At a glance
| SQL support | T-SQL in dedicated and serverless SQL pools, with documented differences from SQL Server |
| Storage and compute architecture | Dedicated pools: control and compute nodes over Azure Storage in 60 distributions; serverless SQL and Spark over the data lake |
| Workload types | Enterprise data warehousing, ad hoc lake queries, Spark data engineering, ETL pipelines |
| Integrations | ADLS Gen2, Power BI, Microsoft Entra ID, Azure Machine Learning, SSMS and VS Code |
| Deployment | Azure PaaS in Azure regions only |
| Pricing model | DWU-hours for dedicated pools, per TB processed for serverless SQL, vCore-hours for Spark, storage per TB-month |
Synapse is still a supported service and a reasonable place to keep dedicated SQL pools that already work well, especially for T-SQL teams. For a new warehouse, Microsoft's own documentation recommends starting with Fabric Data Warehouse, so evaluate Fabric (and other platforms) before building anything new on Synapse.
Overview
Azure Synapse Analytics brings several engines together in a Synapse workspace managed through Synapse Studio. Dedicated SQL pools (formerly SQL DW) are provisioned massively parallel T-SQL warehouses sized in Data Warehouse Units, with data spread across 60 distributions by hash, round-robin or replication. Serverless SQL pools run pay-per-query T-SQL over Parquet, CSV, TSV and JSON files in the data lake. Apache Spark pools handle data engineering and ML, and pipelines use the same integration engine as Azure Data Factory.
The service follows Microsoft's Modern Lifecycle Policy and is listed as In Support. Microsoft's current direction, however, is Microsoft Fabric: it publishes migration guides and a Fabric Migration Assistant for dedicated SQL pools, Spark and pipelines. Moving is a migration rather than an in-place upgrade, with documented data type and T-SQL differences.
Key features
- Dedicated SQL pools from DW100c to DW30000c, with pause, resume and scaling
- Serverless SQL pools billed per TB processed over files in the data lake
- Apache Spark pools with Delta Lake support in the same workspace
- Pipelines and data flows built on the Azure Data Factory integration engine
- Workload management with workload groups for concurrency and resource isolation
- Synapse Studio with Git-based CI/CD and database projects for dedicated pool objects
Pros and cons
Pros
- T-SQL throughout, familiar to SQL Server teams
- Serverless SQL pools for pay-per-query exploration of lake files
- Dedicated pools can be paused to stop compute charges
- Documented migration path and tooling to Microsoft Fabric
Cons
- No longer Microsoft's recommended starting point for new data warehouses
- Synapse Data Explorer is retired and Synapse Link for Cosmos DB is closed to new projects
- Dedicated pools bill per hour while running, and concurrency is capped by service level
- Distribution, indexing and workload management need manual design and tuning
Pricing
Each engine is billed separately. Dedicated SQL pools are billed per hour at the chosen service level (DW100c to DW30000c) while running, pay-as-you-go or with one- or three-year reserved capacity; pausing stops compute charges. Storage is billed per TB per month, plus geo-redundant recovery storage. Serverless SQL pools are billed per TB of data processed, with a 10 MB minimum per query. Apache Spark pools are billed per vCore-hour, and pipelines by activity runs and integration runtime hours. The Azure pricing page shows figures only after choosing a region; Microsoft's Azure Retail Prices API lists, for East US in November 2026, USD 1.51 per hour for DW100c (USD 15.10 for DW1000c), USD 5.00 per TB processed for serverless SQL pools and USD 23.00 per TB per month for dedicated SQL pool storage. Prices vary by region.
Pricing checked on the vendor's official page on 7 November 2026. Prices change; confirm before buying.
Who it is best for
Existing Azure customers with dedicated SQL pools that meet their needs and who want to plan a move to Fabric on their own schedule.
Who should look elsewhere
- New warehousing projects on Azure, where Microsoft recommends Fabric Data Warehouse: see Databricks vs Microsoft Fabric.
- Teams that need a warehouse on more than one cloud: see Snowflake vs Azure Synapse.
Verdict
Azure Synapse Analytics is supported and has no announced retirement date, so existing deployments do not need to move in a hurry. For new work, though, Microsoft itself points to Fabric. Inventory your T-SQL, data types and Spark jobs against Fabric's documented differences, cost Fabric and other platforms on your workload, and migrate when the business case is clear.
Frequently asked questions
Is Azure Synapse Analytics being retired?
Not as a whole. The Microsoft Lifecycle site lists Azure Synapse Analytics as In Support under the Modern Lifecycle Policy, with no retirement date. Synapse Data Explorer (preview) was retired on 7 October 2025, and Synapse Link for Azure Cosmos DB is no longer supported for new projects.
Can I still create new dedicated SQL pools?
Yes. The Synapse FAQ states that new dedicated SQL pools are deployed into a Synapse workspace by default, and a standalone dedicated SQL pool (formerly SQL DW) can still be created. Microsoft recommends Fabric Data Warehouse for anyone new to data warehousing.
How much does a Synapse dedicated SQL pool cost?
It is billed per hour at its service level while running. Microsoft's Azure Retail Prices API listed DW100c at USD 1.51 per hour in East US in November 2026, with storage at USD 23.00 per TB per month. Prices vary by region, and reserved capacity lowers the rate.
Sources
- Microsoft Lifecycle: Azure Synapse Analytics
- Microsoft Learn: What is dedicated SQL pool (formerly SQL DW)?
- Microsoft Learn: What is Azure Synapse Analytics?
- Microsoft Learn: Azure Synapse Analytics FAQ
- Microsoft Learn: Migrate from Azure Synapse Data Explorer to Fabric
- Azure Synapse Analytics pricing
Checked November 2026.
How we research reviews and what star ratings mean: our editorial method.