Zscaler closed the year ended July 31, 2026 with $898 million of fourth-quarter revenue (+25%), ending ARR of $3.77 billion, and a full-year free-cash-flow margin of 23% — while GAAP still printed a $63 million annual net loss. The growth signal sits on the balance sheet first: deferred revenue of $2.93 billion and remaining performance obligations of approximately $7.4 billion. This post models FY2022–FY2026 annuals in a public Beancount ledger so deferred revenue, RPO, and the cash-flow print can be compared line for line with CrowdStrike and Palo Alto Networks.
The Headline Numbers
Q4 is the print the market traded on September 3; the ledger models the FY2026 annual from the 10-K/8-K (conventions fold a lone Q4 into the year when the annual exists). The first table is the annual; the second is the quarter.
| Metric | FY2026 | FY2025 | YoY |
|---|---|---|---|
| Revenue | $3,353M | $2,673M | +25% |
| Gross profit | $2,575M | $2,055M | +25% |
| Loss from operations | $(133)M | $(128)M | — |
| Net loss | $(63)M | $(41)M | — |
| Operating cash flow | $1,130M | $973M | +16% |
| Free cash flow | $779M | $727M | +7% |
| FCF margin | 23% | 27% | −4 pp |
| Metric | Q4 FY2026 | Q4 FY2025 | YoY |
|---|---|---|---|
| Revenue | $898M | $719M | +25% |
| ARR (period-end) | $3,771M | — | +25% |
| Net new ARR | $246M | — | +24% |
| GAAP operating loss | $(15)M | $(32)M | improved |
| Non-GAAP operating margin | 24% | 22% | +2 pp |
| Deferred revenue (total) | $2,926M | $2,468M | +19% |
Revenue at +25% matched ARR growth at +25% — the subscription company print where the P&L amortizes prior bookings and the growth signal lives in ARR and RPO first. Excluding Red Canary, full-year revenue grew 20% to $3,209M and ending ARR grew 20% to $3,630M, so roughly five points of the headline are acquired. GAAP still loses money; free cash flow of $779M at a 23% margin is the cash story the income statement does not show.
Revenue Deep Dive
Zscaler does not publish a GAAP segment split in the earnings tables — subscription and support are the business. What management did name on the Q4 call is where incremental bookings are forming.
On the earnings call, CEO Jay Chaudhry said: "Security for AI bookings increased more than 50% sequentially in Q4 on top of a strong Q3." CFO Kevin Rubin added that the company "saw a 75% increase in pipeline in Security for AI" entering fiscal 2027 (sequential, not year-over-year). Those are product-ramp and demand signals, not scarcity-pricing claims — and they land on deferred revenue and RPO before they fully amortize into Income:Revenue.
The IR release frames the same idea without the sequential booking number: "Our continued innovation across Zero Trust SASE, Agentic SecOps, Data Security, and Security for AI is driving increased platform adoption." Tie that to the ledger: recognized revenue is the annual credit of −3353 MUSD on Income:Revenue. The commercial stock is ARR at $3.77B. What the balance sheet asserts is deferred revenue — $2,481M current and $445M noncurrent — on Liabilities:Current:DeferredRevenue and Liabilities:NonCurrent:DeferredRevenue. RPO of approximately $7.4B (about 45% current) is the contracted book including amounts not yet invoiced.
Red Canary contributed $141M of ARR at year-end and is the goodwill story: goodwill rose from $418M to $1,218M, and acquired intangibles from $47M to $214M. Platform adoption is not only organic seat expansion; the ledger shows the acquisition's footprint in noncurrent assets.
The Margin Story
| Period | Revenue | Gross margin | GAAP op. margin | GAAP net margin | FCF margin |
|---|---|---|---|---|---|
| FY2023 | $1,617M | 78% | −15% | −13% | — |
| FY2024 | $2,168M | 78% | −6% | −3% | — |
| FY2025 | $2,673M | 77% | −5% | −2% | 27% |
| FY2026 | $3,353M | 77% | −4% | −2% | 23% |
Gross margin held near 77% while GAAP operating loss stayed in the low-single-digit percent of revenue — stock-based compensation of $822M in FY2026 is the bridge between the GAAP loss and the non-GAAP print. CFO Kevin Rubin called Q4 "a record 24%" non-GAAP operating margin. That is the pricing-and-mix claim the release will own; the GAAP ledger does not invent it — it keeps the SBC in the operating expense lines (R&D $903M, SG&A $1,805M).
Free-cash-flow margin compressed from 27% to 23% even as FCF dollars rose, because Q4 capex and internal-use software jumped to $218.5M (versus $78.7M a year earlier). Cash generation is intact; the margin math moved with investment timing, not with a collapse in collections.
What the release does not say. There is no "demand exceeds supply," no "sold out," no sustained ASP-increase claim in the IR release. The bullish language is AI-security bookings, pipeline, and record non-GAAP operating margin — product ramp and operating leverage, not scarcity pricing.
The One Big Question: Can Security-for-AI Bookings Show Up in Deferred Revenue Before Guidance Slows?
FY2027 guidance is the tension. ARR is guided to $4.396–$4.426B (about 17% growth) and revenue to $3.908–$3.938B — a deliberate step down from the +25% FY2026 print. Against that, management is arguing that Security for AI is a new logo and upsell engine whose bookings and pipeline are accelerating into the slower year.
The ledger's answer is mechanical. If AI-security bookings are real and multi-year, they must thicken deferred revenue and RPO relative to recognized revenue. At July 31, deferred revenue was $2.93B against $3.77B of ARR and ~$7.4B of RPO — RPO/ARR about 2.0×, the same shape Palo Alto's peer table assigns Zscaler. A bookings boom that does not move those stocks is a story the balance sheet will refuse to tell.
Deferred Revenue, RPO, and the Security Peer Set
Three July 31, 2026 (or nearest reported) prints — the same columns every post in this peer set uses:
| Metric (period-end) | Palo Alto Networks (FY2026) | CrowdStrike (Q2 FY2027) | Zscaler (FY2026) |
|---|---|---|---|
| ARR / NGS ARR | $9.10B NGS ARR | $5.84B ARR | $3.77B ARR |
| Deferred revenue (total) | $14.76B | $4.84B | $2.93B |
| Remaining performance obligations | $21.2B | $10.7B | $7.4B |
| RPO / ARR | 2.3× | 1.8× | 2.0× |
Zscaler is the pure-play SASE column: smaller deferred-revenue stock than either peer, RPO still about 2× ARR, and a balance sheet that stayed acquisition-light until Red Canary. CrowdStrike's deferred revenue and RPO are the endpoint-growth shape; Palo Alto's are the platform-plus-CyberArk heavyweight. Line for line, the three ledgers assert current and noncurrent deferred revenue under the same account names so this table is not a metaphor.
Tracking a $7.9B Balance Sheet in Plain Text
Double-entry forces every dollar of deferred revenue, convertible notes, and Red Canary goodwill to land somewhere. That is why we model every company the same way — see how we model every company. Signs follow Beancount convention (income credits negative, expenses debits positive); Equity:Adjustments absorbs net income so the transaction sums to zero while retained earnings are set by the balance-sheet assertion.
; Check: −3353 + 778 + 903 + 1805 + 46 − 116 + (-63) = 0 ✓
2026-07-31 * "Zscaler, Inc." "FY2026 Income Statement"
Income:Revenue -3353 MUSD
Expenses:CostOfRevenue 778 MUSD
Expenses:ResearchAndDevelopment 903 MUSD
Expenses:SellingGeneralAdministrative 1805 MUSD
Income:OtherNet -116 MUSD
Expenses:IncomeTax 46 MUSD
Equity:Adjustments -63 MUSDThe balance-sheet number that carries the subscription narrative is deferred revenue's twin lines — −2481 and −445 MUSD on the current and noncurrent accounts — against goodwill's jump to 1218 MUSD after Red Canary. Cash fell from $2,389M to $928M while short-term investments rose to $2,546M: liquidity reshuffled, not evaporated.
The Multi-Year Arc
| Year | Revenue | Gross margin | GAAP op. margin | Net income | Deferred revenue (total) | Goodwill |
|---|---|---|---|---|---|---|
| FY2022 | $1,091M | 78% | −30% | $(390)M | $1,021M | $79M |
| FY2023 | $1,617M | 78% | −15% | $(202)M | $1,440M | $89M |
| FY2024 | $2,168M | 78% | −6% | $(58)M | $1,895M | $417M |
| FY2025 | $2,673M | 77% | −5% | $(41)M | $2,468M | $418M |
| FY2026 | $3,353M | 77% | −4% | $(63)M | $2,926M | $1,218M |
Five years compounded revenue from $1.1B to $3.4B while GAAP operating losses compressed from −30% to −4% of sales. Deferred revenue nearly tripled. Goodwill was flat after FY2024's step-up until Red Canary reopened the M&A account in FY2026. The compounding story is subscription liability growth and operating-loss compression — not GAAP net income, which remains slightly negative.
The Verdict: Bull vs. Bear
Bull Case
- ARR +25% to $3.77B with net new ARR +24% keeps the commercial engine compounding into FY2027 guidance.
- Security for AI bookings "+50% sequentially" and pipeline "+75%" (call) are explicit product-ramp signals tied to a new logo motion.
- Record 24% non-GAAP operating margin and 23% FCF margin show cash and operating leverage even while GAAP loses money.
- Deferred revenue $2.93B and RPO ~$7.4B give multi-year visibility at ~2× ARR.
- Peer-set deferred-revenue accounts are identical across Zscaler, CrowdStrike, and Palo Alto — the comparison is literal.
Bear Case
- FY2027 ARR/revenue guidance at ~17% is a clear deceleration from +25%; the AI bookings story has to prove it can re-accelerate later, not just decorate a slower year.
- GAAP net loss widened to $63M from $41M — non-GAAP records do not retire stock-based compensation.
- FCF margin compressed four points as Q4 capex spiked; investment timing can mask underlying cash conversion.
- Red Canary adds $800M of goodwill in one year — integration and retention risk sit on the balance sheet before they sit in ARR quality.
- The release does not claim demand-exceeds-supply or ASP strength; the bull case leans on call commentary the numbers have not yet fully confirmed in deferred-revenue acceleration.
Our Take: FY2026 is a cash-flow and deferred-revenue year dressed in an AI-security narrative. Own the thesis only if you believe Security for AI bookings thicken RPO and deferred revenue faster than FY2027's ~17% ARR guide implies — the public ledger will show which side is winning every July 31.
