Google BigQuery Now Supports Cross-Engine Apache Iceberg Tables
Google Cloud has made a significant move in the data lakehouse space. At the Apache Iceberg Summit last month, Google announced new interoperability features for Apache Iceberg in BigQuery. Now, teams can create, update, and query the same Apache Iceberg BigQuery tables across engines like Spark, Flink, and Trino, without duplicating a single byte of data.
The centerpiece of this update is a serverless Iceberg REST catalog, currently in preview. Previously, teams had to choose between Iceberg tables managed by the Google-managed REST catalog or those managed directly by BigQuery. That meant teams relying on Apache Spark for ETL couldn’t write through BigQuery or use its storage management features. This update removes that forced choice.
Beyond catalog access, Google is also extending BigQuery infrastructure to handle managed metadata, automatic table maintenance, transactions, and change data replication. These are tasks that teams typically handle manually in Iceberg deployments, and they add up to real operational overhead.
Then at Google Cloud Next ’26, Google went further. The company expanded Apache Iceberg BigQuery interoperability to a cross-cloud lakehouse, adding support for querying Iceberg catalogs across AWS, Azure, Databricks, and Snowflake. AI workflows are included too. The goal, according to Google, is to let organizations keep data in open formats while using any combination of processing and analytics tools.
Centralized access controls are also part of the package. Permissions can now be managed consistently across query engines, a meaningful step for governance teams managing multi-engine environments. Additionally, BigQuery ObjectRefs are now generally available, enabling teams to combine structured Iceberg data with unstructured files in Cloud Storage for multimodal and AI workloads. The Knowledge Catalog (formerly Dataplex), still in preview, handles metadata, lineage, and access control across systems.
Practitioners are paying close attention. David Colbert, commenting on a LinkedIn discussion about the announcements, points out that open formats solve storage portability, but control plane choices determine long-term optionality. Teams often hit friction fast on compaction, metadata management, and orchestration. Google’s managed approach directly targets that friction.
Google is not alone in this space. AWS analytics services including EMR, Glue, Athena, and Redshift all provide native Apache Iceberg support. However, Shashank Muthuraj, cloud engineer at Red Oak Strategic, notes that Apache Iceberg has grown from a Netflix engineering project into the undisputed standard for open data lakehouse architecture in under seven years. The industry alignment behind it is unprecedented.
Some observers frame Google’s strategy more boldly. Precious Pendo, writing about the Next ’26 announcements, argues that Google is betting enterprise AI value will go to whoever owns the reasoning layer over data, not just the storage layer.
For now, the core managed Iceberg table support in BigQuery is generally available. The broader open interoperability features and the REST catalog announced at Iceberg Summit 2026 remain in preview.
