Before any large language model answers a question inside an enterprise, someone has to store, govern, and serve the underlying data. That unglamorous infrastructure layer is where Snowflake competes, and the fiscal second quarter ended July 31 made the strongest case yet that the market has been underpricing how central that position becomes as AI workloads scale.
The Market Context
Enterprise AI spending has moved from proof-of-concept to production. That transition has a predictable consequence: the data platforms sitting beneath production workloads start generating real consumption revenue, not just contract signings. Snowflake’s product revenue reached $1.33 billion in its fiscal first quarter, up 34% year over year, a notable acceleration from 30% in the previous quarter. The second quarter extended that sequence. Product revenue came in at $1.49 billion, representing 37% year-over-year growth, marking a third straight quarter of acceleration.
The Research
Total revenue reached $1.55 billion, up 35% year on year, while adjusted EPS of $0.62 per share came in well ahead of analysts’ consensus estimates of $0.45. That earnings gap is not the kind of number that comes from favorable seasonality. Non-GAAP operating income hit $237 million, producing a 15.3% margin.
Non-GAAP operating income reached $237 million, producing a 15.3% operating margin, compared with 11.1% a year earlier. That 420-basis-point expansion reflects disciplined cost structure alongside genuine revenue growth, which is not common in software companies still reinvesting aggressively in AI capacity. Net revenue retention stood at 126%, while remaining performance obligations reached $9.00 billion, representing 30% year-over-year growth.
Management pointed to growth in its AI coding agent named CoCo as a key driver of consumption, alongside CoWork. These products are increasing spend per existing customer rather than simply attracting new accounts, which is the mechanism behind the retention rate and the margin expansion simultaneously.
The Hidden Insight
The second-order implication that deserves attention is what AI is doing to Snowflake’s switching costs. The proliferation of AI coding tools is making it easier for enterprises to migrate from legacy data platforms to modern ones like Snowflake, with migrations that previously required months of engineering work now being compressed. Faster migration lowers the barrier to entry, but it also locks customers into the data governance layer more quickly. Once production AI workloads run inside a governed Snowflake environment, the cost of moving them rises sharply. The platform is capturing customers at an earlier stage of commitment and deepening that commitment faster than prior product cycles allowed.
Investment Opportunities
For fiscal Q3, management guided for $1.588 billion to $1.593 billion in product revenue, raised its full-year product revenue forecast to $6.07 billion from $5.84 billion in May, and projected a 14.5% non-GAAP operating margin, wider than the 13.5% figure it had previously guided.
Wall Street responded accordingly. Goldman Sachs raised its price target to $436 from $300 and maintained a Buy rating following the better-than-expected Q2 results. Argus moved to $450, and Wells Fargo set a street-high figure at $525. According to aggregated analyst data, the average 12-month price target for SNOW is closer to the low $400s, with the consensus rating in the Buy to Strong Buy range.
Risks and Counterarguments
The counterarguments deserve honest weight. Insiders sold about $353 million in shares over the past 90 days following the report. Management also trimmed its full-year non-GAAP product gross margin outlook to 74% due to AI workload mix. Competition from Microsoft’s Fabric platform represents a structural threat that a single strong quarter cannot resolve.
Research Conclusion
Three consecutive quarters of accelerating product revenue, a 126% retention rate, margins expanding at the same pace as growth, and guidance that cleared Street expectations on every line, all at a stock price roughly 13% below the post-earnings high. The fundamental case has not weakened since September 2, 2026. What investors should watch between now and the December 2, 2026 earnings report is whether CoCo and CoWork consumption held through October, and whether hyperscaler competitive activity intensified. Those two data points will determine whether the current gap between SNOW’s price and analyst targets closes from the top, or whether it was the market correctly staying cautious.
