Power BI real-time streaming semantic models retire

What should you do? decision tree

1. you have no push or streaming semantic models and no plans for any
   Do nothing now; note the date and avoid starting new real-time work on this feature.
2. you only run existing streaming models and do not need new ones
   Keep them as they are, since Microsoft says they are unaffected, and document each one with its owner and data source.
3. you plan new real-time dashboards before October 31, 2027
   Create them before the cutoff only if short-lived, otherwise start on Real-Time Intelligence in Fabric.
4. you will need new real-time models after the cutoff and have no Fabric capacity
   Scope a Fabric trial or capacity, and compare it against a separate real-time BI tool.
5. none of the above fits
   Inventory every streaming and push dataset and tile now, then decide per use case whether to keep, rebuild or retire it.

Options

• WaitKeep existing streaming and push models as they are; Microsoft says they are unaffected. Just stop planning new ones past the cutoff date.
• OptimizeInventory all push, streaming and PubNub datasets and streaming tiles, and flag which ones will need rebuilding or cloning later.
• Re-architectBuild new real-time use cases on Real-Time Intelligence in Fabric, as Microsoft recommends, and size the capacity it needs.
• MigrateFor heavy real-time needs, move those dashboards to a BI tool with native live or streaming query support.

Alternatives editorial picks, not sponsored