Databricks cost planner
What does the next year cost?
Set cloud, starting spend, workload mix, and monthly growth. The year updates as you go.
Plan inputs
Use quick inputs, import a CSV, or edit products in Advanced.
Next 12 months
Oct 26–Sep 27 · USD
About $170,258 per year. Planning range $144,719 to $204,309.
Planning range $144,719–$204,309
First month
$11,996.70
Month 12
$16,606.23
29% of the first-month bill is cloud machines.
Monthly spend
Oct 26–Sep 27 · USD
- Forecast
- Planning range
Future values are estimates. The shaded planning range reflects assumptions; it is not a calibrated probability interval.
Monthly baseline
- Pipelines and jobs$5,400
- Dashboards$4,800
- Models$1,797
- Databricks usage $8,488
- Cloud machines $3,509
Anyone with a link can read this plan. Export → JSON saves a copy for Menu → Open forecast.
What these prices include
First-month coverage: 3 priced · 0 approximate · 0 unresolved · 0 documented free
Inspect rates and assumptions
Pipelines and jobs$5,400.00
priced · Bundled public price book · 2026-10-19
View rate provenance source (opens in a new tab)105.43 nodes (104.43 workers + 1 driver) × 0.69 DBU/h × 173.3 h × $0.15/DBU from the Databricks public price card for this region.
VM $0.192/h × 173.3 h × 105.43 nodes, before spot and a VM commit.
Dashboards$4,800.00
priced · Bundled public price book · 2026-10-19
View rate provenance source (opens in a new tab)12 DBU/h × 3.3 × 173.3 h × $0.7/DBU from the Databricks public price card for this region.
Serverless SQL includes compute in the DBU rate.
Models$1,796.70
priced · Bundled public price book · 2026-10-19
View rate provenance source (opens in a new tab)Token DBUs use the model-serving list rate. Cache reads and regional processing uplifts are not included.
Reference unit costs use the installed catalog before contract adjustments. Custom rates and imported billed or effective amounts can differ.
What drives this
- 45% pipelines, 40% dashboards, 15% models.
- Growth is 42.6% per year with flat seasonality.
Price book dated 2026-10-17, catalog v2026-10-17.3. Estimates only. The calculation stays in this browser. Not a Databricks quote.
Budget and scenarios
Set a spending target, schedule changes, and compare your options.
For the 12 months shown in this forecast.
Add a budget to see forecast headroom and the first month spending could exceed your budget schedule.
A planning assumption on eligible on-demand machines. Spot machines keep their separate rate. This does not model a prepaid contract balance.
Plan launches and usage changes
Each change starts in its selected month and continues afterward. Changes to the same workload multiply together. Use −100% to retire a workload.
No planned changes yet. Add a launch, expansion, or retirement to see when it affects your forecast.
Saved scenarios
0 / 20 savedSave a baseline, change the plan, then compare. Saved scenarios stay in this browser. To back one up, open it and choose Export → JSON in the forecast. Each file saves the active plan.
Keep a starting point
Save this plan before exploring growth, budgets, or usage changes.
Commitment break-even what-if
Enter a USD allowance at list value and the eligible list spend from your contract. Eligible spend follows the monthly profile of this forecast. This separate what-if includes unused prepaid value and pay-as-you-go overage; it does not change the main plan.
No DBCU conversion, rollover, tax, or cancellation terms are inferred. Enter list spend explicitly; billed or effective net cost is not an eligible list-spend measure.
Enter an allowance and eligible monthly spend to compare the commitment with pay as you go.
Ways to pay less
Each option compares with the current plan. Savings overlap; do not add them together. Applying an option recalculates the remaining choices.
$59,510a year
Commitment discount, plus 30% off machines
Both numbers are planning assumptions until a contract says otherwise.
$44,571a year
Ask for a 37% commitment discount
A planning assumption for the Databricks portion, not an offer.
$23,477a year
Use spare cloud capacity for pipelines
About 70% of the worker machines. The driver stays on demand.