EDB unveils Agentic Lakehouse vision at Seoul AI summit

At the EDB Postgres AI Summit Seoul 2026, data company EDB unveiled its ‘Agentic Lakehouse’ vision, a unified data platform designed to integrate transaction processing, real-time analysis, and AI capabilities. The architecture aims to reduce reliance on proprietary commercial databases and specific cloud providers by connecting operational databases (OLTP), real-time analysis (ClickHouse), and petabyte-scale data warehousing (WarehousePG) under a single SQL interface and security model.

EDB also previewed ‘WarehousePG 19’, a next-generation data warehouse engine scheduled for release in the first half of 2027. Key features include property graphs for fraud detection and network analysis using standard SQL, large-scale streaming data storage, and ‘Persistent Agent Memory’ technology to manage relational, vector, and graph memory within a single ACID database to reduce generative AI token costs.

Several South Korean companies shared successful migration and integration cases. IBK Industrial Bank of Korea demonstrated a parallel transition of 15 systems to open source, while Kyobo Book Centre integrated operational, analytical, and AI-driven book recommendation services on a single Postgres platform. Additionally, Shopcast reported reducing quarterly settlement batch processing time from 40 minutes to 7 minutes after transitioning to the EDB Postgres Lakehouse.