Aug 13, 2026: Snowflake Data Clean Rooms updates

Clean Rooms API Version: 17.5

The following updates are now available in Snowflake Data Clean Rooms:

  • Improved error handling and validation for RUN: Error handling now covers a broader range of template run failure cases, surfacing clear, actionable error messages, along with faster pre-flight validation. These improvements apply to collaborations created after the June 18, 2026 release.
  • Faster API performance: Improvements have been made to reduce latency for GET_STATUS, VIEW_REGISTERED_TEMPLATES, VIEW_REGISTRIES, and GET_CONFIGURATION.
  • Fixed freeform SQL view resolution for data providers: Incorrectly formatted freeform SQL view names in the response of VIEW_DATA_OFFERINGS for the data provider have been corrected.
  • ML Jobs content hash validation: ML Jobs code specs now automatically compute and verify a SHA-256 content hash for every file in stage_code_dir at registration. Code specs registered before the June 18, 2026 release don’t have content hashes and can’t be added to a collaboration. Re-register the spec as a new version to enable hash validation. For details, see ML Jobs in Data Clean Rooms and the ML Jobs troubleshooting guide.
  • Default ML Jobs container runtime image upgraded to 2.8.1: The default ML Jobs container runtime image is now 2.8.1. Collaborations that don’t pin image_tag use the new default on the next patch. To trigger a patch, add a template with the code spec to the collaboration. All pinned image_tag values are preserved unchanged. To avoid being affected by future default image changes, pin image_tag in your ML Jobs code spec. For details, see the ML Jobs code spec field reference.
  • General performance improvements and bug fixes.
  • Updates to private preview features.