Amazon has acquired DuckLabs, the Amsterdam-based firm behind the analytical database DuckDB, and launched the first DuckDB-powered capability in Aurora PostgreSQL on October 17. The acquisition allows Amazon to embed DuckDB's analytical processing directly into its cloud database service, eliminating the need for separate data pipelines when combining real-time operational information with archived records stored in S3.
The integration enables a single query to merge live operational data in Aurora PostgreSQL with historical records stored in Apache Iceberg tables managed through AWS Glue Data Catalog, as well as Parquet and Iceberg data in Amazon S3 and S3 Tables, according to the announcement. Before this feature, joining recent transactional data in Aurora with historical records in S3 required reverse ETL pipelines that copied data, raised infrastructure expenses, and demanded continuous engineering work to maintain synchronization. The feature is now live in all commercial regions and AWS GovCloud (US) at no extra cost, with customers paying only for the additional Aurora compute and the S3 requests their queries produce. DuckDB and all components of the "Duck Stack," including DuckLake and Quack, will stay free and open-source under the MIT license, with the non-profit DuckDB Foundation maintaining its oversight of the projects.
Hannes Mühleisen and Mark Raasveldt, who created DuckDB and co-founded DuckLabs, will remain in charge of the team and the open-source project's technical direction as part of AWS. "Joining AWS gives us the resources and reach to bring this technology to many more developers and organizations, and to pursue ideas at a scale that would have been difficult for us to reach alone," Mühleisen and Raasveldt wrote in a blog post. Mai-Lan Tomsen Bukovec, technology vice president at AWS, said DuckDB was built to address the challenge of fast performance for the more than 90% of data queries that run on 1 terabyte of data or less, with optimizations especially relevant for agentic AI workloads where agents examine data and run small exploratory queries before choosing how to proceed.
Andy Warfield, distinguished engineer at AWS, explained that the teams strengthened their partnership over two years, with the S3 team making initial contact while working on S3 Tables. The collaboration deepened as the teams explored AWS Lambda as a natural fit for launching DuckDB queries rapidly and refined connections between DuckDB and AWS database and analytics engines through DuckDB's developing ATTACH and CONNECT commands. Warfield characterized the deal as a chance to expand the co-founders' influence rather than to absorb their work, noting that their vision aligns closely with Amazon's approach to data and would give them the widest possible opportunity to shape not just DuckDB but how analytics should be constructed and delivered in the cloud. Esra Kayabali, principal solutions architect at AWS, said the feature streamlines workloads that previously demanded building custom data pipelines, with the processing remaining inside Aurora without extra network transfers or ETL pipelines.
The acquisition positions Amazon to offer unified analytics directly within its operational database, removing the traditional boundary between transactional and analytical workloads that has long required separate systems and complex synchronization. The company can now target the vast majority of analytical queries that operate on smaller datasets while keeping DuckDB's open-source community intact under foundation governance, addressing concerns about the pace of innovation following the acquisition. The cloud giants have long struggled to balance the flexibility developers want with the simplicity enterprises demand, and embedding purpose-built analytics engines into managed databases may reshape expectations for what a database should do without requiring organizations to stitch together multiple services.