Databricks says it solved the decades-old data pipeline problem that's been slowing AI agents

Databricks announced Lakehouse//RT and LTAP, two products designed to reduce the need for separate real-time serving, transactional, and analytical data systems. Lakehouse//RT brings millisecond query latency to governed Delta and Iceberg tables, while LTAP stores Postgres-native transactional data directly in those formats from the point of write.

The launch reflects a broader push to simplify enterprise data stacks as AI agents require faster access to fresh, governed data. By shifting unification to the storage layer rather than relying on separate engines and ETL pipelines, Databricks is positioning its lakehouse architecture as a foundation for agent-driven applications.