Snowflake Inc.
Information Technology · Application Software
Structural read: consumption-revenue data warehouse riding the AI-workload tailwind, but the platform sits between hyperscaler-native stacks ($MSFT Fabric, $GOOGL BigQuery, $AMZN Redshift) and open-table standards (Iceberg, Delta). Cortex AI brings LLM functions inside SQL; Polaris catalog opens the lakehouse layer.
Margin question is whether AI inference workloads land on $SNOW credits or get pulled to native hyperscaler runtimes.
- Product revenue growth re-accelerating on AI workloads; net revenue retention >125%
- Cortex AI + Snowpark Container Services expand TAM into ML/LLM inference
- Iceberg + Polaris position $SNOW as neutral layer above hyperscaler lock-in
- $5B+ remaining performance obligations gives multi-year revenue visibility
- Founder-mode CEO (Sridhar Ramaswamy, ex-Neeva/Google ads) refocused on AI
- $MSFT Fabric, $GOOGL BigQuery, $AMZN Redshift bundle data + AI cheaper at hyperscaler layer
- Open-table standards (Iceberg, Delta) commoditize the storage moat $SNOW historically owned
- Consumption pricing means revenue tracks customer query volume - soft macro = soft growth
- Stock-based compensation runs ~40% of revenue; GAAP profitability still distant