The AWS DuckDB acquisition everyone's talking about is actually the DuckLabs acquisition
AWS didn't acquire DuckDB. AWS acquired DuckLabs — the small, bootstrapped company in Amsterdam that employs the core team of 30-something people who built it. DuckDB itself stays with the DuckDB Foundation, under the MIT license, just like it's always been.
That distinction matters, and most of the coverage today is getting it wrong. But the nuance isn't a loophole, and the worry in the HN comments isn't baseless either. Let me walk through what this actually changes, what doesn't change, and where the real risk sits.
The numbers first
- Bootstrapped from day one. DuckLabs was founded ~5 years ago without venture capital. A technical, founder-owned company focused on the technology.
- 1M+ downloads per day. DuckDB's adoption curve went from "cool academic project from CWI" to "default analytical database for data engineers" in about 5 years.
- 10,000+ commits since v1.0. The project shipped v2.0 nine days ago with DuckLake, Quack (a server protocol for DuckDB), a stable extension ABI, and a new Foundation advisory board.
- 30 employees in Amsterdam. Not a big company. A tight group that solved a hard problem: make analytical queries fast on a single machine without requiring a cluster.
What the DuckLabs acquisition actually means
DuckLabs got to a point where their own growth was going to outpace their ability to support it as a small bootstrapped company. The founders said this explicitly:
"We worried that DuckDB's growth would eventually outpace our ability to support it. That our small company could become a bottleneck for the project."
They had a choice: scale DuckLabs into a much larger sales and support organization — pulling focus away from the engineering — or find a partner who could carry that weight. They chose AWS, where they'd been collaborating for over a year (since the S3 Tables preview in 2024).
What is and isn't changing
Doesn't change:
- DuckDB stays MIT license. The Foundation holds the IP.
- The team stays in Amsterdam. Same people, same building.
- DuckDB continues to serve every vendor and platform, not just AWS.
- Quack, DuckLake — all the open-source projects stay open.
- The v2.0 roadmap with the advisory board, stable ABI, and extension ecosystem proceeds as planned.
Changes:
- The DuckLabs team now has AWS salary and benefits instead of bootstrapped company runway.
- AWS's infrastructure and distribution scale apply to whatever managed services AWS builds around DuckDB.
- The foundation gains a formal advisory board (announced in v2.0, timed or not) with community stakeholder input.
- Extensions signed by third parties will be allowed in DuckDB — the "open extension stack" was announced as part of this deal.
Why the DuckDB acquisition pattern worries everyone on HN
Everyone on HN this morning is saying the same thing: "we just can't have nice things." The worry is understandable. We've seen this movie before:
- Sun acquired MySQL. MySQL is still around. But it's also not the default choice it was in 2008.
- Oracle acquired Sun. Java survived. But the governance model changed.
- Microsoft acquired GitHub. GitHub is arguably better now. But the skepticism was real.
- IBM acquired Red Hat. Red Hat's culture shifted.
The pattern isn't "big company destroys acquisition." The pattern is "big company's incentives slowly diverge from the community's."
AWS's incentive is to sell AWS services. If DuckDB becomes a gateway drug to S3, S3 Tables, and a managed DuckDB service, that's great for everybody — as long as playing well with others doesn't become a lower priority than playing well with S3.
The Foundation structure matters here. The DuckDB Foundation owns the IP, not DuckLabs, not AWS. AWS can't change the license without the Foundation's approval. And the Foundation just added an advisory board specifically to give community stakeholders a voice. That's not nothing. It's also not a guarantee — governance structures are only as strong as the people willing to enforce them.
The MotherDuck elephant
You can't talk about this acquisition without talking about MotherDuck. Jordan Tigani (MotherDuck CEO, ex-Google BigQuery) wrote a post called "DuckDB outgrows its nest" today. His quote from the announcement:
"Amazon putting its weight behind DuckDB is going to add a ton of momentum and strengthen the ecosystem. This is great news for those of us who believe in DuckDB as the platform on which the future of analytics is being built."
That's diplomacy. The real story is that MotherDuck — a company that built its entire business on DuckDB-compatible cloud analytics — just saw its upstream vendor get acquired by its biggest competitor (AWS has Redshift, Athena, and now a direct line into the DuckDB core team).
MotherDuck's actual countermove happened yesterday: they acquired Tower, a data agents company. And they've been investing heavily in their own managed DuckLake lakehouse, Flights (AI-native runtime), and Dives (dashboards). The message is clear: MotherDuck doesn't depend on DuckLabs the company; it depends on DuckDB the open-source project, and that stays open.
But if you're a startup building on DuckDB today, this acquisition means your primary platform vendor and your cloud provider just merged. That's a concentration risk you need to factor into your planning, whether or not you think it matters today.
The real winner: the DuckDB core
Here's the take that's not getting enough airtime: the DuckDB core team of 30 people just got AWS-scale resources without having to become a sales organization. They get to keep building DuckDB, Quack, and DuckLake — but now with AWS's infra budget instead of bootstrapped runway.
DuckDB v2.0 shipped nine days ago with:
- Quack: a server protocol that turns DuckDB from an in-process library into a networked database
- DuckLake: a lakehouse format built on S3-compatible storage
- Stable extension ABI: extensions no longer need to rebuild for every DuckDB release
- 10,000+ commits of improvements across the board
That's a lot of momentum. AWS isn't buying a stalled project — they're buying a rocket ship that was going to take off regardless, and they want a seat on it.
What to watch
Three things to track over the next 12 months:
Does the Foundation's advisory board actually have teeth? The members are supposed to provide input on technical direction. If the board is stacked with AWS-friendly picks, the Foundation structure becomes theater.
Does DuckDB's S3 integration get better than its other storage integrations? If querying data in GCS or Azure Blob becomes a second-class experience relative to S3, that's how vendor capture looks in practice.
Does the open extension ecosystem stay neutral? DuckDB's extension system is one of its best features. If AWS starts prioritizing first-party extensions over community ones, that's a yellow flag.
The bottom line
For the practicing data engineer: you don't need to migrate off DuckDB today. The MIT license isn't changing. The Foundation owns the IP. The team is still the same people.
But open-source acquisitions have a half-life of trust measured in years, not days. AWS knows this. That's why they're making all the right promises on day one. The question isn't whether they mean them today — it's whether they still mean them in 2028, when a VP wants to know why DuckDB keeps shipping features that make it easier to use on GCP.
Watch the extension stack. Watch the advisory board. Watch the S3 integration. That's where the real story will show up, not in the press releases.
Sources: DuckLabs announcement, AWS announcement, DuckDB v2.0 highlights, MotherDuck "DuckDB outgrows its nest", Hacker News discussion
Top comments (0)