CrowdStrike Revenue Grew 29% to $3.95B in Its Outage Year — Net Retention Fell 7 Points
The fiscal year containing the July 2024 global IT outage closed at $3.95 billion in revenue, up 29%. The real scar is in ARR: net retention fell from 119% to 112%, and net new ARR declined for the first time.
On this page
At 04:09 UTC on 19 July 2024, CrowdStrike shipped a routine content update to the Falcon sensor for Windows. Channel File 291 carried a logic error — CrowdStrike’s own external root-cause analysis, published 6 August 2024, traced it to an IPC template type that defined 21 input fields against sensor code supplying 20 — and machines running the sensor blue-screened on boot. Microsoft’s David Weston put the count at roughly 8.5 million Windows devices, under 1% of Windows machines worldwide. Airlines, hospitals, broadcasters and payment networks stopped. Delta Air Lines cancelled more than 7,000 flights over five days and later sued CrowdStrike over it.
The fiscal year containing that day closed at $3.95 billion in revenue, up 29%. Which is the most misleading true sentence you can write about CrowdStrike, and the reason this post exists: in a subscription business at this scale, the revenue line is the last place a shock shows up, and the smallest place it shows up at all.
The thesis. CrowdStrike did not lose customers over the outage. It lost expansion. Gross retention held at 97%; dollar-based net retention fell from 119% to 112%, and net new ARR declined year over year. The subscription model absorbed a globally televised failure and converted it into seven points of net retention.
The revenue line never bent
Recognised subscription revenue in any quarter is mostly the amortisation of contracts signed months or years earlier. That makes it a beautiful shock absorber and a terrible early-warning system. CrowdStrike’s audited revenue series, straight from its 10-K filings, shows no discontinuity at all where the outage sits:
| Fiscal year (ends 31 Jan) | Revenue | YoY growth |
|---|---|---|
| FY2019 | $249.8M | +110% |
| FY2020 | $481.4M | +93% |
| FY2021 | $874.4M | +82% |
| FY2022 | $1.452B | +66% |
| FY2023 | $2.241B | +54% |
| FY2024 | $3.056B | +36% |
| FY2025 | $3.954B | +29% |
Source: CrowdStrike 10-K filings via SEC XBRL company facts. FY2019 growth is measured against fiscal 2018 revenue of $118.8M. The July 2024 outage falls in the second quarter of FY2025, the last row.
Deceleration from +110% to +29% over seven years is the ordinary arithmetic of a bigger base, not a wound. Subscription revenue specifically grew 31% in fiscal 2025 to $3.76 billion, per the 4 March 2025 results release. Free cash flow hit a record $1.07 billion on $1.38 billion of operating cash flow. GAAP swung to a small net loss of $19.3 million for the year — against $89.3 million of net income in fiscal 2024 — but non-GAAP net income rose to $987.6 million from $751.8 million. Nothing in that paragraph looks like a company that broke the internet.
What actually broke was ARR
Annual recurring revenue is the leading indicator, because it re-prices the whole book to today’s contract terms rather than yesterday’s. That is where the outage is visible. Ending ARR reached $4.241 billion at 31 January 2025, growing 23%, against $3.435 billion growing 34% a year earlier. The dollars added tell it more bluntly: net new ARR was $806.7 million in fiscal 2025 versus $875.5 million in fiscal 2024 — the company added less new recurring revenue in absolute terms than the year before, off a base a third larger.
7 pointsof dollar-based net retention lost between January 2024 and January 2025 — 119% to 112%
97% at the fiscal 2025 close and again in Q1 FY2026, per both results releases. The fiscal 2025 10-K states the company “maintained high dollar-based gross retention rates following the incident.” Enterprises did not rip out the endpoint agent.
112% as of 31 January 2025, down from 119% a year earlier. Every point of that gap is expansion that did not happen, or contracted, inside accounts that stayed.
Why net retention falls without churn
Dollar-based net retention = current-period ARR from last year’s customers ÷ prior-period ARRCrowdStrike’s definition, fiscal 2025 10-K: renewals, expansion, contraction and churn, excluding new customers and excluding incident-response and proactive services.
The mechanism has a name in the filings. CrowdStrike offered what it calls customer commitment packages, and the 10-K lists exactly what went into them:
- Discounting — straight price concession on renewal.
- Additional modules — more of the platform, at no incremental ARR.
- Professional services — delivery cost absorbed by the vendor.
- Flexible payment terms — cash timing traded for goodwill.
- Subscription period extensions — the same contract value stretched over more months.
The last one is the trap, and it is the item most operators would not have modelled. ARR annualises contract value; extending a customer’s term without raising the total spreads the same dollars across a longer period, so the annualised figure goes down. The 10-K says so in as many words: the packages “have resulted, and are expected to continue to result, in increased contraction, due to elongated subscription terms, and decreased upsell dollar values.” A goodwill gesture that costs nothing in cash costs real points of net revenue retention, and it keeps costing them for the length of the extension.
The direct cash bill, by contrast, was almost trivial at this scale. The fiscal 2025 10-K accrues $60.1 million of July 19 Incident expenses net of insurance receivable, booked across sales and marketing ($21.4M), research and development ($6.8M) and general and administrative ($31.9M). That is 1.5% of a year’s revenue. The expensive part of the incident was never the invoice; it was the growth rate.
A note on our own numbers
Where GetLatka’s row disagrees with the filings
Our database records $3.44 billion of revenue for CrowdStrike, captured December 2024, and $2.56 billion captured January 2023. Both figures are real and both are mislabelled: $3.435 billion was CrowdStrike’s ending ARR at 31 January 2024, and $2.56 billion its ending ARR at 31 January 2023. Reported revenue for those years was $3.056 billion and $2.241 billion. The older row in our series, $874.4 million captured April 2021, matches fiscal 2021 revenue exactly — so the set is a mix of two different metrics. Trust the filings. It is a useful reminder that run-rate and recognised revenue diverge by 10–15% at a company like this, and that the gap is mostly timing, not accounting mischief. Where our data does hold up: GetLatka’s headcount series records 10,342 employees as of July 2025, against 10,118 full-time employees in the 10-K at 31 January 2025, and a pre-IPO funding ladder totalling roughly $1.09 billion — a $26M Series A in February 2012 at a $24.85M valuation through a $200M Series E in June 2018 at $2.8 billion, before the June 2019 NASDAQ listing. The full tracked record sits on CrowdStrike’s GetLatka profile.
One year on
The first quarter of fiscal 2026, reported 3 June 2025, is the cleanest read available on whether the damage is structural. Revenue was $1.10 billion, up 20%, against $921.0 million a year earlier. Ending ARR grew 22% to $4.44 billion at 30 April 2025. Gross retention stayed at 97%. Cash from operations set a record at $384.1 million, and the board authorised a share repurchase of up to $1 billion.
But net new ARR in the quarter was $193.8 million, against $212 million in the same quarter a year before — fewer new recurring dollars, on a base 22% larger, three quarters after the incident. The company’s answer is Falcon Flex, a flexible-commitment structure that accounts had used for more than $3.2 billion in total deal value by the end of Q1 FY2026, growing more than 6x year over year; management pointed to it when guiding to net new ARR re-acceleration in the second half of the year.
We started the fiscal year with record Q1 large deal and MSSP momentum alongside sustained 97% gross retention and consistently strong net retention as the market consolidates on Falcon as its cybersecurity platform of choice for the agentic AI era.
George Kurtz, founder and CEO, CrowdStrike
Full-year fiscal 2026 guidance, issued the same day, sits at $4.744–4.806 billion of total revenue against a stated long-term goal of $10 billion in ending ARR. On the two figures CrowdStrike reported for fiscal 2025 — 29% revenue growth, and $1.07 billion of free cash flow on $3.954 billion of revenue, a 27% margin — the company cleared the Rule of 40 by sixteen points in the year of the outage, by our arithmetic on its own reported numbers.
And yet the incident is not closed in the accounts. The guidance table in CrowdStrike’s 3 June 2025 release still carries, among the items excluded from non-GAAP results, the line: “costs (recoveries) associated with the July 19 Incident and related matters, net.” A single Friday morning in July has its own permanent name in this company’s income statement, and the filings still list it as an open risk.
SourcesCrowdStrike Q4 and full-year fiscal 2025 results release, 4 March 2025; CrowdStrike Form 10-K for fiscal 2025 (filed March 2025); CrowdStrike Q1 fiscal 2026 results release, 3 June 2025; CrowdStrike Q4 fiscal 2023 results release, 7 March 2023; CrowdStrike Q1 fiscal 2025 results release, 4 June 2024; annual revenue series from SEC XBRL company facts for CIK 0001535527; CrowdStrike external technical root cause analysis of Channel File 291, 6 August 2024; Microsoft corporate blog, 20 July 2024; Delta Air Lines statements and reporting on its October 2024 suit; CrowdStrike’s GetLatka profile for headcount and pre-IPO funding.
