List prices as of Oct 17, 2026 12:57 UTC
Serverless vs. classic jobs compute
Compare the cost of a scheduled job on a classic jobs cluster (DBUs + VMs + startup/idle time) with serverless jobs compute, using current list $/DBU for your cloud and region.
📘 Guide: Databricks serverless vs classic compute: which is cheaper?
Classic cost per month
—
How it works: classic cost = nodes × (instance DBU/h × list $/DBU + VM $/h) × (runtime + billed startup/idle). Serverless cost = serverless DBUs × serverless $/DBU. The VM price is already included in serverless. We can't know how many DBUs serverless will use for your job, so we show the break-even. VM prices are auto-filled with on-demand Linux list prices (Azure pricing feed, AWS EC2 on-demand pricing, Google Compute Engine pricing pages); type your own price to override, e.g. for discounts or spot. List prices only: no committed-use discounts, spot, or promotions.
Get the monthly pricing digest
Every Databricks price change, once a month. Free.
We only use your email for the monthly digest. Unsubscribe anytime with one click. Privacy
Sponsor slot available. Reach data engineers and FinOps teams while they're actively comparing Databricks costs. Learn more →