The Numbers
Snowflake added $158 million in net new product revenue last quarter. That is the largest sequential add in company history, and it happened at a $6 billion run rate with 37% year-over-year growth. Three quarters ago, the sequential add was $68 million. The stock moved from $306 to $370 overnight.
Product revenue guidance went from $5.84 billion to $6.07 billion for the full year. That is a $230 million raise in a single quarter. The implied full-year growth rate jumped from 31% to 36%. Re-acceleration at this scale almost never happens.
The AI Bet
CEO Sridhar Ramaswamy said roughly half the acceleration came from AI products (the CoCo coding agent, CoWork knowledge agent, and Cortex AI Gateway), and the other half came from those products pulling more core platform consumption behind them. CoCo is now in 9,100 accounts, adding 2,000 in the quarter. CoWork hit 5,800 accounts. Total customer count is 14,554, which means 60% of the base is touching the AI agent.
The pricing decision matters for comp: AI consumption runs on the same meter customers already have. No separate SKU, no separate sales motion, no split quota. Every AI query runs on data already in Snowflake, so the AI dollar and the core dollar arrive together. That makes quota credit cleaner, but it also means AI is not a standalone upsell motion for AEs.
The Margin Trade
Snowflake deliberately dropped non-GAAP product gross margin from 76% to 75%, and guided the full year to 74%. CFO Brian Robins was direct: AI workloads carry lower contribution margin because Snowflake buys the inference. The company raised operating margin guidance on the same call, which means they are betting the volume makes up for the per-unit hit.
This is the same wall Figma hit (five-point margin drop on AI credits) and Canva ran into on its free tier. Box held the line at 81% because it sells the content layer and does not buy inference. Snowflake buys inference, so margin compresses.
GTM Context
Snowflake has turned over the CRO role twice in 14 months: Mike Gannon in March 2025, Jonathan Beaulier in March 2026. That is a lot of change at the top of a sales org posting record quarters. NRR sits at 126%, and only 49 net new customers crossed $1M in the quarter, which signals the growth is coming from expansion in the existing base, not new logo volume.
RPO (remaining performance obligations) grew 30%, but current RPO grew 42%, which means customers are committing to shorter contract durations with higher near-term spend. That is consistent with consumption-based pricing: enterprises are paying for what they use rather than locking in multi-year commits.
What This Means for Sales Teams
If you are selling into the cloud data platform space, Snowflake just validated that AI agents can drive consumption at scale without splitting the sales motion. If you are looking at enterprise AE roles, the comp here ties to platform consumption plus AI pull-through, not discrete AI deal sizes. And if you are tracking ANZ tech, Snowflake has 79 employees across Auckland, Sydney, Melbourne, Brisbane, Perth, and Canberra, with customers like Canva, which means the regional footprint is real but still focused.
The guide implies deceleration from here: Q3 revenue is expected to add about $99M sequentially, and Q4 drops to $63M, putting Q4 growth back near 35%. Snowflake has beaten and raised every quarter this year, so treat the guide as conservative. But the company is not forecasting 37% to hold.