Charter selected Harmonic's vCMTS architecture in 2023 and expanded the partnership in November 2025 to its full service area plus DOCSIS 4.0 remote PHY devices. VISN's Q1 2026 10-Q records the consequence — Charter dropped below the 10% disclosure threshold — but does not name the cause: Charter's spending continued; the supplier changed.

VISN is now post-divestiture pure-play Aurora (broadband access / CMTS for cable MSOs). Comcast: 39% of net sales, accelerating from 21% (2024) to 35% (FY2025). Aurora reporting-unit goodwill: 9% fair-value headroom at the October 2025 test, with explicit impairment warning. RUCKUS sold to Belden ($1.846B, ≈12.5x adj EBITDA, H2 2026 close) — a premium Belden rationally paid for a 60%+ GM, 20%+ EBITDA-margin asset they integrate into a 38%/17% profile. The capital realization is real; the residual is a single-customer cyclical with a binary architectural option attached.

What the market is pricing

At $11.72 (post the April 27 $10/share special distribution) on 225.5M shares: market cap $2.64B. Net cash $255M; PV of RUCKUS proceeds at 78% close ≈ $1.30B. Implied Aurora EV ≈ $1.09B, or 5.4x $201M annualized adj EBITDA. HLIT, the closest cable-equipment comp, runs 8-11x. Broadband infrastructure 10-14x.

The honest read on the multiple: the market sees the structural risk. We thought it didn't — that the post-distribution rally was capital-return positioning and the architectural cliff was unrecognized. That was wrong. RSI 70 captures the rally; 5.4x captures the bear. A probability-weighted SOTP across cycle states (continues / decelerates / pulls back / cuts) and architectural states (Comcast stays / dual-supplies / migrates / delays) lands at fair value ≈$13.00 — a $1.30/share gap, Sharpe ≈0.3-0.4 over an 18-month horizon.

What's not fully priced is unknowable from inside the filing. The 5.4x discount could reflect a 30% probability of Comcast architectural migration, or a 50% probability. We assume something closer to the lower end given the November 2025 Charter precedent is fresh and binary outcomes 18-24 months out tend to under-discount in cyclical multiples.

Judgment

Not a trade today. The structural read is correct, the multiple has done most of the work, and the catalyst window is too far out to fight extended momentum on either leg. The pair (long HLIT / short VISN) is the cleanest expression of the architectural alpha but both legs trade extended (HLIT RSI 78, IV rank 222%; VISN RSI 70).

Activation triggers:

• VISN to $9-10 — asymmetry restores; Sharpe → 0.6-0.7 on the long
• HLIT cools to RSI <50, IV rank <100 — pair entry becomes viable
• Aurora goodwill impairment 8-K — counter-intuitively bullish. The 5.4x already prices ambient cliff risk; an impairment crystallizes the bear, anchors the multiple, and removes the binary uncertainty. The cleanest long entry is after the bad news has a number on it.
• Comcast publicly commits to integrated CMTS — kills the architectural binary; Aurora rerates to 8-9x cycle multiple
• RUCKUS-Belden close 8-K — collapses the event leg; Aurora standalone repricing on observed numbers

Risks (ranked)

1. Comcast follows Charter to vCMTS at DOCSIS 4.0 — binary, 2027+. If adverse, Aurora reprices 30-50% over a 2-3 year deployment with only 9% goodwill cushion to absorb.
2. Cycle deceleration as Comcast nears 80%+ mid-split coverage — 2026-2027. Goodwill impairment likely on a 10-15% Aurora revenue decline.
3. RUCKUS deal break or material slip — ≈22% probability per timing risk. Removes ≈$7/share from SOTP.
4. Comcast capex cut — lowest probability, highest severity.
5. Stranded costs post-RUCKUS — filing warns "additional restructuring charges could be material."

Catalysts

Evidence

Independent signals only. LRs are not multiplied — they're correlated through shared underlying events. Read each as the marginal update from that signal alone.

LR signal 0.9. Bearish lean from a structural finding the filing doesn't name. The multiple has done most of the work; the remaining edge is a $1.30/share weighted gap, Sharpe sub-threshold at current entry. The work surfaces real PnL when triggers fire, not on the print itself.