Data + AI Summit Keynote 2026 | Day 1

Summary

Key Points

• —Databricks' Data and AI Summit, the largest ever, emphasizes its open-source roots with projects like Spark, Delta Lake, MLflow, Unity Catalog, Postgres, and the newly open-sourced Omnigent. 3:12
• —The presentation showcases impactful AI applications from customers such as Insulet for diabetes management, Tonal for personalized workouts, and Merck for transformer-enabled drug discovery. 7:05
• —Databricks asserts that Artificial General Intelligence (AGI) is already present, but the primary challenge for enterprises is integrating AI effectively by providing it with comprehensive context from their diverse and often siloed data. 9:34
• —LakeFlow unifies data ingestion and transformation into an open lakehouse format (supporting both Delta and Iceberg), featuring real-time capabilities with Zero Bus and Spark Real-Time Mode, alongside visual data preparation via LakeFlow Designer. 18:32
• —Genie Ontology automatically constructs a comprehensive knowledge graph from all organizational data, including external sources, to provide AI agents with critical context for faster, more accurate, and cost-efficient responses. 39:14
• —Genie 1, Genie Agents, and Genie Code are presented as AI co-workers that leverage Genie Ontology to perform tasks like document generation, autonomous workflows, data engineering, and machine learning, with Zero Ops automating data pipeline maintenance. 46:43
• —The platform addresses four critical challenges for enterprise AI: delivering relevant context, ensuring robust security and control, optimizing costs, and offering choice to avoid vendor lock-in across data, models, and cloud environments. 59:12
• —Databricks introduces Lakehouse RT, a new SQL warehouse powered by the Raiden engine, which delivers sub-second latency and massive concurrency for real-time analytics directly on lakehouse data. 126:28
• —LakeBase offers a fully managed, serverless, and branchable Postgres database that operates on the lake, enabling cost-effective operational workloads with built-in cross-cloud disaster recovery. 137:49
• —The ultimate vision is an “agentic system of record” achieved through Altap (Lake Transactional Analytical Processing), which unifies OLTP and OLAP by allowing a single copy of data to serve both operational and analytical needs without performance compromise or complex CDC pipelines. 168:36