Databricks Kills Data Pipelines With Unified Architecture

Databricks unveiled a new data architecture at its Data + AI Summit in San Francisco aimed at eliminating the long-standing separation between transactional and analytical systems. Called Lake Transactional/Analytical Processing, the platform stores operational and analytical data in a single copy on a data lake, removing the need for ETL pipelines and change data capture processes. The company also introduced Lakehouse//RT, a real-time analytics engine powered by a new execution engine called Reyden, promising response times as low as 10 milliseconds. Databricks framed both announcements as essential infrastructure for AI agents, which require fast, accurate data access to function effectively.