Snowflake Adds Postgres Support to Unify Data for AI
Snowflake has launched Snowflake Postgres and new open data interoperability capabilities, aiming to make enterprise data "AI-ready." The update allows users to build applications on Postgres and run them on the Snowflake platform, reinforcing its strategy to be a central lakehouse for diverse enterprise data. As an example of adoption, Visionaize is partnering with Snowflake to accelerate AI-driven operations.
- The introduction of fully managed Snowflake Postgres is built on technology from Snowflake's acquisition of Crunchy Data, a provider of enterprise Postgres services. This move allows applications that use standard Postgres to be moved to Snowflake without any code changes. - This new offering is distinct from Snowflake's existing Unistore workload, which was introduced in 2022 to unify transactional and analytical data using proprietary Hybrid Tables. The addition of a PostgreSQL-compatible service was driven by customer demand for a familiar, open-standard relational database for their applications. - Snowflake Postgres instances run in their own isolated private networks with attached disks, a different architecture from the standard Snowflake model of separated compute and storage, to provide better performance for transactional workloads. - The service integrates with open data formats by using pg_lake, a set of open-source PostgreSQL extensions that enable reading and writing directly to Apache Iceberg tables, reducing the need to move data between transactional and analytical systems. - This launch is part of a larger trend of data platform convergence, with competitors like Databricks also expanding from their data lake and AI/ML roots into data warehousing and transactional capabilities to capture a wider range of enterprise workloads. - The move reinforces Snowflake's strategy to be the central platform for AI-driven applications, allowing developers to build and run the transactional components of their apps directly on Snowflake, where the data for AI models and analytics already resides. - Alongside the Postgres support, Snowflake also recently introduced Cortex Code, an AI-powered coding agent designed to understand an enterprise's specific data and governance context to accelerate development.