Databricks Expands into White Space
Charles Chretien, co-founder of Prequel, on the modern data stack’s ROI problem
This is why Databricks has had more room than Snowflake to turn one product into a broader platform. Starting at the data processing layer means Databricks already touches raw pipelines, notebooks, ML workflows, and SQL, so each nearby product can be sold into the same engineering team and run on the same underlying engine. Snowflake started at the warehouse, which is closer to the finished analytics use case and leaves fewer natural adjacencies before it has to move upstream into harder territory.
-
Databricks built outward from managed Spark into BI and warehousing with Databricks SQL, which reached general availability on December 15, 2021. That path is downstream from compute and processing into higher level tools that use the same data and execution layer.
-
Snowflake expanded first into app building and analyst tooling through Streamlit and Cortex, but those sit close to the warehouse and mostly deepen usage inside analytics. Its 2025 Crunchy Data deal shows the next growth step is moving left into operational Postgres, not just adding more dashboards on top.
-
The revenue mix shows the practical difference. Databricks has become a multi line business with large contributions from its core platform, SQL warehousing, and AI products, while Snowflake still gets the vast majority of revenue from warehousing and has much smaller expansion lines.
The next phase is a race to own more of the stack before data even reaches the warehouse. Databricks is likely to keep pulling adjacent ML and application workloads onto its platform, while Snowflake will keep adding database and developer products so new apps are born inside its ecosystem instead of only landing there for analysis later.