AWS Snaps Up DuckLabs as DuckDB Moves Into the Cloud Giant's Orbit

Amazon Web Services has acquired DuckLabs, the company behind the wildly popular open-source analytical database DuckDB. Financial terms weren't disclosed, which is standard practice for acquisitions where the seller doesn't want you doing the maths on how much they left on the table.

For the uninitiated: DuckDB is an in-process OLAP database written in C++. There's no server to run, no daemon to babysit, no separate process eating your RAM. It embeds directly into your application. The Python package, for instance, can run SQL queries straight against Pandas dataframes without moving a byte. Developers love it. Analysts love it. AWS apparently loves it enough to write a cheque.

DuckDB's origins are decidedly academic. It came out of Centrum Wiskunde & Informatica, a mathematics and theoretical computing research centre in Amsterdam, went public in 2022, hit 1.0 in 2024, and has been on a tear ever since. The more recent DuckLake project extended its reach into data lakes, using DuckDB as a metadata store to sidestep the painful "small changes" problem that has long plagued that space.

AWS wasn't exactly a stranger here. The two teams had already been working together on S3 Tables support in DuckDB, announced in December 2024. DuckLabs CEO and co-founder Hannes Mühleisen told us the collaboration had already produced "first-class support for S3 Tables in DuckDB" before any acquisition discussions presumably turned serious.

AWS VP and distinguished engineer Andy Warfield had been openly enthusiastic about DuckDB for at least a year. When DuckLake launched, he said the announcement had been "passed around broadly across the teams" and that engineers had been playing with it. That's AWS-speak for genuine excitement, not just marketing throat-clearing.

Now, the obvious concern: AWS buys open-source project, open-source project quietly suffocates under enterprise priorities. Mühleisen seems aware of this optic. He's keen to expand the role of the independent DuckDB Foundation, which will continue to oversee the project under its MIT licence. He's floated the idea of a Technical Advisory Board within the Foundation, giving organisations that build on DuckDB a formal seat at the table and a communication channel that sits outside AWS's direct influence.

This matters because DuckDB isn't just an AWS tool. It integrates with Google BigQuery and Google Sheets via extensions, among other things. Having the Foundation remain genuinely independent gives those stakeholders somewhere to go that isn't an AWS Slack channel.

The governance details, including voting rights and how exactly the community's voice feeds into development decisions, haven't been pinned down yet. Mühleisen and CTO Mark Raasveldt, who co-authored the database and leads the Foundation alongside Mühleisen, are still working through that.

Warfield, in the obligatory press statement, called DuckDB "completely open source and easily embeddable" and pledged continued development under the Foundation's stewardship. Whether those commitments hold five years from now when AWS's priorities shift is anyone's guess, but the structure at least gives the community something to point to if things go sideways.

The DuckLabs team stays in Amsterdam. The code stays open. The acquisition stays suspiciously cheap-looking given how much momentum DuckDB has built.