CrowdStrike delivered the kind of quarter that would normally settle any argument about whether its post-outage recovery is real. Revenue reached 1.47 billion dollars, up 26 percent from a year earlier. Net new annual recurring revenue hit a record 332.8 million dollars, comfortably clearing the bar investors had been told to watch for a genuinely convincing quarter, after a smaller beat the prior quarter was followed by a stock decline of roughly 10 percent despite clearing estimates. Management responded by raising its full year net new ARR growth guidance to 34 percent at the midpoint, a jump of more than six percentage points from the figure it gave three months ago.
The market rewarded it. Shares, which had already closed the regular session up about 2 percent ahead of the print, climbed roughly 10 to 12 percent in extended trading and held there through the end of the after hours session, a stable, monotonic move with no reversal. That is a meaningfully different pattern than the one investors had been bracing for at CrowdStrike's valuation, where a beat that fails to clear an unusually high bar has, in recent quarters, been enough to send the stock lower anyway.
A Quarter That Reads Very Differently in GAAP Terms
Look past the adjusted numbers and the picture gets more complicated. Non-GAAP earnings per share came in at 31 cents, ahead of the roughly 29 cents analysts expected. GAAP diluted earnings per share was 1 cent. CrowdStrike's GAAP operating result was still a loss of 33.2 million dollars, even as non-GAAP operating income reached 371.6 million dollars, a 46 percent increase from a year earlier. The gap between those two pictures, roughly 400 million dollars, is driven overwhelmingly by stock based compensation, which rose to 399 million dollars this quarter from 276.7 million dollars a year ago, a faster increase than revenue itself.
This was still, on balance, progress. The company's GAAP net income of 5.3 million dollars marks a swing from a loss in the year ago quarter, and subscription gross margin improved on both a reported and adjusted basis. But a meaningful piece of that swing to profitability came from a specific, one-time item: CrowdStrike recorded a net recovery of 14.5 million dollars this quarter tied to costs from the July 2024 Falcon outage, a sharp reversal from a 35.7 million dollar net cost in the year ago period. That is roughly a 50 million dollar swing in the company's favor from a single reconciling item that has nothing to do with new sales or renewed customer demand. It is a legitimate result, insurance recoveries are real cash, but it means this quarter's razor thin GAAP profit should not be read as a sign that the underlying cost structure has become durably leaner.
Falcon Flex and the Case for Durability
The more encouraging signal sits in how customers are buying. Falcon Flex, CrowdStrike's consumption based licensing model, ended the quarter with 2.29 billion dollars in annual recurring revenue, more than double where it stood a year ago, and customers converting from standard subscriptions to Flex are increasing their annual spend by more than 40 percent on average. Module adoption also crept higher, with 26 percent of subscription customers now running eight or more Falcon modules, up a point from the prior quarter. Executives pointed to a newer product category, described internally as AI detection and response, as a business they believe could eventually rival the scale of the company's original endpoint detection line, a claim that is aspirational rather than something this quarter's numbers can independently confirm.
On the call, an analyst from UBS asked whether CrowdStrike had begun charging for its newer AI products on a usage or token basis rather than a flat subscription fee. Management confirmed that token based pricing already exists within the Flex licensing structure, a detail that matters because it signals CrowdStrike is willing to price its AI security products the way infrastructure providers price compute, tying revenue more directly to usage rather than seat counts. Separately, an analyst from RBC pressed on competitive positioning, and the answer, according to the available call material, emphasized CrowdStrike's proprietary threat data advantage without directly naming the company's most obvious cybersecurity competitors. Whether that reflects the actual substance of the call or a gap in the secondary material available for this article is not something this publication can resolve, but it leaves a natural question, how CrowdStrike's expanding AI security push stacks up against rivals building similar capabilities, without a direct answer this quarter.
What Comes Next
CrowdStrike's own guidance suggests confidence that this quarter's acceleration continues. The company is guiding to 6.18 billion dollars in ending annual recurring revenue for the current quarter and has raised its full year revenue, ARR, and earnings guidance across the board, on a split adjusted basis, even as the specific net new ARR growth metric saw by far the largest upward revision. That asymmetry, a big raise to the growth metric management has chosen to emphasize most, alongside smaller raises everywhere else, is itself a signal of what the company wants investors focused on.
The burden of proof from here sits less on whether CrowdStrike can grow, this quarter answered that convincingly, and more on whether growth converts into GAAP profitability that does not depend on stock based compensation staying elevated or on favorable one-time items repeating. Investors should watch whether GAAP operating results move toward breakeven and then positive without help from unusual reconciling items, whether the 34 percent net new ARR growth guide holds up as the comparisons get tougher in the back half of the year, and whether Falcon Flex conversions continue at their current pace as the easier, earlier-adopting customers are worked through. The bull case strengthens if GAAP profitability improves on its own terms in coming quarters. The bear case strengthens if this quarter's swing to profitability turns out to have been substantially a function of the outage related recovery rather than a repeatable improvement in the underlying cost structure.