Cogent - Investor Relations for a Day: FAQs
Our published Cogent materials attract a lot of inbound inquiries to discuss the company. Since we started publishing articles on Cogent in mid-2023 we’ve fielded about 50 inbounds from other firms. Consequently, we have a pretty good sense for what people are asking.
In that light, we will have a little fun with this Insight and, for a day, role play Cogent’s investor relations department, which doesn’t exist. This Insight assumes the reader has some pre-existing knowledge (and confusion) about Cogent and will be formatted as an FAQ.
1. How is the Sprint acquisition going?
This is an incredibly loaded question. Three high-level answers:
From an infrastructure perspective, the wave network has been completed. Cogent can do 10, 100, and 400 gbps optical transport (wavelengths) anywhere-to-anywhere across a footprint of over 900 data centers in Canada, the US, and Mexico. This is an incredible achievement which should fuel significant growth in revenue and cash flow over the coming years.
From an asset perspective, the fallow assets are still yet to be monetized. No monetization of the IPv4 address space, and no monetization of the facilities converted to data centers. However, Cogent expects to ask for “best and final” binding offers for the data centers in the next month or so. Hopefully we will get some deals announced within 1H 2025.
From a business perspective, significant synergies have been achieved and, as of the Q4 2024 earnings call (Feb 27, 2025), Sprint was operating at breakeven – ahead of plan by a few months. The company expects Sprint to generate positive EBITDA for full year 2025.
2. I don’t understand the reported financials. Can you help me unpack them?
We will attempt to de-mystify Cogent’s reported results, which is not an easy task. One of the keys to this process is the following disclosure on Cogent’s Q2 2024 earnings call:
“Yeah, and for the past 15 months since the acquisition, we’ve been spending about $5 million a month on OpEx related to integration projects. We do not break that out separately. We did not add it back to EBITDA because we had the added complexity of the T-mobile add back and did not want to, I think, overly complicate our accounting. We anticipate that $5 million a month drag on EBITDA to be a 3-year drag. That was what we said it would take us to go from the negative 190 of acquired EBITDA to, if you just looked at those customers, about $85 million to $90 million of positive EBITDA. You know, I think we are on track to do that. I think that drag will continue, you know, beyond this year and then, you know, well into next year. But it will start to taper off. We are ahead of schedule in terms of taking costs out, which gives us the ability to turn off some of that extraordinary expense sooner.”
There are additional exceptional and non-recurring costs that Cogent has been paying to ready its data center portfolio for sale. From the Q3 2024 earnings call:
“Our cost of goods sold increased by $5.3 million from the last quarter due to 2 major expenses, the additional costs involved in converting former Sprint switch sites in the Cogent data centers that were not capitalized and certain vendor contract termination costs.”
There are other moving parts as well, such as:
How the Core Cogent business has been trending
Changes in payments from T-Mobile for the IP Transit Agreement
Changes in payments from T-Mobile from the Commercial Services Agreement
Changes in Audit and Bad Debt Expenses
Waves revenues ramping up
In the table below, we attempt to capture the moving pieces since the beginning of 2024:
Recurve Capital LLC Estimates
Key assumptions include:
Core Cogent is growing EBITDA MSD y/y
Commercial Services Agreement goes to $0 immediately
Data Center cleanup opex goes away after Q2 2025, when the refurbishment project is completed.
Integration expenses begin to tail off across 2025 and 2026. They should be at or near zero in 2027.
Waves revenue grows at 90% incremental margins. We assume a material ramp in waves revenue across 2025 and 2026 as Cogent ramps its installation capabilities.
We can see significant tailwinds to reported adjusted EBITDA as some of these non-permanent costs come out of the cost structure. Most companies would adjust these out in their definitions of adjusted EBITDA, but as shown in the quotes above, Cogent does not.
Above is a view of EBITDA, but the view of revenue is confusing – largely because Cogent is eliminating undesirable revenue streams.
If we assume Core Cogent (ex Sprint) has grown in the MSD range, Sprint’s implied run-rate revenues are now about $330m. Cogent inherited a $670m cost structure and outlined $220m of synergies, $200m of which have been achieved. The math doesn’t square - $330m of revenue less $470m of cash costs should result in -$140m of EBITDA.
To our understanding, the difference is that Cogent is eliminating more undesirable revenue than originally anticipated. There was more pass-through revenue and negative margin revenue – shockingly, to the tune of about $140m. The post-synergy Sprint may achieve slightly lower aggregate EBITDA (maybe $80m instead of $90m), but it will operate at significantly higher margins.
3. Can you help me understand what’s going on with the wavelength network? Why did backlog decrease if they are supposed to grow to $500m ARR by mid-2028?
We wrote a longer piece about Cogent’s network architecture, which can be found here.
At a high level, Cogent has taken an empty Sprint network and repurposed it with modern technology to be a scalable network that can be provisioned using standardized workflows at each endpoint and in the network operations center. This architecture connects to 906 data centers, most of which are carrier-neutral facilities where customers exchange traffic. The most recent wave network map can be found here.
Cogent brings ubiquity of endpoints, diverse physical paths, faster installation times, and cheaper prices to this market. This will allow it to be disruptive. So why did “funnel” orders decline?
Long story short, backlog decreased because Cogent hadn’t groomed opportunities out of the funnel - ever. Even the oldest opportunities from 18 months earlier.
First, let’s define the funnel as Dave Schaeffer describes it:
Tier 1: Orders contractually finalized, yet to be installed (i.e. not in the installation queue).
Tier 2: Opportunities with terms finalized, but not yet contractually signed into orders.
Tier 3: Opportunities with terms not yet finalized, not contractually signed.
The tiers are strongest to weakest, but even tier 3 is a known, discussed opportunity, quoted for customers going from endpoint A to endpoint Z, but there may be some final language to hammer out.
Cogent had about 2,300 wave opportunities in its funnel at year-end 2023 and “over 3,400” as of Q3 2024.
Source: Recurve Capital Estimates, Cogent Earnings Calls
It’s important to understand the nuance and context around funnel. The network wasn’t completed until year-end 2024. The sales force wasn’t pushing very hard on wave sales because the engineering teams were focused on building the network, not on installations – and customers don’t want to wait forever for service to be installed. Also, these are not as easy as plugging in a port to the internet – this is a dedicated, unprotected service from a specific point A to a specific point Z, along a specific routing path. Each sale is necessarily a more detailed technical discussion, i.e. not necessarily a quick sales engineering discussion that can be finalized in one short call.
The 2,300 opportunities in the funnel at year-end 2023 likely had little tolerance to wait around for >12 months. Most opportunities need to be provisioned within 3-9 months or else customers will look elsewhere.
What we learned is that Cogent spoke to about 2,500 opportunities in the 3,400 funnel (from Q3 2024). They started from the oldest to the newest, meaning they spoke to customers about many opportunities that were at least 12 months old, i.e. those likely were fulfilled elsewhere. Of the 2,500, 1,500 were groomed off. During some of those conversations, other routes were added back to the funnel. It cannot talk to every opportunity in parallel because Cogent has a 30-day installation SLA and it only has capacity to install about 500 wave orders per month. Therefore, it must work through the funnel in sequence.
We believe that about 40% of the YE2023 backlog remained. We estimate that another 60% of the ~500 bookings from 1H 2024, 20% of Q3 2024, and 10% of Q4 2024 will be groomed off due to the extended installation times (opportunities booked in Q4 2024 must wait until it’s their turn chronologically unless the need is urgent). As such, we believe the funnel will be groomed again to the tune of about -600 contracts in Q1 2025 – but then the grooming exercise largely will be complete, other than normal business-as-usual pushouts and cancellations.
We have heard that Cogent is ramping up its installations and we expect it to reach its goal of 500 contracts/month sometime in Q2. Dave told investors at a conference that 372 wave orders were in the installation queue at the end of February, suggesting that the 500/month target is not far off on the horizon.
We expect booking activity to accelerate, and we have heard that this is happening thus far in 2025. New bookings were around 1,000 in Q3 and 900 in Q4 – all while Cogent was giving extended installation estimates. We expect installations to ramp to 500/mo imminently, and for new bookings to ramp to that level within the next 3-6 months. The bookings ramp will allow for the excess funnel to clear and for Cogent to revert to a more normal “book and burn” type of cadence in the wavelength business.
4. Is Cogent over-leveraged and is its balance sheet and dividend strategy too risky?
Our short answer is “no,” but it wouldn’t be surprising if others disagreed with us. Cogent’s current elevated leverage is a function of a few different moving parts, all happening simultaneously and causing maximum under-earning in conjunction with maximum net leverage:
Investing $100m of capex into retrofitting data centers, while…
Investing $50m in upfront capex to integrate Sprint’s and Cogent’s networks, while…
Investing significant opex resources (see above) in integration expenses, while…
Taking losses on Sprint
The $100m of capex on DC retrofits should earn a quick payback. Cogent hopes to realize $1B of value from that portfolio. We assume a sizable discount to that, but the broader point is that there should be a significant deleveraging event on the horizon when Cogent monetizes this investment, presumably by the middle of 2025.
Investing $50m of capex to integrate the networks was essential to unlocking the new Wavelength growth vector. However, that work has not yet translated into material Wavelength revenue acceleration. This will be the primary story over the coming years. Cogent will earn pretty quick returns on this as well if it achieves even 50% of its medium-term revenue target ($250m, 90% margins vs. $50m of capex).
Integration opex, to the tune of about $45-50m/year currently and $60m/year previously, is essential for Cogent to (a) operate the wave network and (b) extract significant cost synergies from Sprint. Recall that these expenses will go away once the integration is done, but they are not added back to adjusted EBITDA, i.e. they create the illusion of significant EBITDA pressure. If we assume Cogent will spend about $150-200m on integration opex in aggregate but this spending unlocked (a) the wave business, and (b) Sprint transitioning from -$190m of EBITDA to +$80m, it is clearly money worth spending. But the non-existing Cogent IR team should have called this out and adjusted it out in its definition of Adjusted EBITDA.
Any rational business owner would spend the capex and opex described above because the paybacks are enormous, the IRRs are exceptional, and they put Cogent on a path for significant future growth and cash flow generation. Cogent’s revenue growth should begin accelerating at the same time that its capex programs roll off and the non-recurring operating expenses begin to wind down, which should translate into meaningful and immediate deleveraging.
On an operating basis, we aren’t concerned about leverage or Cogent’s capital allocation strategy. When we include Cogent’s fallow asset portfolio, we are even less concerned. Cogent could immediately raise $500M+ of cash within weeks if it wanted to. Specifically, Cogent could:
Sell a large portfolio of its unleased IPv4 addresses (it has >20m unleased, even in a “fire sale” situation they likely could sell for ~$30/address).
Take the quickest/easiest bids on its data centers to raise cash immediately.
Sell long haul dark fiber to hyperscalers. We believe a dark fiber contract for Cogent’s nationwide long haul footprint could command over $100M upfront and over $25M/year in maintenance payments.
In summary, Cogent is sitting at peak leverage with near-maximum “under-earning” in its recent historical financials, but is also on the cusp of (a) material wavelength revenue acceleration, (b) asset monetization, and (c) capex/opex rolling off, all while having significant fallow assets that could be monetized in a “just in case” scenario.
Conclusion
We know Cogent has been very confusing to unpack over the last two years. Few public companies take on acquisitions like Sprint because of the risk and complexity of creating massive investor confusion. However, we are at a pivotal inflection point: the Wave network is done, Sprint is operating above breakeven, and leverage has peaked. From here, we should see accelerating revenue, very healthy EBITDA growth, and asset disposals, all of which will contribute to reducing Cogent’s leverage position. We should be on the cusp of Cogent becoming the best and most disruptive growth story in its sector.
