Palo Alto Networks (PANW) FY2026 10-K: CyberArk, $24.4B of Goodwill, Negative Tangible Book
Listen to this report
Of the $24.36B Palo Alto Networks paid for companies in fiscal 2026, $24.34B landed in two line items nobody can touch, count or ship. Tangible book value is now negative $1.5B, and next year's scheduled amortization is 1.6 times this year's entire operating income.
RoboSystems initiating coverage · Computer peripheral equipment (SIC 3577) · Source filing: FY2026 Form 10-K for the year ended July 31, 2026, filed September 10, 2026 (accession 0001327567-26-000023, CIK 0001327567). Palo Alto Networks' fiscal year ends July 31, not December: FY2026 covers August 1, 2025 to July 31, 2026, so every period label below is a July year and is not comparable to a December filer's FY2025 without adjustment. All figures read directly from the filing's XBRL via RoboSystems unless attributed otherwise. Not investment advice. No price target.
1. The Hook
Palo Alto Networks $PANW closed four acquisitions in fiscal 2026 for $24.36B of total consideration, more than twice the $11.48B of revenue it earned that year. Here is where the money went, from the company's own purchase price allocations:
| FY2026 acquisition | Closed | Consideration | Goodwill | Identified intangibles |
|---|---|---|---|---|
| CyberArk Software Ltd. | 2026-02-11 | $21,061M | $14,800M | $6,279M |
| Chronosphere, Inc. | 2026-01-29 | $2,951M | $2,364M | $565M |
| Koi Security Ltd. | 2026-04-14 | $231M | $169M | $35M |
| Portkey, Inc. | 2026-05-29 | $117M | $110M | $15M |
| Total | $24,360M | $17,443M | $6,894M |
Goodwill plus intangibles is $24,337M of a $24,360M bill. 99.9%. Twenty-three million dollars of a twenty-four billion dollar year bought something you could put on a balance sheet and point at.
That is not an accusation, it is what buying software companies looks like. The interesting part is what it did to the balance sheet it landed on. Goodwill went from $4.57B to $22.01B. Intangibles went from $763M to $7.02B. Together they are $29.03B, which is 59.9% of total assets and, more to the point, more than the entire $27.49B of shareholders' equity. Subtract the estimates from the book and what is left is negative $1.54B. Every dollar of Palo Alto Networks' book value, and then about a billion and a half more, is an accounting judgment about the future.
The filing also says, in one tagged sentence most readers will never reach, what those judgments have produced so far: since their acquisition dates, CyberArk and Chronosphere combined contributed revenue of $930M and an operating loss of $797M.
2. Company Snapshot
Palo Alto Networks is a Santa Clara, California cybersecurity company that sells network security, cloud security and security operations as a platform: firewalls and the subscriptions attached to them, plus Cortex, Prisma and, as of February 2026, CyberArk's identity security. It reports one operating segment and one reportable segment, so there is no segment table to hide in. It sells almost entirely through a two-tier channel, and the concentration is real: two distributors each represented 15% of total revenue in FY2026, 30% combined, and one distributor was 19% of gross accounts receivable at year end. No single end customer reached 10%.
Revenue is increasingly the recurring kind, which is the good news in this document.
| Revenue line | FY2026 | FY2025 | FY2024 | FY26 growth |
|---|---|---|---|---|
| Product | $2,280M | $1,802M | $1,603M | +26.5% |
| Subscription | $6,239M | $4,974M | $4,188M | +25.4% |
| Support | $2,961M | $2,445M | $2,236M | +21.1% |
| Total revenue | $11,480M | $9,221M | $8,027M | +24.5% |
Geographically it is a US company with a European business attached: United States $7,108M, other Americas $571M, EMEA $2,428M, Asia Pacific and Japan $1,373M. Remaining performance obligations stood at $21.2B, of which $9.3B is expected to be recognized in the next twelve months. Note also that the company effected a two-for-one stock split on December 12, 2024, so any pre-2025 per-share figure you compare against needs halving first.
3. The Financial Story
Revenue accelerated and every profit line went the other way. Growth rose from 14.9% to 24.5%. Gross margin fell from 74.3% to 73.4% to 70.4%. Operating income fell 44% to $695M. Net income fell 73% to $307M. Diluted earnings per share went from $3.64 to $1.60 to $0.40.
| FY2026 | FY2025 | FY2024 | |
|---|---|---|---|
| Revenue | $11,480M | $9,221M | $8,027M |
| Gross profit | $8,077M | $6,770M | $5,968M |
| Gross margin | 70.4% | 73.4% | 74.3% |
| Operating income | $695M | $1,243M | $684M |
| Operating margin | 6.1% | 13.5% | 8.5% |
| Other income (expense), net | $(159)M | $353M | $304M |
| Income tax charged | $229M | $462M | $(1,590)M |
| Net income | $307M | $1,134M | $2,578M |
| Diluted EPS | $0.40 | $1.60 | $3.64 |
| Operating cash flow | $4,553M | $3,716M | $3,258M |
Two of those columns need a warning label before anyone draws a three-year trend line.
The first is FY2024's $2,578M of net income, which is not an operating result. It contains a negative $1,590M income tax line, a benefit rather than a charge, and the filing says why: the company released the valuation allowance on its US federal, non-California state and UK deferred tax assets. The FY2024 effective tax rate was negative 160.8%, and the valuation allowance release alone was worth negative 341.9 percentage points of it. In the tax note the item shows up as a $2,173M deferred foreign benefit. Strip it out and FY2024 pretax income was $988M, which is a normal year. A chart running $2,578M to $1,134M to $307M tells a story of collapse that did not happen.
The second is FY2026's $(159)M of other income (expense), net, down $512M from a positive $353M. That swing is almost half of the entire $1,060M decline in pretax income, and it has nothing to do with whether the acquisitions were good. Inside it sits a $620M non-cash loss on the change in fair value of the 2030 convertible senior notes that Palo Alto Networks assumed from CyberArk and elected to carry at fair value. Those notes convert at $280.75. The stock is far above that. The $620M charge exists because the share price went up. It is a bull signal being reported as a loss, and it is also non-deductible, which is why the effective tax rate jumped to 42.7% from 28.9%: the fair-value change added 22.0 points to the rate on its own.
What is left after those two adjustments is a real, explainable operating decline. Operating expenses grew $1,855M against $1,307M of gross profit growth. The drivers are all acquisition mechanics:
- Amortization of purchased intangibles rose to $640M from $166M. $418M of it runs through cost of revenue, which is most of the three-point gross margin decline, and $222M through sales and marketing.
- General and administrative more than doubled, up 103% to $899M, on accelerated vesting of equity awards tied to the deals, CyberArk severance and acquisition costs. CyberArk acquisition-related expense alone was $56M.
- Share-based compensation reached $1,815M, 15.8% of revenue and 2.6 times operating income. $177M of that came from CyberArk and Koi replacement and accelerated awards.
Now the part where the two versions of this company diverge completely. Operating cash flow rose 22.5% to $4,553M, or 39.7% of revenue, the third straight year at roughly 40%. Free cash flow after $440M of capital expenditure was $4,113M, a 35.8% margin. Against $307M of net income that is a 14.8x gap. Someone reading the income statement and someone reading the cash flow statement are describing different businesses, and the cash reader has the better of it, because the gap is almost entirely non-cash charges that are real accounting but not real outflows: $1,815M of stock compensation, $855M of depreciation and amortization, $620M of convertible note revaluation.
The stock-versus-cash framing needs one correction, though. It is commonly said that this was paid for in stock. Two-thirds of it was: equity consideration was $18,862M of non-cash issuance, and shares outstanding went from 668M to 815M, up 22.0% in a single year, with 112M shares issued for CyberArk alone. But the company also wrote a genuinely large cheque. Cash paid for business acquisitions, net of cash acquired, was $4,663M - which is more than the entire $4,553M of operating cash flow the company generated that year. Cash on hand barely moved, from $2,269M to $2,514M, because the investment portfolio funded the difference: the company sold and matured $5,693M of investments against $3,694M of purchases, a net $2.0B liquidation, and the filing's own MD&A attributes part of other income to "gains on sales of our investments to fund acquisitions." Add $1,000M of buybacks (7M shares at a weighted average $147.70) and the year consumed every dollar of operating cash and drew down the securities account on top.
Two forward obligations deserve to be on the page. First, the amortization schedule, which is not a forecast but a fixed charge the filing already tabulates:
| Scheduled amortization of intangibles | Amount |
|---|---|
| FY2027 | $1,081M |
| FY2028 | $989M |
| FY2029 | $908M |
| FY2030 | $892M |
| FY2031 | $667M |
| FY2032 and thereafter | $2,480M |
| Total | $7,017M |
FY2027's $1,081M is 1.56 times FY2026's entire $695M of operating income, and it is larger than FY2026's $640M of amortization by $441M. GAAP operating income does not get easier next year by default; it gets harder by roughly $441M before anything else happens. Second, $8,242M of unrecorded purchase commitments, of which $7,686M is cloud capacity running out past 2032, against $2,514M of cash.
And one thing that has not happened. Palo Alto Networks tests goodwill for impairment annually in the fourth quarter, which for FY2026 ended July 31, 2026, after CyberArk closed. The filing states plainly: "We did not recognize any impairment losses on our goodwill, intangible assets, or other long-lived assets during the years ended July 31, 2026, 2025, and 2024." The spree is in this document. The write-down is not, and may never be. That is the honest state of the evidence.
4. Valuation: What It Is Worth as a Normal Business
At $373.94 per share on September 15, 2026 (market price, not from the filing) against the 818M shares outstanding the 10-K cover reports as of August 31, 2026, the market capitalization is $305.9B. That share count matters: secondary sources were quoting market caps between $213B and $306B on the same day because of inconsistent handling of the post-CyberArk share count, and the filing's own cover page settles it. Net of $7.9B of cash and investments and $1.8B of convertible notes, enterprise value is roughly $299.8B.
| Multiple | FY2026 basis | Value |
|---|---|---|
| Price to sales | $11,480M revenue | 26.6x |
| Enterprise value to revenue | $11,480M revenue | 26.1x |
| Price to earnings (GAAP) | $307M net income | 996x |
| Enterprise value to free cash flow | $4,113M FCF | 72.9x |
| EV to FCF less stock compensation | $2,298M | 130.4x |
| Free cash flow yield | $4,113M FCF | 1.34% |
Scenario discounted cash flow. Assumptions stated plainly: a five-year explicit free-cash-flow growth period off the FY2026 base, then a terminal value, with net cash and investments added and convertible notes deducted. Bear uses 6% growth, a 10.0% discount rate and 2.5% terminal growth; base uses 13%, 9.0% and 3.0%; bull uses 20%, 8.5% and 3.5%. The table runs twice, because whether stock compensation is a real cost is the single largest judgment call in this name.
| Scenario | Implied value, FCF of $4,113M | Implied value, FCF less stock comp of $2,298M |
|---|---|---|
| Bear | $87 per share | $52 per share |
| Base | $139 per share | $81 per share |
| Bull | $214 per share | $123 per share |
| Current price | $373.94 | $373.94 |
Peer and cross-sector re-rating. Applying a range of enterprise-value-to-free-cash-flow multiples to the FY2026 result, 30x implies $158 per share, 45x implies $234, 60x implies $309 and 75x implies $385. The stock trades at 72.9x today, so the market is already paying a multiple at the very top of the large-cap software range, on cash flow that is flattered by a $1.8B stock-compensation add-back.
What today's price implies. Solving the base-case discount rate and terminal growth for the free-cash-flow growth that justifies $305.9B, the market is pricing a 40.7% compound annual free cash flow growth rate for five years on the unadjusted figure, or 58.8% if stock compensation is charged as the cash-equivalent cost it economically is. Free cash flow grew 18.6% in FY2026. Framing: these are implied values under the stated assumptions, not price targets and not advice. They exist to say what a buyer at today's price is assuming, which is that the acquisitions work, quickly and completely.
One more market fact worth stating, because it cuts against everything above. In the CyberArk exchange on February 11, 2026, Palo Alto Networks' own shares were valued at $165.07 each ($18,488M of stock for 112M shares). At $373.94 the stock is 2.27 times that level, seven months later. The market has already looked at this balance sheet and voted, emphatically, that the estimates are good. That is not a reason to agree with it. It is a reason to be precise about which side of the argument the evidence is actually on.
5. Risks
The acquisitions are not yet paying. The filing's own tagged numbers say CyberArk and Chronosphere together produced $930M of revenue and a $797M operating loss since their acquisition dates. The pro forma table is starker: had both deals closed at the start of each period, FY2026 revenue would have been $12,312M with a $114M net loss, and FY2025 would have been $10,486M with a $37M net loss. On that basis revenue grew 17.4%, not 24.5%, and the company lost money in both years. A material part of the headline growth rate is acquired, not earned.
The demand signal is softer than the revenue line. Total deferred revenue rose 15.7% to $14,756M while revenue rose 24.5%, and $776M of that increase was CyberArk's deferred revenue balance acquired on day one rather than billed. Deferred revenue is the leading indicator for this model and it is growing more slowly than the lagging one. Working capital is a deficit of $1,280M, widened from $465M.
The balance sheet has no cushion under the estimates. With goodwill and intangibles at $29.03B against $27.49B of equity, a goodwill impairment of roughly 12% would erase tangible book entirely on paper, and the CyberArk goodwill alone is $14.8B, or 70.3% of what was paid for it. The company's own risk factors name the exposure directly, flagging that it may fail to realize anticipated synergies or accurately forecast an acquisition's financial impact, "including accounting charges and any potential impairment of goodwill and intangible assets," and it names CyberArk and Chronosphere specifically.
Dilution is the funding mechanism, so the share count is a live variable. Shares rose 22.0% in one year. The $1.0B buyback retired 7M shares against 147M issued. Stock compensation is running at $1,815M a year, 15.8% of revenue.
Concentration and the tax reserve. Two distributors are 30% of revenue and one is 19% of gross receivables. Unrecognized tax benefits rose to $719M from $572M. And on the tax question specifically, because it is worth checking rather than assuming: Palo Alto Networks was charged $229M of income tax in FY2026 and paid $199M in cash. Those are close, there is no gap to explain, and the distinction is drawn here only because it is the kind of thing that is usually left blurred.
6. The Bottom Line
Palo Alto Networks spent more than two years of revenue buying its way into identity security and observability in a single fiscal year, funded it with 22% dilution and every dollar of operating cash flow it generated, and booked 99.9% of the price as goodwill and intangibles. The cash engine is genuinely strong and getting stronger: $4.55B of operating cash flow, 39.7% of revenue, three years running at roughly that level. The GAAP earnings statement has, for the moment, stopped describing it, and about half of this year's collapse is a tax-benefit base effect in FY2024 and a convertible-note mark that exists because the stock went up.
What to watch is narrow and specific. First, whether deferred revenue and the $21.2B of remaining performance obligations start compounding at the revenue growth rate rather than well below it, because that is where acquired growth turns into earned growth. Second, whether the $930M revenue and $797M operating loss from CyberArk and Chronosphere converge toward the synergy case in FY2027, when both are in the numbers for a full year. Third, the fourth-quarter fiscal 2027 goodwill test, which is the first one that will have a full year of CyberArk operating results behind it rather than five months. The filing shows the spree. It cannot yet show whether it worked.
Every figure above was read out of Palo Alto Networks' FY2026 Form 10-K (accession 0001327567-26-000023) via RoboSystems, from the filing's own XBRL facts and note text, except the September 15, 2026 share price, which is market data and attributed as such. Nobody wrote this by hand, and the same pipeline reads any of the roughly ten thousand companies that file with the SEC.