Microsoft brings row-level security to Fabric's high-speed Direct Lake models
Microsoft has added row-level security to Direct Lake semantic models in Fabric. This lets data analysts use DAX expressions in Power BI Desktop to define roles that hide rows from specific viewers, while administrators see everything. The filters stay in place even if the model falls back to DirectQuery mode for unsupported features. Direct Lake handles large datasets without the delays of data imports, a common pain in Power BI. Previously, enabling security meant slower performance or complex workarounds. Admins can now monitor impacts through Fabric's capacity metrics app.
Before, Direct Lake offered blazing-fast dashboard loads on huge datasets, but row-level security was missing. Analysts had to import data or switch to DirectQuery, risking hours-long refreshes on messy sources like shared Excel files. Beginners stuck copying data into Excel to avoid these headaches. Now, DAX roles secure views without sacrificing Direct Lake speed in most cases. Fallback to DirectQuery only hits on edge features, preserving the 'seconds to load' edge for boss demos. This closes the gap between Excel hacks and pro Power BI sharing in mid-size teams.
Analysis
This hands you the keys to build secure dashboards your boss will actually use, without the import nightmares or data leaks that kill credibility. But fallback mode turns 'instant' into 'Excel-slow' if your DAX roles trigger it, so test ruthlessly. Right now, take one report from your shared drive, switch it to Direct Lake in Power BI Desktop, add a simple department-filter RLS role, publish to Fabric, and time the visuals before/after sharing.
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)".