Google Cloud’s BigQuery supports querying data across Databricks, Snowflake, and AWS Glue catalogs in place, without migrating underlying datasets.

Google Cloud describes a cross-cloud architecture using BigQuery and its borderless Lakehouse, based on the open Apache Iceberg REST catalog specification. Supported external catalogs include Databricks Unity Catalog, Snowflake Horizon Catalog, and AWS Glue Data Catalog.

The architecture separates metadata synchronization from data access. Lakehouse runtime catalog periodically synchronizes schemas, snapshot pointers, and table definitions according to a user-defined refresh schedule, while compatible query engines read Parquet files directly from remote object storage, including Amazon S3.

Storage access uses temporary credentials issued by the remote platform. Query processing runs on Google Cloud compute, such as BigQuery or Managed Service for Apache Spark; remote Databricks clusters, Snowflake virtual warehouses, and AWS compute instances are not invoked during execution.

The article demonstrates a single query joining data across all three catalogs and natural-language exploration through BigQuery conversational analytics. Google contrasts this with copying datasets into a central warehouse, citing synchronization lag, recurring storage costs, variable network egress fees, and pipeline breakage as drawbacks of the traditional approach; it does not claim that querying in place eliminates every cost.