Choice of SQL Warehouse

[Brianben]

Hi community,

I am studying the documentation about different kind of SQL warehouse (https://docs.databricks.com/aws/en/compute/sql-warehouse/warehouse-types#:~:text=A%20classic%20SQL%2....)

I saw that the serverless SQL warehouse is better than that of classic SQL warehouse in several aspect. May I know in what situation is it more suitable to choose classic SQL warehouse? 

May I know is there any difference in between if there are some PowerBI dashboard that connect to Databricks? 

Please help clarify for use, thanks in advance. 

[Advika]

Hello @Brianben!

Classic SQL warehouses are better for cost-sensitive 24/7 workloads, stable query patterns, and older workflows that depend on traditional data warehouse setups or external Hive metastores. They also allow some manual configuration, giving more control over infrastructure.
On the other hand, serverless warehouses offer better performance, faster startup, and autoscaling. They are fully managed by Databricks, meaning less direct control.

For Power BI, serverless is generally the better choice due to quicker response times and better concurrency. However, a classic warehouse might be more cost-effective if you need a steady connection for continuous report updates.

For more insights, you can also check out this post: How to integrate Databricks and Power BI for real-time streaming analytics.