Snowflake has released new open source interoperability features, including the pg_lake extension and expanded Apache Iceberg and dbt support, positioning the company more directly within the open data lakehouse ecosystem.
Snowflake has announced the general availability of Snowflake Intelligence, a platform that enables employees across skill levels to query enterprise data using natural language. The system automatically interprets business semantics, generates SQL queries and provides verified outputs marked with a “green shield” to indicate certified data sources. Over 1,000 customers have already used the platform during testing, deploying more than 15,000 AI agents.
Snowflake Intelligence integrates models from Anthropic and OpenAI and supports multimodal data synthesis to help organisations understand not only what is happening, but why. As Jeff Hollan, Head of Cortex AI Agents, said: “It is not just to get basic information but to understand why a trend is happening.”
The platform’s agentic capabilities allow automated actions, while enterprise governance rules are enforced throughout. “We focus every day on making sure Snowflake is trusted and that our customers can entrust us with their data,” said Christian Kleinerman, Executive Vice President of Product.
To support developers, Snowflake has made Cortex Agents generally available and introduced secure integration through a Model Context Protocol Server.
Workspaces now support Git and Visual Studio Code, while dbt Projects on Snowflake are generally available for end-to-end workflow management.
In a significant open-source step, Snowflake has introduced a fully managed PostgreSQL engine and released the pg_lake extension to open source. The extension enables PostgreSQL to read and write directly to Apache Iceberg tables without extract-transform-load steps, creating a direct interface to open lakehouses.
As Kleinerman noted: “The pg_lake extension is for anyone who wants to turn Postgres into an interface to manage an open lakehouse.”
