CCOI — Cogent Communications: Asset Monetization Trade at $15

by psind707

Published: May 5, 2026 | Price: $15.01 (-35% intraday post-Q1 print) | Rating: Speculative LONG | PWFV: ~$22 | Time horizon: 12-24 months

TL;DR

CCOI dropped 35% on Q1 2026 earnings despite an EPS beat. The market is pricing structural decline in the legacy business and discounting the asset monetization plan. At $15, the setup is closer to "deep value with binary catalyst" than "short the cliff." This is not an AI infrastructure play — management itself confirmed the data center portfolio is "not well suited for AI training." The trade is the recap + 24-data-center sale executing.

Probability-weighted fair value: ~$22 (~47% upside), with 35% downside scenario to $6 if the DC sale fails again.

Q1 2026 Print (May 4, 2026, BMO)

The headline beat masked weak underlying signals: wavelength velocity slowed from 92.5% in Q3 2025 to 90.8% YoY (still strong but decelerating), off-net continued double-digit decline, and FCF stayed negative. Wavelength run-rate is still only $54M annualized — 5.7% of revenue.

Is CCOI an AI Infrastructure Play? No.

The wavelength business is the only AI-adjacent line, and it's tertiary in scale:

• CCOI wavelength run rate: $54M annualized
• LUMN Private Connectivity Fabric backlog: $13B

Same conceptual playbook (lighting hyperscaler-grade fiber off the Sprint network) but at a fraction of the scale, much later, and from a much weaker balance sheet. The math doesn't bridge the legacy melt fast enough — off-net is declining ~$60M/yr while wavelength is adding ~$25M/yr.

If you want AI fiber exposure, LUMN is the trade. If you want optical components, COHR/LITE.

The Real Story: Asset Monetization

4-Step Recapitalization Plan (JPM Credit Conf, March 2026)

1. Refinance $750M unsecured 2027 notes → secured pari passu with 6.5% 2032s
2. Move ~$569M IRU-backed obligations to "Infrastructure" entity
3. Sell 10 data centers to a global infrastructure fund — 100% of proceeds contributed to Cogent Group as collateral
4. Pro forma secured leverage: 3.91x (under 4x covenant test)

Schaeffer explicitly framed the DC proceeds as "nice to have, not must-have." The recap can close before the DC deal. He's also voluntarily committing 100% of the proceeds back into Group (where the lenders sit), not paying out to equity. Dividend cut 98% from $1.015 to $0.02 quarterly. Buybacks paused until 4x net leverage achieved.

Asset Stack

• 99 Cogent core data centers
• 86 Cogent edge data centers
• 482 owned technical buildings, 1.6M sq ft (Sprint legacy)
• 15.2M IPv4 addresses (~$450M+ market value at $30/addr; $380M already securitized in ABS)
• 24 facilities flagged for sale: 1M sq ft, 109 MW

Buyer

Schaeffer confirmed the 10-DC LOI is with one named "major global infrastructure fund." Specifics:

• 8 of 10 sites toured with third-party consultant
• Data room access granted
• Proposed price: "substantially more than $144M" (the prior 2-DC deal value)

Realistic shortlist based on size + DC platform track record: Stonepeak, DigitalBridge, KKR Infrastructure, Brookfield Infra, Blackstone Infra, Macquarie/IFM. DigitalBridge and Stonepeak are highest-conviction candidates.

Locations

24 facilities total ~1M sq ft / 109 MW (avg 4.5 MW/site, max ~10 MW). Tier-1 metro footprint:

• Northeast: NYC, Boston, Hackensack NJ, Pittsburgh, Pennsauken NJ
• Southeast: Atlanta, Charlotte, Miami, Boca Raton, New Orleans
• Midwest: Chicago, Dallas, Irving TX, Detroit, Kansas City (Sprint legacy hub)
• West: Phoenix, Salt Lake, LA, San Francisco/Stockton CA
• Mid-Atlantic: Reston/Ashburn VA, Washington DC

The 10 in the current LOI haven't been disclosed individually — likely a curated portfolio with the best power-to-fiber ratios.

Why NOT AI-Grade

Schaeffer himself: "this portfolio is not well suited for AI training."

• Power density too low: 109 W/sq ft vs AI training needs 50-100 kW/rack
• Average 4.5 MW per site — modern AI campuses are 100-1,000 MW
• Tier-1 metro locations skew to latency-sensitive end users, not rural cheap-power AI training markets

Actual use cases: AI inference / edge compute, cloud edge nodes, connectivity-rich colocation, enterprise managed services.

Pricing Math

Realistic 10-DC range: $400-700M. "Substantially more than $144M" probably means $300-500M. $700M+ would be a clear bull surprise.

Peer Valuation Comparison

CCOI trades at almost 2x LUMN's EBITDA multiple at spot with worse leverage and weaker AI positioning. The DC sale closing is what mechanically re-rates the credit profile from 7.4x net leverage to 4-5x. That's the difference between covenant risk and investment-grade-ish.

Three-Scenario Fair Value

Probability-weighted: ~$22. At $15 spot = ~47% upside to PWFV with 35% probability of going to $6.

Risk/Catalyst Map

Bull Catalysts

• 10-DC LOI converts to signed contract (Q2-Q3 2026)
• $750M unsecured-to-secured refi closes
• Wavelength continues 90%+ YoY growth into 2027
• Short squeeze: 13.76% of float short
• Second LOI on next 10 DCs announced
• Hyperscaler wavelength anchor contract (LUMN PCF-style)

Bear Catalysts

• Third LOI failure (already failed twice)
• T-Mobile $100M/yr cash runs off Q4 2027 mechanically
• Off-net decline accelerates beyond -17% YoY
• Covenant pressure forces fire sale or equity raise
• Wavelength growth decelerates below 60% YoY
• Founder-CEO continues insider selling (3 sales Q1 2026, 0 buys)

Operational

• Capex ~$185M/yr against negative FCF — cash burn continues
• Schaeffer's commercial real estate side business creating personal stress on his Cogent ownership

Investment Conclusion

Speculative LONG at $15 with 0.5-1% position size in a stagflation-aware portfolio.

This is special situation / event-driven, not fundamental valuation. The base business is melting; the equity value is in asset sale execution + recap closing.

Single-Leg Options View

• Stock over options is the cleaner expression at $15. Long stock captures the 3.4:1 reward/risk on PWFV with optionality on short squeeze (13.76% short interest).
• Jan 2027 $20 calls — captures recap close + DC sale catalyst window. IV crushed today, premium got cheaper. Asymmetric IF you believe execution. Size 0.25-0.50% of book. Closer to credit option than equity option.
• Jan 2027 $10 puts — bankruptcy hedge if you're long calls. Cheap insurance against scenario where DC sale fails AND covenants trip. 0.10-0.15% of book.
• Avoid Jan 2027 calls outright at higher strikes — wavelength growth already in price; better to own stock if bullish.

Sizing Discipline

• Initial: 0.5% of book at $15
• Add on first signed DC contract: +0.5%
• Cap: 1% of book
• Stop: $9 (clear breach of covenant scenario)
• Trim: $25+ pre-recap close, $35+ post-recap

Honest Caveat

I anchored too hard on "melting telco + leverage = avoid" before recognizing the asset monetization weight. At $23, AVOID was defensible. At $15, with $400-700M of contracted asset sales pending and 13.76% short interest, this is a different trade. The AI-infra angle remains tertiary — LUMN is the cleaner expression. But the equity at $15 is being priced as if the DC sale won't close, and that's the asymmetry.

This is research notes for personal use, not investment advice. Verify all figures independently before acting.