Snowflake is a consumption-priced data cloud that is now trying to become the governed operating system for enterprise AI agents, and the second-quarter print is the first clean evidence that the AI layer is changing the growth rate rather than just the marketing. Product revenue rose 37% from a year earlier, the third straight quarter of acceleration, and management framed the step-up as coming from both the core warehouse and a meaningful lift in AI workloads. The investment debate is whether that acceleration is a durable consumption multiplier or a mix shift the market is already capitalizing as if it were permanent.
The tension sits in the gap between a raised full-year product-revenue outlook of $6.1 billion and a still-unprofitable GAAP income statement that management itself dates to a first profit in late fiscal 2028. Non-GAAP operating margin, the adjusted operating profit that strips stock-based pay and acquisition items, reached 15% in the quarter. The company lifted the full-year operating-margin guide even as it cut the product-gross-margin outlook on a heavier AI mix. Stock-based compensation remains large enough to keep GAAP results deeply negative, and second-quarter free cash flow is still a mid-single-digit slice of revenue against a full-year adjusted-cash-flow guide that assumes a back-half billing season.
Remaining performance obligations, the contracted revenue not yet recognized, totaled $9.0 billion and grew slower than product revenue, a reminder that consumption, not bookings, is the real engine. Net revenue retention, the expansion of an existing customer cohort after churn, held at 126% as large accounts added AI workloads on top of migrations. The next several quarters resolve whether AI keeps compounding platform consumption without permanently cheaper mix, and whether cash conversion catches the profit-and-loss story.