
Azure Databricks
- Features
- Ease of use
- Ease of management
- Quality of support
- Affordability
- Market presence
- Information technology and software
- Agriculture, fishing, and forestry
- Energy and utilities
What is Azure Databricks
Managed Spark at scale
Unified analytics and ML workflows
Deep Azure integration controls
Cost management can be complex
Requires specialized platform skills
Azure-centric deployment constraints
Plan & Pricing
Pricing model: Pay-as-you-go (usage-based) Free tier/trial: 14 days of free Databricks Units via Azure free account; new Azure customers can also use the Azure free account USD 200 credit for up to 30 days. See notes below. How charges are calculated: Azure Databricks charges for (1) Databricks Units (DBUs) billed by usage (per-second granularity) and (2) underlying Azure VM compute, managed disks, storage, and other Azure resources. DBU consumption depends on workload, tier, and VM size; some features use DBU multipliers. Pricing varies by region, workload, and selected tier. Workloads / tiers (structure): Standard and Premium tiers (note: Standard tier scheduled for retirement Oct 1, 2026). Workloads shown include All-Purpose Compute, Jobs Compute, Jobs Light Compute, SQL Compute, SQL Pro Compute, Serverless SQL, Interactive/Automated Serverless Compute, Model Training, Model Serving, Database Serverless Compute, and Lakeflow Spark Declarative Pipelines. Some SKUs are only available in specific tiers. Pre-purchase / commitment options: Databricks Commit Units (DBCU) pre-purchase for 1- or 3-year terms (advertised up to ~37% savings vs PAYG) and Azure Savings Plan / Reserved VM Instances / Spot VMs can be used to reduce costs. DBCU purchase tiers (multiple DBCU quantities listed) provide tiered discounts. Storage and other billed items: Databricks Storage Units (DSU) are billed separately and certain features (data transfer, NAT gateway, Private Endpoint usage) may have additional charges; some charges are waived for Private Link data processing. Promotions / notes: Pricing pages show region- and workload-specific per-DBU and per-VM prices (users must select region/currency to view numeric rates). The pricing page also lists a limited-time promotional price (50% off) for specific serverless/server options with stated expiry dates. Example / guidance: Microsoft Learn pricing docs list DBU multipliers and example DBU-based estimates (for example, sample model-training DBU counts and an illustrative cost using a US East example DBU rate). Official per-DBU numeric rates are shown on the Azure Databricks pricing page when region/currency/workload filters are selected. Important: Official pricing pages display numeric per-DBU and per-VM prices dynamically based on region, currency, and workload. Because the vendor page requires selecting region and display options to show exact numeric rates, no universal flat per-DBU USD price is published on the static overview page. For precise dollar amounts pick your target region/currency on the official Azure Databricks pricing page or use the Azure Pricing Calculator.
(References: Azure Databricks pricing detail page and Azure free account / Databricks Units info on Microsoft Learn.)