focused. But again, we'll evaluate each and every quarter based on market conditions.

No more questions. I would like to turn it back over to Dave for closing remarks.

Well, thank you all very much for staying late. We normally do this in the mornings, but we needed to accommodate some schedules. So hopefully, this was not too inconvenient. We thank everyone for their attention, and we're available to do any follow-up calls. Take care. Talk to you soon. Bye-bye.

Thank you for your participation in today's conference. This does conclude the program. You may now disconnect.