Sep 29 — Oct 6, 2026
This week centered on operationalizing enterprise AI, led by the general availability of Unity AI Gateway and the introduction of the low-latency ai_decide() SQL function. Supporting updates placed heavy emphasis on transactional database branching in Lakebase and open table catalog portability.
1.Native SQL Classification Debuts with the ai_decide() Function
Databricks introduced ai_decide(), a specialized SQL function tailored for deterministic decision-making and categorization rather than token-heavy generative responses. SDK releases for Python and Java immediately integrated client-level support, while community videos demonstrated utilizing these System 1 style models to generate calibrated probabilities and classifications directly within batch transformations.
Sources
- Introducing ai_decide: make fast decisions on your governed dataNews · databricks-blog · Sep 30
- Let AI Decide: SQL Decision Making with Databricks ai_decide()Video · databricks MVP Hubert Dudek · Oct 3
- The Shift to System One | System One Models - Part OneVideo · Advancing Analytics · Oct 2
- v0.144.0Release · databricks/databricks-sdk-py · Sep 30
2.Unity AI Gateway Hits GA as Enterprise Agent Governance Deepens
Databricks officially rolled out the general availability of Unity AI Gateway, offering unified observability, guardrails, and cost limits across multi-model agent systems. Concurrently, Model Context Protocol (MCP) support expanded with Meta's advertising server arriving in the Marketplace, complemented by new SDK securables and APIs built to manage agent tools and skills.
Sources
- How Enterprises Govern AI Agents Across Multiple ModelsVideo · Databricks · Oct 1
- Meta’s ads MCP server comes to Databricks: Put your customer intelligence to work in advertising campaignsNews · databricks-blog · Oct 6
- Demo: Building a Governed AI Agent with Unity AI GatewayVideo · Databricks · Oct 1
- v0.157.0Release · databricks/databricks-sdk-java · Oct 1
3.Lakebase Postgres Operationalizes Database Branching and Instant Recovery
Databricks highlighted major performance and operational patterns for Lakebase Postgres, introducing zero-copy branch-based restores capable of recovering large-scale databases in seconds. Ecosystem tooling aligned with this model as the Databricks CLI added lifecycle flags for branches and engineering demos detailed parallel coding agent workflows driven by ephemeral database clones.
Sources
- Lakebase Postgres branch-based restores for fast recovery at scaleNews · databricks-blog · Oct 1
- Parallel Coding Agents with Lakebase | Claude Code + GitHub ActionsVideo · Databricks · Oct 1
- A practical guide to cost optimization with Lakebase PostgresNews · databricks-blog · Sep 30
- v1.19.0Release · databricks/cli · Sep 30
4.Open Catalog Interoperability Expands via REGISTER APIs and Iceberg Standards
Efforts to eliminate storage lock-in advanced with new REGISTER and UNREGISTER APIs designed to transfer table ownership across catalogs without moving underlying object data. In tandem, Delta Lake 4.4.1 introduced stricter safeguards against unmanaged table cleanups, while broader catalog-level governance standards moved toward unified cross-engine read restrictions.
Sources
- Unlocking Data Portability: Preventing Catalog Lock-in with REGISTER and UNREGISTER APIsNews · databricks-blog · Oct 5
- Read Restrictions and Catalog Labels: Unifying governance across engines and catalogsNews · databricks-blog · Oct 1
- Delta Lake 4.4.1Release · delta-io/delta · Oct 5
- Your data, your storage, your rules: a 2026 guide to storing Unity Catalog managed tablesNews · databricks-blog · Sep 29