Cogent Communications (CCOI) – My Thesis
Entry / Exit Plan (not advice): 👉 Accumulate around $17/share 👉 Long-term exit 10–15 years
Dated: 25-Nov
Why I'm Writing This
This is not a stock recommendation. It is my personal investment thesis to articulate why I believe Cogent is mispriced. Writing it helps clarify assumptions, risks, and long-term expectations.
Please read my general investment disclaimers. Important: Not financial advice.
The Hard Truth
Starting With the Negatives
I want to deliberately open with the negatives because these explain why the stock has fallen sharply. They set the emotional context for why an opportunity may exist.
The Three Negatives
No insider ownership
Dave Schaeffer was forced to sell all his stock due to personal real-estate leverage.
High leverage
~$2.5B of debt on a company with ~$850M market cap.
"Dumb pipe" economics
Price per Mbps declines every year.
These issues look serious, and the market has priced the stock accordingly. However, each of these negatives will be explained and contextualised later in this document.
The Core Thesis
For now, we move to the core thesis — what the company actually does and why it matters.
The Cogent Thesis (Understanding the Business)
Cogent Is One of the Largest Internet Backbone Operators in the World
Cogent sells high-capacity internet connectivity to:
- Data centres
- Enterprises
- Cloud providers
- Telecom carriers
- AI and hyperscale compute clusters
It operates a global network capable of reaching ~98% of all internet destinations. Its business is extremely simple: sell bandwidth (IP transit, dedicated internet, and now optical wavelengths) at the lowest cost in the industry.
The Company Was Built Around One Radical Idea
More than 20 years ago, Cogent's founder, Dave Schaeffer, realized something that most telecom companies resisted:
The internet is a dumb pipe. Everything — voice, video, enterprise applications — eventually becomes IP traffic.
This insight shaped Cogent's entire business model:
- No expensive legacy services
- No bloated product portfolio
- No multiple network types
- Extremely lean organisational structure
While incumbents like AT&T, Verizon, Lumen, and others carry decades of legacy systems and products, Cogent focused on being the cheapest, simplest, and most standardised bandwidth provider.
This is why Cogent consistently wins in markets with heavy price competition — it is structurally the lowest-cost producer.
The Sprint Wireline Acquisition: A $20 Billion Network Bought for $1
In 2023, Cogent acquired Sprint's entire wireline business from T-Mobile for $1. Sprint spent over $20B building this national long-haul fibre backbone.
Cogent effectively acquired:
- A nationwide fibre network
- International subsea routes
- Dozens of data centres
- A fully redundant long-haul optical transport system
- Massive rights-of-way and dark fibre
The infrastructure is impossible to replicate today at rational cost. This acquisition has two transformative outcomes:
Cost advantage → Cogent's wave products now run at 90–100% gross margin.
Scale advantage → Cogent instantly becomes a major long-haul backbone provider.
The Wavelength Business (Waves) Is the Growth Engine
Optical wavelengths are point-to-point 10G / 100G / 400G circuits connecting data centres. They are essential for:
- Cloud storage
- AI training clusters
- AI inference
- High-performance computing
- Inter-data-centre traffic
- Hyperscale infrastructure
Cogent sells these wavelengths:
- ~20% cheaper than Lumen/Zayo
- With faster installation times
- On a modernised, simplified network
- At extremely high margins (because the core network was acquired for nearly free)
The demand for waves is exploding because:
- AI clusters need huge east–west bandwidth
- Cloud regions need redundant paths
- Hyperscalers always multi-source from multiple providers
Cogent fits perfectly as a third or fourth provider, adding resiliency at a low incremental cost.
AI Megacycle = Bandwidth Supercycle
AI training clusters depreciate at millions per month. A single day of downtime is catastrophic. Therefore:
- Hyperscalers buy multiple redundant wave paths
- They need low-latency, high-throughput, reliable fibre
- They must link hundreds of data centres
- They need scale and cost efficiency
Cogent's architecture, prices, and simplicity make it extremely attractive in this environment.
Emerging Markets Provide Multi-Decade Growth
Cogent has entered India — one of the world's fastest-growing internet markets. Growth in India, Southeast Asia, and similar markets will be driven by:
- Cloud expansion
- Mobile broadband
- Streaming
- AI/data centre buildout
Cogent's low cost structure and global peering position make it well suited for these markets.
The Real Story Behind the Negatives
(This combined section addresses both the concerns and their proper context.)
Insider Ownership – Forced Sale, Not Business Weakness
Dave Schaeffer never took a salary; he took stock. He pledged that stock against loans tied to commercial real estate. When real estate collapsed, margin calls forced him to sell all shares. This had nothing to do with Cogent's fundamentals. Dave remains wealthy and aligned with long-term value creation, and he will rebuild ownership through future stock-based compensation.
High Debt – But Low-Risk, Asset-Backed Leverage
Telecom companies operate with leverage — that's normal. Cogent's debt is backed by:
- long-haul fibre
- metro fibre
- data centres
- right-of-way
- dark fibre
- acquired Sprint assets
These were obtained at distressed prices, but are worth billions. Cogent has paused buybacks and dividends to prioritise deleveraging. The business remains profitable even during the messy Sprint integration.
Dumb Pipe Economics Aren't a Problem When You're the Lowest-Cost Producer
Yes, bandwidth prices fall every year. But usage grows far faster. And Cogent is the lowest-cost bandwidth provider because it:
- acquired Sprint's network for effectively nothing
- runs a lean, standardised network
- avoids legacy product baggage
- targets high-volume, high-utilisation markets like data centres and waves
In commodity markets, the lowest-cost producer wins.
Valuation View
Current Snapshot
Share price
Market cap
Debt
EV today
Assumed DC sale**: +$500M → adjusted EV ≈ $2.85–3.0B
**Sprint deal also brought in certain Data Centres which Cogent has put on the market. $500M is conservative figure (after tax) on proceeds from these sales.
ARR vs EV
Current ARR: ~$1.0B
Expected ARR: ~$1.5B (wave ramp)
EV / ARR ≈ 2.0x — nearly unheard-of for a recurring-revenue backbone business.
EV vs EBITDA
Expected EBITDA (post-transition): $700–750M
EV / EBITDA ≈ 4×
Comparable fibre networks trade at 8×–12×. The market is pricing in failure — while the underlying economics improve every quarter.
The Long View
Horizon
This is a 10–15 year investment thesis, not a short-term trade. Cogent owns a $20B fibre network acquired for $1, is scaling a 90–100% margin wave business, and benefits from the global AI bandwidth supercycle. Deleveraging, asset monetisation, and ARR expansion can all lead to significant re-rating over time.
Subscribe to My Whatsapp Channel
If you like this content, you can subscribe to my channel to get notified when new articles are published.
I don’t post often, so there shouldn’t be any spam.
If you would like to browse all my earlier investment theses, you can find them here: