Cloud data platform Snowflake on April 8 announced a "data autonomy" strategy to strengthen interoperability aimed at minimising data movement, simplifying governance and improving enterprise AI systems' access to data.

The data autonomy strategy focuses on helping organisations access, manage and analyse data anywhere without moving it between platforms.

The company said existing architectures force data movement, driving operational complexity, security risks and higher costs, and reducing the efficiency of AI workloads.

The core is expanded support for Apache Iceberg version 3. Iceberg V3 supports the Variant data type for semi-structured data without a fixed structure such as JSON and XML, geospatial data types, row-level lineage tracking, improved delete operations through deletion vectors, and nanosecond timestamps. It works across both Snowflake managed tables and external Iceberg catalogs, providing a portable data experience regardless of environment.

Snowflake is also strengthening governance portability. Using Apache Polaris, which it released as open source 2 years ago, Snowflake supports governance policies such as access controls and semantic context moving with data rather than being tied to a specific platform.

James Rowland-Jones (제임스 롤랜드-존스), a director of product management at Snowflake, said: "Today, the only way to safely share data with fine-grained access controls with external engines is to materialize intermediate results through APIs. It is inefficient and expensive. We are changing this structure with Polaris."

The announcement also includes pg_lake, an open-source PostgreSQL extension that Snowflake unveiled in November last year. It allows PostgreSQL databases to directly query data lake formats such as Parquet and CSV and write data to Iceberg tables without ETL pipelines.

Snowflake also supports new standards such as OpenLineage, which tracks data movement, and Open Semantic Interchange, which standardises business definitions such as metrics and dimensions. It is part of efforts to help AI interpret data better.

Snowflake has made more than 9,000 contributions to open-source projects over the past 2 years and is also participating in the development of Iceberg version 4.