Azure Databricks connector enters public preview for Power Platform

Microsoft and Databricks released a connector in public preview on 11 June. It gives Power Apps, Power Automate, and Copilot Studio direct read and write access to Databricks tables. The connector keeps Databricks governance and access controls intact. No custom API work is required. Network setup such as VNet peering may still be needed for private access, and write-back remains limited in the current preview.
Until now the only realistic data store for these builders was SharePoint lists. That choice created hard ceilings on row counts, delegation warnings, and slow galleries once real usage began. The new connector offers a governed alternative, but it also introduces new questions about cost, network rules, and when the extra capability is actually required versus when it simply adds another layer the builder cannot maintain.
Analysis
Ignore any internal pressure to 'modernise' your current SharePoint app by swapping in Databricks until you can name the exact limit you are hitting and the licensing delta it creates. Run one controlled test: connect a small canvas app to a single Databricks table, measure load time and delegation behaviour, then compare it against the same query on your existing list.
Pulse published by Collab365 Spaces, reviewed by Helen Jones on . Cite as "Azure Databricks connector enters public preview for Power Platform", Collab365 Spaces. 1 source referenced.