Phase Space AI

Valuation

Palo Alto Networks [PANW]

Palo Alto Networks, Inc. [PANW] — Valuation

As of 2026-07-29 · spot $314.335 · Two outputs, both required and both given. They disagree, and §4 names why that disagreement is legitimate rather than an error.

---

1. Verified inputs

Read directly from the Q3 FY26 10-Q balance sheet and income statement.

Cross-checks passed: net loss ÷ diluted shares = −177 ÷ 801 = −$0.221 vs filed −$0.22 ✓; prior quarter 432 ÷ 711 = $0.607 vs filed $0.61 ✓. The 2:1 split of 2024-12-16 is correctly reflected (confirmed in the raw price series as a −48.5% single-session move).

---

2. The exit multiple — growth-matched anchor

PANW must be anchored on its ORGANIC growth of 14.2%, not its reported 31.1%. Anchoring on 31.1% would import the acquisition into the comparator selection and then value the acquired revenue at a growth multiple it does not possess — a compounding of the screen's original error.

At 14.2% organic, PANW sits squarely in the mature-enterprise-software band. Comparator set, every member's growth and EV computed from its own filings on this date:

The set brackets the subject on growth (7.0% to 20.6% versus 14.2%). Growth-matched: SATISFIED.

Usable EV/EBIT anchors: ADBE 11.6x · ORCL 14.9x · INTU 16.2x · MSFT 18.8x · CRM 20.8x · VEEV 27.7x. Base 17.5x (median) · Low 11.6x · High 27.7x.

Note the cross-sectional fact before any DCF. WDAY grows organically at 13.5% — within 0.7pp of PANW's 14.2% — and trades at 4.20x sales against PANW's 23.62x. PANW is priced at 5.6 times the sales multiple of the closest growth-matched name in the set. FTNT, the nearest cybersecurity comparator at 20.1% growth and a 31.1% GAAP operating margin, trades at 15.52x sales. PANW trades at 23.62x with a 9.6% GAAP margin.

Terminal operating margin: 20% base, with the full range run. Evidenced: −3.4% (FY22) → 5.6% → 8.5% → 13.5% (FY25) → 9.6% TTM (diluted by acquisition accounting), against a guided 28.9–29.2% non-GAAP. Twenty percent sits between the FY25 peak and the non-GAAP figure and assumes the purchased-intangible amortisation rolls off. Because the verdict does not depend on this input at all (§3), I run it from 9.6% to 35% rather than defending a point estimate.

---

3. Implied-path test — the Valuation Criteria

assets/reverse_dcf.py. Terminal value is 100% of EV by construction, so the reverse DCF is the primary long-horizon output.

Solve 1 — required revenue CAGR

Held fixed: terminal EBIT margin 20%, WACC 9.5%, 5 years.

The run-rate case is the fair one and it is presented because TTM understates PANW's base. TTM contains only one quarter of CyberArk and Chronosphere, so it structurally overstates the required growth. Annualising the Q3 run-rate and using the company's own guided diluted share count is the correction. It moves the margin from −46.2pp to −43.1pp. The conclusion does not change.

Solve 2 — required exit multiple at demonstrated growth

Held fixed: revenue CAGR 14.2% (demonstrated organic), WACC 9.5%.

Against an anchor set whose maximum observation is 27.7x: - At a 20% GAAP terminal margin the price requires 95.7x — 3.5 times the highest growth-matched anchor. - Even crediting PANW's full non-GAAP 29% margin, the requirement is 59.8x — 2.2 times the highest anchor.

Read the "implied change" column carefully: the price requires only a 19% multiple compression over five years while revenue grows 14% a year. The market is pricing PANW to remain a ~60–96x EBIT asset in 2031.

Sensitivity 1 — the exit multiple

Nothing inside the anchor range comes close. The verdict only approaches neutral at a 60x exit multiple — 2.2x the highest observation in a growth-matched set of eight.

Sensitivity 2 — the terminal margin, which cannot rescue it either

Exit multiple held at 17.5x, demonstrated growth 14.2%:

Verdict

FAIL. Margin −46.2pp on the TTM base, −43.1pp on the fairer run-rate base, and FAIL at every terminal operating margin from 9.6% to 35% — including at a margin above the company's own non-GAAP figure.

This is not a marginal call and the memo will not hedge it. The brief anticipated a cluster of names "mildly expensive rather than clearly wrong," and three of the five are exactly that. PANW is the exception. Where CRM lands at −0.9pp and WDAY at −0.0pp — cases where a judgement input decides the answer and the analysis honestly cannot — PANW fails by 29 to 72 percentage points across every input combination tested. No judgement input rescues it. Saying otherwise would be manufacturing balance rather than conviction.

For scale: a 60.4% five-year revenue CAGR takes PANW from $10.6bn to $112.7bn of revenue by 2031. That is larger than Oracle's entire current revenue base of $67.4bn.

---

4. Twelve-month target — and why it points the OTHER WAY

4a. Own-multiple history, with percentile stated

P/S at fiscal quarter ends, explicit split factors applied (3:1 on 2022-09-14 and 2:1 on 2024-12-16, both confirmed in the raw price series):

2019-01 9.4 · 2020-01 7.8 · 2021-01 6.2 · 2022-04 9.2 · 2022-10 8.9 · 2023-01 10.2 · 2023-04 12.7 · 2023-10 14.9 · 2024-01 13.6 · 2024-07 15.3 · 2024-10 17.3 · 2025-01 15.3 · 2025-04 15.8 · 2025-10 13.8 · 2026-01 12.2 · 2026-04 21.5

PANW has never been this expensive on sales in its clean, split-consistent history. The previous high was 21.5x last quarter; before the acquisitions it had not exceeded 17.3x.

Data note: two observations in the 2021-10/2022-01 tail read 34.7x and 34.2x. Those are artifacts of an as-reported share count that predates the 3:1 split interacting with an adjusted price, and I exclude them rather than let them flatter the current percentile. The clean-window figure above is the honest one.

4b. Regime-change test

The own-multiple history spans a 15.0pp growth-regime change, so mean reversion within it is declared UNIDENTIFIED per valuation.md, exactly as for the other four names.

4c. The stated target — and an important asymmetry in what "flat multiple" means

I apply the same uniform rule as the other four names: the identified 12-month component is NTM revenue and the share count at an unchanged multiple.

• NTM revenue $14,500m. Derivation: Q4 FY26 guidance $3,345–3,355m (the first quarter with a full period of both acquisitions) × 4 = ~$13.4bn, plus ~8% mid-year growth → ~$14.5bn. Company guidance, disclosed.
• Shares in 12 months ~845m. The company's own Q4 FY26 diluted guidance is 830–840m, plus ~1% of SBC issuance and zero buyback. This is the only name of the five where the forward share count is higher.

"Flat multiple" is a conservative floor for CRM, ADBE and WDAY — which sit at the 5th to 10th percentile of their own histories. For PANW it is the opposite: holding 24.15x means holding a multiple above every observation in the company's clean history for a further year. That is not a neutral assumption and it must not be presented as one. I report the arithmetic and I decline to call $414 a target.

Stated 12-month position: the 12-month output for PANW is UNIDENTIFIED, with a defensible range of −21% to +32%, and the range is that wide because the entire question is whether a 100th-percentile multiple holds for twelve more months. Nothing in the filings answers that.

4d. The two horizons genuinely disagree, and that is the point of having both

The 12-month arithmetic is favourable and the 5-year implied path fails by 46pp. Both are correct, because they measure different things:

• Next twelve months: reported revenue grows ~37% as the acquisition anniversaries through the P&L. Estimates rise. Momentum is the strongest in the cluster (+436% over five years). Nothing forces the multiple to compress inside a year.
• Next five years: the anniversary is a one-time step. After it, PANW is a 14%-organic grower carrying a 23.6x sales multiple, and the required exit is 60–96x EBIT against a 27.7x anchor maximum.

This is precisely the horizon distinction valuation.md was written to enforce — the file exists because a 5-year instrument provably cannot be calibrated to a 12-month one (best achievable correlation +0.08). PANW is the clean demonstration: a name can be a reasonable twelve-month hold and an unownable five-year asset simultaneously. Reporting only one of these would misrepresent it in either direction.

4e. External sanity band

No professional target is on file for PANW in this book's reference set. The check cannot be run, and that is stated rather than substituted.

---

5. Downside case with a named cause

Type: MEASURED. Logged, scored, and it does not reject the name — the Valuation Criteria already does.

Scenario — multiple compression toward its own history as the acquisition anniversaries and organic growth becomes visible in the reported line. Probability 45% — the highest in the cluster, and the only one above one-in-three.

Named cause and mechanism, with a date. Reported growth is ~31% today and ~37% over the next twelve months purely because CyberArk and Chronosphere are consolidating into a base that excluded them. In Q3 FY2027 — the quarter ending April 2027 — the acquisitions fully anniversary, and reported growth converges to organic. On today's trajectory that is roughly 14%, printed by a company trading at 23.6x sales with a negative GAAP operating margin, zero buyback, and a share count that has risen 22%.

Nothing needs to go wrong operationally. The bear case is the arithmetic of the anniversary arriving. It has a date, which is unusual for a compression thesis and is why the probability is set at 45%.

Quantified: PANW reverts to the median of its own clean multiple history (14.44x sales) on NTM revenue of ~$15.5bn by then, across ~855m shares → approximately $262 per share, a 17% decline from spot. In the harsher case where organic growth also slips below 13% and the multiple reverts to the pre-acquisition high of 17.3x on a lower revenue base, roughly $200–215, a 32–36% decline. A move to the growth-matched cross-sectional multiple — FTNT's 15.52x sales, on a 31% GAAP operating margin PANW does not have — implies ~$285; a move to WDAY's 4.20x would imply ~$80, which I record as arithmetic rather than a forecast because no 12-month path delivers it.

Not a going-concern case, and not remotely close. Net cash of +$5,640m, a 38.5% trailing adjusted free-cash-flow margin up 430bp, $3,716m of FY25 operating cash flow, and $1,352m of convertible notes as the only debt. PANW's cash generation is genuinely excellent and it is the strongest single fact in its favour. Stated explicitly because criteria.md requires a going-concern case to be flagged where it exists, and here it emphatically does not. The risk is entirely valuation, not solvency — which is a materially better kind of risk than CRM's levered version of the same problem.

Interim control: inverse-volatility sizing. PANW's 252-day volatility is 41.6%, mid-range in this cluster (ADBE 37.7%, CRM 40.6%, CRWD 47.6%, WDAY 49.0%). Note the limitation honestly: inverse-vol sizing does not protect against this bear case. The Downside Criteria's interim control works because "a fat-left-tail name is almost always a high-volatility name" — PANW is not. It is a mid-volatility name whose left tail is a multiple that has to compress. Inverse-vol sizing would size PANW third-largest of the five while the implied-path test ranks it worst. That divergence is the one place in this cluster where the sizing mechanism and the analysis actively disagree, and the book should know it.