Phase Space AI

Trade Construction

Zscaler [ZS]

Zscaler [ZS] — Trade Construction & Risk

As of 2026-07-29. Spot $153.72. No position verdict is issued here — the book decides.


1. Liquidity Criteria — BINDING — PASS

Spot $153.72 (Alpaca latest trade, 2026-07-29)
Market cap $25.8bn on 168.0m fully-diluted shares
Enterprise value $24,052m (net cash +$1,772.7m, verified)
Trailing 252-day realised volatility 64.4%
Public float $17.9bn
Exit plausibility Mega-cap-adjacent, S&P 500 constituent, index-fund bid. No exit constraint at any size this book would take

PASS. Liquidity is not the constraint on this name.


2. Vehicle ladder — EQUITY (rung 1)

Per criteria.md, equity is the default and requires no argument. I am not stepping above it. But the required disclosure is given in full, and it contains a finding.

Required disclosure Value
1. Chosen vehicle Common equity. Ladder rung 1.
2. Implied vol minus trailing realised vol +1.1pp — measured, see below
3. Quoted size at the specific strike 238 × 141 contracts on ZS 17-Jun-2027 $150 call
4. Argument for stepping above equity None offered. Therefore the vehicle is equity.

The measured LEAP spread — and why it is a genuine finding

ZS LEAP, 17-Jun-2027 (10.6 months) Strike $150 Strike $160
Bid × size $40.72 × 238 $37.57 × 36
Ask × size $43.66 × 141 $39.89 × 97
Implied volatility 65.5% 66.4%
Delta 0.666 0.628
Trailing 252-day realised vol 64.4% 64.4%
IV − RV +1.1pp +2.0pp

This is the tightest implied-over-realised spread in this project's record. criteria.md documents that every chain pulled on 2026-07-29 showed IV materially above RV — five large-software names at an 11–15pp premium, CRDO at 108% implied against 90.7% realised. ZS at +1.1pp is a different regime, and it is a real observation, not a data error (it is corroborated by the $160 strike at +2.0pp and by ZS's genuinely elevated 64.4% realised vol after a 52% twelve-month drawdown).

So the LEAP is not badly priced. I still do not recommend it, for reasons the ladder names: 1. The thesis does not support paying for convexity. The Valuation Criteria is FAIL at −18.7pp. Buying convexity on a name whose 5-year implied path fails is paying premium to lever an unattractive expected value. 2. The 12-month target is +18.4% at a held multiple. A $150 call at $42.19 mid needs ZS at $192.19 (+25.0%) to break even at expiry — above the base target. The equity captures the base case; the option does not. 3. A LEAP is embedded leverage in an unlevered book, and the inverse-volatility sizing rule does not capture the change in risk profile. 4. The $160 strike is thin on the bid: 36 contracts. Chain depth is not strike depth (the PANW 18×8 lesson). The $150 strike is fine at 238×141; the strike one increment up is not.

Conclusion: equity, and the +1.1pp measurement is recorded because it is the datum, not because it changes the vehicle.


3. Sizing

Inverse-volatility sizing at 64.4% realised vol places ZS in the smallest sizing tier this book uses — ZS's vol sits above CRWD's 47.6% and PANW's 41.6% and just below NET's 60.4%. That is the interim protection the Downside Criteria relies on while it remains MEASURED.

Correlation constraint — this is the binding issue, not single-name risk. The book already carries PANW, CRWD and NET, all underwritten in reports/companies/, all Valuation-Criteria FAIL, all in cloud/network security. Adding ZS and S would make five correlated cybersecurity names. The concentration and correlation limits in portfolio_book.json govern here and the book must apply them; this memo flags the exposure and does not resolve it.


4. Entry, exit and invalidation — levels, and what refutes the thesis

Level Basis
Spot $153.72 2026-07-29
12-month base target $181.94 (+18.4%) 7.47x forward EV/S, 50th pctile of the 2026 regime
Bull $206.89 (+34.6%) 8.50x, 68th pctile
Bear $149.83 (−2.5%) 6.03x, 3rd pctile of the 2026 regime, on 12% growth
Deep bear ~$105 (−32%) 8% growth, 4.5x — below anything printed in 2026
52-week high ~$341 (spot is 45.1% of it) Momentum record

Invalidation gap — the specific, falsifiable conditions

The thesis is refuted, and the position exited regardless of price, if any of these prints:

  1. FY27 formal guidance (early Sept 2026) comes in below 16%. The 16–17% early guide is the base of the entire 12-month build. A cut is a direct falsification, not a disappointment.
  2. Organic net-new ARR growth turns negative. It was +9.5% implied for Q4 FY26 ex-Red Canary. Zero or below means the upsell engine is no longer offsetting the new-logo shortfall.
  3. NRR prints below 110%. The last figures were 114% then 115%. Below 110% and the upsell defence against PANW's bundle is failing, which is the bear mechanism.
  4. Total RPO growth falls below revenue growth. RPO at +29.8% against revenue at +25.4% is the accounting- quality support for the whole name. Inverted, the backlog is no longer confirming.
  5. Calculated billings growth falls below 18% on a TTM basis (currently +24.1%). Since ZS no longer reports billings, this must be reconstructed each quarter from the balance sheet — the method is documented in ZS_Research.md §2a and validates to 0.06%.

Monitoring that is now the analyst's job because the company stopped doing it

ZS retired calculated billings and DBNRR as reported metrics. Both must be reconstructed or extracted every quarter:


5. Downside Criteria (MEASURED — logged, blocks nothing)

Named cause: new-logo acquisition failure compounding through a sales-leadership vacuum while PANW's bundle takes the greenfield. Full argument in ZS_Research.md §7.

Scenario Target Return P
Bear — growth settles ~12%, multiple to the 2026 low $149.83 −2.5% 0.20
Base — 16.5% guide delivered, multiple held at the 2026 median $181.94 +18.4% 0.50
Bull — new-logo fix lands, growth re-accelerates above 20% $206.89 +34.6% 0.30
Deep bear (tail, inside the 0.20) ~$105 −32%

Going concern: not applicable, and not argued. $3,539.1m of cash and short-term investments, $1,772.7m net cash, $890.2m trailing free cash flow. The bear case is a de-rating on decelerating growth, not solvency.

For Brier scoring, logged to the ledger: the bear cause is new-logo stall, not macro. This is a falsifiable attribution — management itself has removed macro from the explanation (macro mentions: 27 in FQ2 FY23 Q&A → 0 in both FY26 calls).