Fabric Data Warehouse

by
Andrew Snodgrass

by
Andrew Snodgrass

Andrew analyzes and writes about Microsoft's data management, business intelligence, and machine learning solutions, as well as aspects of licensing... more

Fabric Data Warehouse is the relational data warehousing component in Microsoft Fabric. Its tables are stored in OneLake as Delta Parquet files, and a distributed, serverless T-SQL engine is layered over them to give them structure. It also allows developers to load, update, and query data using views, stored procedures, functions, and multitable transactions. The same engine powers the read-only SQL analytics endpoint that provides access to OneLake lakehouses. Compute comes from the workspace’s Fabric capacity, with custom SQL pools to isolate workloads. The warehouse succeeds Azure Synapse Analytics dedicated SQL pools, but its T-SQL is only partly compatible with SQL Server and migrating from Synapse or SQL Server is a major effort.

Product: Fabric Data Warehouse

Licensing: Fabric

Near-Term Updates and Roadmap (2026–2027)

Microsoft’s 2026 work on Data Warehouse aimed at the obstacles that stall migrations from Synapse Analytics and SQL Server, including workload isolation, T-SQL coverage, and migration tooling. However, compatibility with SQL Server remains the product’s defining weakness, and the cost model changed underneath customers in 2026, so metering now favors sustained workloads over short, scattered queries.