SNOW Is Up 38% Before the Bell

Snowflake came in last night and just about put the bears in a difficult position. Revenue of $1.39 billion, up 33% year-over-year. Product revenue of $1.33 billion, up 34% – accelerating from 30% in Q4 and 26% a year ago. EPS came in at $0.39 against a consensus of $0.32. And non-GAAP operating margin hit 12%, which was more than 300 basis points above what analysts had modeled at 9.2%.

That’s not a narrow beat. That’s a full-stack outperformance on every line that matters.

But the earnings weren’t even the loudest thing in the room. Simultaneously, Snowflake announced a $6 billion, five-year strategic commitment to Amazon Web Services – its largest infrastructure deal to date, and more than double its previous arrangement of $2.5 billion signed in 2023. For context, at IPO in 2020, that AWS commitment was $1.2 billion. The trajectory here is not subtle. The money goes toward AWS’s custom Graviton general-purpose chips and cloud-based GPU infrastructure for AI workloads. This is the compute stack that agentic AI runs on – not just language models answering queries, but AI systems coordinating tasks, executing workflows, and pulling from governed enterprise data environments simultaneously.

Slight tangent, but it matters: Snowflake’s pivot toward agentic AI isn’t just a positioning slide. The Natoma acquisition – an enterprise Model Context Protocol platform – tells you what they’re actually building. MCP is the plumbing that lets AI agents securely connect to enterprise tools like email, Slack, and Jira while staying inside a governed data environment. Snowflake isn’t just where your data lives anymore. They’re reaching for the governance layer on everything AI agents do inside an enterprise. That’s a different business than the one the market was pricing three months ago.

The customer numbers reinforced the story. Remaining performance obligations – essentially pre-booked future revenue – came in at $9.21 billion, up 38% year-over-year. Net revenue retention hit 126%, ticking up from 125% the prior quarter. There are now 779 customers generating more than $1 million in trailing 12-month product revenue, a 29% increase year-over-year. These aren’t vanity metrics. NRR at 126% means existing customers are spending meaningfully more, and the $9.21 billion RPO gives visibility that a pure consumption model typically doesn’t offer.

CFO Brian Robins credited Cortex Code – internally nicknamed