This repository organizes and summarizes all major product announcements from the Databricks Data + AI Summit 2025, grouped by key platform categories.

• Data + AI Summit Homepage: Watch sessions on-demand
• Official Announcements Blog: All platform announcements
• Databricks Documentation: Main documentation portal

	Data + AI Summit 2025 Main Keynote Covers the Data Intelligence Platform vision and major announcements around Mosaic AI, Lakebase & Databricks Apps.
	Data + AI Summit 2025 Day 2 Keynote Dives deeper into product announcements like Lakeflow, Unity Catalog, Databricks SQL & AI/BI.

• Agent Bricks: Auto-optimized domain-specific agents with evaluation and tuning built-in.
• MLflow 3.0: Unified experimentation, observability, and governance for GenAI and ML.
• AI Functions in SQL: Run GenAI functions (e.g. ai_parse_document) directly in SQL.
• Vector Search: Storage-optimized vector store with compute/storage separation.
• Serverless GPUs: Instant A10G GPU access for fine-tuning and inference.
• Model Serving Enhancements: High-QPS, lower-latency serving with in-house inference engine.
• AI Gateway: Central control for routing, safety, and usage across LLM endpoints.

📁 mosaic-ai/

• Lakeflow: Unify pipelines for ETL and orchestration.
• Lakeflow Designer: No-code interface for building ETL pipelines.
• Declarative Pipelines: Open-sourced DLT framework powered by Spark 4.0.
• Managed Connectors: New support for SQL Server, GA360, ServiceNow, and row-level ingestion API.

📁 lakeflow/

• Databricks Apps: Secure, governed data apps using Python/JS frameworks.
• Genie: Conversational analytics with AI/BI Dash, powered by UC metrics and LLMs.
• AI/BI Dashboards: LLM-infused dashboards with UC-backed metrics.
• Databricks One: Secure, streamlined access for data consumers.

📁 ai-bi/

• UC Metrics: Semantic layer to define consistent metrics across dashboards & models.
• Attribute-Based Access Control (ABAC): Fine-grained, dynamic policy enforcement.
• Tag Policies: Governance via tag-based rules.
• Data Classification: Auto-detect and classify sensitive data.
• Anomaly Detection: Identify abnormal usage patterns in governed datasets.
• Iceberg Support: Enhanced governance and compatibility with Apache Iceberg tables.

📁 unity-catalog/

• Lakebase: Fully managed Postgres integrated into the Lakehouse.
• Lakebridge: Migration toolkit for 25+ legacy data warehouses.
• Power Platform Connector: Real-time access to Databricks from Power Apps, Copilot, etc.

📁 databricks-sql/

• Community Edition: Free, hands-on Databricks learning workspace.

📁 free-edition/