Databricks announced two new products at its Data + AI Summit on Tuesday, aiming to solve a decades-old data pipeline problem that has hindered the performance of AI agents. The offerings are designed to collapse the infrastructure separating operational and analytical databases, a structural issue that becomes critical when systems must reason continuously on live data.
Lakehouse//RT delivers millisecond query latency directly on governed Delta and Iceberg tables, removing the need for a dedicated real-time serving tier alongside the lakehouse. LTAP, or Lake Transactional/Analytical Processing, stores Postgres-native transactional data in Delta and Iceberg format from the point of write, eliminating the ETL pipelines that have connected operational and analytical systems for decades.
Reynold Xin, co-founder of Databricks, described a simpler data stack as "the holy grail for agents" in a briefing with VentureBeat. He argued that as users "vibe code" more applications, agents reasoning analytically atop those apps require the underlying data infrastructure to keep pace.
The announcements address a longstanding pain point for data professionals, who have struggled to manage both operational and analytical databases without introducing latency and performance degradation. By eliminating the pipeline between data and the systems that act on it, Databricks is betting that its platform becomes the default for enterprise AI workloads.
Some experts may question whether the new products truly eliminate all pipeline complexity or simply shift it, and the company has yet to disclose detailed performance benchmarks or pricing for the offerings.