Snowflake
Cloud data warehouse with separated storage and compute and per-second billing.
Overview
Snowflake made warehouse scaling an operational non-event: resize compute, pay for what runs. That elasticity is also the cost risk, since inefficient queries are billed rather than blocked.
Scored criteria
Query performance
4.6
Cost predictability
3.3
Governance
4.6
Ecosystem
4.7
Ease of operation
4.8
Value for money
3.7
Strengths
- Almost no operational overhead
- Zero-copy cloning transforms development workflows
- Best-in-class cross-organisation data sharing
Limitations
- Consumption costs escalate without query governance
- Warehouse sizing requires active management
- Machine-learning workflows trail lakehouse platforms
Features
Separated storage and computeZero-copy cloningTime travel and data recoverySecure data sharing and marketplaceSnowpark for Python and JavaIceberg table support
Verdict
The lowest-operations warehouse available. Budget for a FinOps owner from the start.
Compare Snowflake
Alternatives to Snowflake
Full list →Databricks
Lakehouse platform unifying data engineering, analytics and machine learning on open formats.
BigQuery
Serverless warehouse with no cluster management and pricing by bytes scanned.
Looker
Governed BI built on a version-controlled semantic layer.
Metabase
Open-source BI that non-technical users can actually operate without SQL.