Amazon Acquires DuckLabs to Power Aurora PostgreSQL

Amazon has acquired DuckLabs, the Amsterdam-based startup behind the open-source analytical database DuckDB. Under the acquisition, co-founders Hannes Mühleisen and Mark Raasveldt will join Amazon Web Services to guide the technical roadmap while continuing to lead the open-source engine. Amazon confirmed that DuckDB and its related "Duck Stack" utilities, such as DuckLake and Quack, will remain free and open-source under the MIT license under the ongoing governance of the non-profit DuckDB Foundation.

The deal brings DuckDB's in-process analytical capabilities directly into Amazon Aurora PostgreSQL as its first user-facing cloud feature. DuckDB was engineered to deliver fast analytical execution for the majority of queries that operate on datasets of 1 terabyte or less. AWS executives highlighted that these optimizations align closely with agentic AI workloads, where autonomous agents frequently execute small exploratory queries across datasets before taking action. The merger expands on approximately two years of prior collaboration between AWS and DuckLabs, which previously focused on S3 Tables, AWS Lambda integrations, and DuckDB's ATTACH and CONNECT commands.

For database administrators and developers, the integration simplifies complex hybrid architectures by eliminating the need for custom reverse ETL pipelines. Previously, combining transactional records in Aurora with historical archives required manual data synchronization and extra infrastructure. Aurora PostgreSQL users can now write a single query that joins live operational tables directly with Apache Iceberg datasets managed in the AWS Glue Data Catalog, alongside Parquet and Iceberg files stored in Amazon S3 and S3 Tables.

By processing queries natively inside Amazon Aurora without extra network hops, the feature reduces infrastructure complexity and operational overhead. The functionality is available across all commercial AWS regions and AWS GovCloud (US) at no additional license fee, with customers charged only for incremental Aurora compute resources and standard S3 storage requests.