Automatic change data feed reached general availability as teams refined external recipient governance.

Automatic change data feed reached general availability in September 2026, giving data providers a standardized method to stream incremental table changes to external consumers [5]. Alongside this milestone, developers spent significant effort standardizing OpenSharing workflows. Community members published patterns for securing cross-cloud shares through additional authentication layers [6] and tackled practical questions regarding how OpenSharing interfaces directly with Databricks environments [7]. Furthermore, technical teams evaluated using OpenSharing protocols to transport machine learning model artifacts across regions to support distributed model serving [2].

Operational management for external recipients remained another major theme. Practitioners explored architectural solutions for automating credential rotation and recipient token lifecycles when serving organizations that do not run Databricks or Unity Catalog [3]. Cost attribution also drew attention, specifically regarding which party absorbs egress network bills when transferring shared data across cloud boundaries, such as between Google Cloud Platform and Microsoft Azure [8].

Application architecture and end-user tooling rounded out the month. Independent software vendors investigated Delta Sharing as a mechanism to store entitlement records securely outside a customer's primary metastore [4]. In addition, community guides emerged demonstrating how to integrate newly shared public datasets into conversational AI workflows using Databricks Genie [1].