Databricks Kills Decades-Old Data Pipeline Problem

Databricks announced two products at its Data + AI Summit designed to eliminate the longstanding divide between operational and analytical databases. Lakehouse//RT delivers millisecond query latency on Delta and Iceberg tables, removing the need for separate real-time serving infrastructure. LTAP stores Postgres-native transactional data directly in Delta and Iceberg format from the point of write, cutting out ETL pipelines entirely. The moves address a structural problem AI agents expose: systems that reason and act on live data cannot tolerate delays between themselves and the information they need.