AWS to acquire DuckLabs, owner of DuckDB, which remains open source
Embedded analytical database team remains in Amsterdam and projects continue under the MIT license, with the nonprofit DuckDB Foundation maintaining governance.

DuckLabs, the company behind DuckDB, announced on August 26 that it will join Amazon Web Services (AWS), with the deal expected to take effect in early September. The announcement was signed by co-founders Mark Raasveldt and Hannes Mühleisen on the DuckLabs blog.
According to the statement, the team remains together in Amsterdam and continues working on DuckDB, DuckLake, Quack, and the rest of what the company calls the "Duck Stack." The point that matters most to those who already use the technology: the open source components remain free and under the MIT license, with the nonprofit DuckDB Foundation maintaining curation of the projects.
What DuckDB is and why devs care
DuckDB is an embedded analytical (OLAP) database. The most common analogy is to call it "SQLite for analytics": it runs inside your process, with no separate server, and is optimized for heavy analytical queries rather than transactions. You can query Parquet, CSV, and JSON files directly with SQL, without spinning up infrastructure, which made it popular in data pipelines, notebooks, and command-line tools.
In practice, you can run something like:
SELECT categoria, SUM(valor)
FROM 'vendas/*.parquet'
GROUP BY categoria;directly in the CLI or inside a Python script, with no database process running in the background. This combination of full SQL, columnar performance, and zero deployment friction is what made the project grow.
The numbers DuckLabs itself cites give a sense of the adoption: more than one million downloads per day. The company was founded just over five years ago as a spin-off of CWI, the Dutch research institute, and grew to more than 30 people in Amsterdam, always under a bootstrapped model (no venture capital) controlled by the founders.
Why DuckLabs decided to sell
The statement is explicit about the reason. The founders say the current model was starting to show its limits: there was concern that DuckDB's growth would outpace a small company's capacity to provide support, turning DuckLabs itself into a bottleneck for the project, the team, and the companies building on top of it.
They also point out that scaling a large sales, support, and operations organization would divert attention from the technical work and the open source community, which were exactly what made DuckDB succeed. Commercial partnerships worked best with highly technical organizations, and reaching a broader audience would require solving more complete problems and investing much more in infrastructure.
According to the text, DuckLabs and AWS had already been working closely together for more than a year before the deal. AWS, according to the statement, has committed to supporting the continued development of DuckDB and its community over the long term.
Andy Warfield, Distinguished Engineer and VP at AWS, said in the statement that DuckDB is "widely used and much loved by S3 customers today" and that he has worked with the DuckLabs team for about two years. The mention of S3 is no coincidence: querying data in object storage is one of DuckDB's central use cases.
What changes and what stays the same
DuckLabs directly listed what remains:
- DuckDB, DuckLake, Quack, and the other open source components remain free and under the MIT license;
- the DuckDB Foundation remains the curator of the projects and holder of the intellectual property for open source DuckDB;
- the team remains in Amsterdam and will continue contributing to the project.
Peter Boncz, of CWI and a board member of the DuckDB Foundation, reinforced that the foundation, created when DuckLabs spun off from CWI, holds all the IP for open source DuckDB and will continue to hold it, and that the foundation will ensure that community voices continue to be heard.
On the expansion side, the company previewed two concrete developments. The DuckDB Foundation will get a technical advisory board, so that prominent community members can weigh in on the project's technical direction. And there's a plan to open up the extension stack, so that extensions signed by other developers and organizations can run on DuckDB, something relevant for anyone who wants to distribute their own extensions without depending exclusively on the official channel.
Reactions from the ecosystem
Two voices that depend directly on DuckDB commented in the statement. Jordan Tigani, CEO of MotherDuck (a company that builds a cloud service on top of DuckDB), said that Amazon's weight will add "a lot of momentum" and strengthen the ecosystem. George Fraser, CEO and co-founder of Fivetran, said Amazon is "the ideal home" for DuckLabs and highlighted the company's track record of working with the entire cloud ecosystem.
From the academic world, Torsten Grust, professor at Universität Tübingen, pointed out that DuckDB's open nature and "hackability" made it one of the leading platforms for database research and teaching, and welcomed the fact that the project remains open source under AWS's umbrella.
What this means for developers building software in Brazil
For Brazilian developers who already use DuckDB in pipelines, notebooks, or as an embedded analytical layer, nothing changes in the short term: the binary remains MIT, free, and maintained by the same team. Those using the open source version don't need to react to anything right now.
The point to watch is medium-term. With DuckLabs inside AWS, it's likely that managed services and deeper integrations with AWS products will appear, especially around S3. This can be good for those already in the Amazon ecosystem, but it puts on the table the classic question raised by hyperscaler acquisitions of open source projects: how much of the engineering effort goes into the open core, and how much goes into features that only make sense inside the buyer's cloud.
The existence of the DuckDB Foundation as IP holder and the public commitment to the MIT license are the guarantees the community has today. What remains open is how the technical advisory board will work in practice, what the pace of investment in the open core will be under AWS, and what commercial services will emerge from the deal. DuckLabs itself says there are "many details still to be defined" and that it will share more as the plans move forward.
Translated from the Brazilian Portuguese original · Read the original
Perplexity swaps DynamoDB for in-house database and cuts latency by 5x
The company behind the AI-powered search engine migrated its serving layer to CobbleDB, an internal database written in Rust, and cut batch read latency by up to 5x while saving at least 20% on storage.