Pay as you go with per-second usage-based billing; public tier list prices for Databricks SQL are not fully disclosed on the pricing pages.
↻ Billing notes
Databricks says pricing is pay-as-you-go with no up-front costs and per-second granularity. The pricing pages also mention a 14-day free trial, committed-use discounts, and custom requirements by contacting sales. For Azure Databricks, Databricks says pricing is set by Microsoft and is provided for convenient reference. The public documents do not disclose a universal seat minimum, renewal schedule, or a single monthly billing cadence for Databricks SQL.
Pay-as-you-go access to Databricks SQL and related platform services
Serverless SQL warehouse experience
Open standards and APIs
Unified governance model
Tools of choice with no lock-in
Public pricing pages reference a detailed price list by cloud and SKU group, but do not expose a complete public tier-by-tier dollar schedule for Databricks SQL in the provided documents.
Committed-use / custom requirements
Not publicly disclosed
Not publicly disclosed
Discounts for committed usage
Potentially flexible commitments across multiple clouds
Custom requirements handled by sales
Eligibility, discount levels, and final terms are quote-based and not published in the provided documents.
Databricks says customers can contact the company for committed-use discounts. The documents do not disclose a numeric discount rate or a fixed add-on price.
Cloud-specific DBU rates and SKU groups
Not publicly disclosed
Databricks pricing is tied to cloud-specific price lists and SKU groups. The provided documents mention DBU calculators and cloud-specific matrices, but they do not disclose a single public add-on price for these components.
⚠ Databricks pricing is consumption-based, so costs can rise with compute usage, runtime, and cluster configuration. The cost-management blog explains that a Databricks Unit (DBU) is the underlying unit of consumption and that SQL warehouses sum the DBU rates of the clusters making up the endpoint. It also notes that cloud instance types each have different DBU rates, which means infrastructure choices can materially change spend.
⚠ The same blog warns that unmanaged compute can lead to spiraling cloud costs and that overly restrictive policies can backfire by making jobs run longer and cost more. In practice, buyers should expect usage spikes from warehouse concurrency, cluster sizing, instance selection, and workload type differences rather than from a simple flat subscription fee.
No flat public monthly price is disclosed in the provided official Databricks pricing pages. Instead, Databricks presents pricing as pay-as-you-go and points buyers to a detailed price list and cloud-specific references. For buyers, that means the final cost depends on usage, cloud, and SKU selection rather than a simple sticker price.
Yes. The official pricing pages say you can start free and specifically mention a 14-day free trial. That makes it possible to evaluate Databricks SQL before committing to ongoing usage. After the trial, ongoing charges follow the usage-based pricing model.
Yes, but the details are not fully public. Databricks says customers can use committed-use contracts to access discounts and other benefits, and it also notes that some promotional discounts may be communicated to eligible customers. Because the documents do not publish a universal discount schedule, most discounting should be treated as quote-based.
The biggest cost drivers are DBU usage, warehouse sizing, cloud instance selection, and workload behavior. The cost-management blog explains that SQL warehouses are a group of clusters and that policies, auto-scaling, and auto-termination can materially affect spend. Buyers should model usage carefully because the same workload can cost more or less depending on compute choices and governance settings.