If your Power BI reports require near real-time data, Serverless SQL can be convenient but expensive at scale. To reduce costs, optimize queries, use aggregation tables, limit data scanned, enable caching, and avoid overly frequent Power BI refreshes. You can also create materialized views or Delta tables with pre-aggregated data so Power BI reads less data. As an alternative, consider using a small all-purpose or SQL warehouse cluster with auto-scaling and auto-stop enabled, which may be more cost-effective for predictable workloads. The best option depends on your query volume, concurrency, and latency requirements, so monitoring Databricks cost reports is essential to identify the biggest cost drivers.