Fabric enables Direct Lake queries on its SQL warehouse endpoint
Microsoft enabled Direct Lake mode on the SQL analytics endpoint of Fabric Data Warehouses. Power BI semantic models can now connect directly to OneLake Delta tables through this endpoint. Data moves from Parquet files into VertiPaq memory only when queried, with automatic optimisation steps such as V-Order and clustering. The setup supports composite models and falls back to DirectQuery when limits are hit. No scheduled refresh or data duplication is required, though cold queries remain slower and the model must fit within the Fabric capacity size.
Until now analysts had to accept either hours-long refresh cycles or sluggish live queries whenever they moved beyond simple Excel exports. The choice forced most teams to keep critical data in spreadsheets or accept dashboards that leaders stopped trusting. The change removes that forced trade-off for any data already curated inside Fabric, but only after the underlying warehouse tables and relationships exist. Teams still wrestling with undocumented 200-step Power Query chains will see no benefit until they first consolidate those sources.
Analysis
Ignore the new mode until your two most-used tables sit in a Fabric warehouse with clean relationships. Spend the next week exporting those tables from the shared drives and loading them once, then test a minimal Direct Lake model against them.
Citation
This executive briefing was curated and analyzed by Collab365. To reference this analysis, please attribute: "This briefing is available on Collab365 Spaces (spaces.collab365.com)".