Aug 14 — Aug 21, 2026
This week's updates centered on operationalizing agentic AI within governed lakehouse environments and strengthening core developer and governance primitives. Databricks doubled down on Genie and Unity Catalog-governed agents, AI Gateway budget and privacy controls, and Spark 4.2 alignment across Delta Lake and Unity Catalog.
1.Databricks Genie and Unity Catalog Expand Governance to AI Agents and MCPs
Databricks is unifying agent development with Unity Catalog, bringing access controls, discovery, and audit logging to Model Context Protocol (MCP) servers, tools, and domain-specific Genie Agents. Blog posts, architecture videos, and Terraform provider updates emphasize moving beyond general coding agents to context-grounded workflows that directly interrogate governed structured and unstructured lakehouse data.
Sources
2.AI Gateway Adds Fine-Grained Spending Limits, Tracing, and PII Guardrails
A wave of hands-on tutorials detailed production safety features inside the Databricks AI Gateway. Platform administrators can now enforce tiered budget thresholds with automated blocking at workspace or per-user levels, inspect LLM telemetry via OpenTelemetry-compatible traces, and deploy account-level PII input guardrails before prompts reach serving endpoints.
Sources
- Configure Budget Thresholds for AI ModelsVideo · Databricks Skill Builder · Aug 21
- Monitor LLM Usage & Prevent Data Leaks with TracesVideo · Databricks Skill Builder · Aug 19
- Set AI Budgets & Control Spending in Databricks!Video · Databricks Skill Builder · Aug 17
- Databricks AI: PII Blocking for Data Privacy!Video · Databricks Skill Builder · Aug 15
3.Delta Lake 4.4 and Unity Catalog 0.6 Add Apache Spark 4.2 Support and Catalog Views
Open-source storage and governance layers released major milestone versions aligned with Apache Spark 4.2. Unity Catalog 0.6 introduces governed metric and SQL views alongside shorter default token lifespans, while Delta Lake 4.4 adds identity columns, generated DDL columns, and broader UC Delta API integration across Delta Kernel and Flink.
Sources
- Busting SQL Migration Myths: How New SQL Features Make Lift-and-Shift to Lakehouse EasierNews · databricks-blog · Aug 20
- UnityCatalog 0.6.0Release · unitycatalog/unitycatalog · Aug 20
- Delta Lake 4.4.0Release · delta-io/delta · Aug 20
- How Databricks Feature Store serves features with sub-second freshnessNews · databricks-blog · Aug 17
4.Developer Tooling Gains Git-Driven App Deploys, Environment Sync, and DAB Controls
The developer ecosystem received major workflow refinements across the CLI, IDE extensions, and SDKs. Databricks VS Code added automated local virtual environment creation to match target cluster runtimes, the CLI enhanced Asset Bundle lifecycle summaries, and SDK releases delivered native Git auto-deploy integration for Databricks Apps.