The most mature FinOps teams running data platforms aren’t just
tracking cloud spend — they’re governing across compute hours, DBUs,
tokens, and provisioned throughput as a unified whole. In this talk,
I’ll share what we’ve seen the most successful Databricks customers do
differently: how they structure budgets across heterogeneous
workloads, how they allocate costs by team and business unit despite
fragmented meters, and how they turn cost visibility into operational
discipline. If you’re responsible for governing a modern data
platform, this is what good looks like.