BP Cut Cloud Spend by 43% and Speed Critical Jobs From Days to ~8 Hours With Databricks SQL

BP is a 117-year-old global energy company spanning upstream exploration and production, refining, trading, retail, and growing low-carbon businesses. Its data estate evolved across classic enterprise warehouse patterns and cloud stacks split by business unit—upstream on Azure and retail on AWS. Hadoop-era systems became too slow at scale, and Redshift-based serving was costly and hard to right-size. BP standardized on Databricks SQL as its warehouse and BI serving layer across clouds, using Unity Catalog and Delta Sharing to reuse governed data without copying it. The team reduced cloud cost 43% and accelerated key processing from days to ~8 hours, while powering Power BI and other tools through Databricks SQL Serverless warehouses.