Big data’s roots are in the hyperscalers. As a result, most tooling we use for analytics today has been built around a scale-out approach where performance is sacrificed for the ability to handle enormous amounts of data. So although as an industry we’re very good at building heavy machinery to clean, format, and process petabytes of data, for modestly smaller workloads the tool chain is tremendously complicated, slow, and unfit for purpose. (When we say modestly smaller, we’re still referring to many terabytes of data, which is adequate for a large swath of analytics use cases.)
This is one reason why DuckDB has witnessed absolutely explosive growth over the last year.
Originally developed at Dutch research institute CWI (Centrum Wiskunde & Informatica), the open-source in-memory OLAP database takes a different approach. DuckDB focuses on a scale-up performance approach that allows it to run orders of magnitude faster than traditional scale-out approaches, and it can do so over terabytes of data. The petite size also lets it run anywhere — on-device, in-browser, or on a server.
To date, DuckDB has largely been an open source project that is downloaded and used by hundreds of thousands data scientists and analysts all over the world. However, this limits its broader potential as a truly serverless analytics engine that can be applied to any dataset, driven from cloud programs, and shared between users. That’s why we were so ecstatic when we met with Jordan Tigani, who told us about his vision with MotherDuck — a hosted cloud-service, built around DuckDB, that combines speed and ease-of-use to make analytics frictionless and ubiquitous.
After spending a decade building one of the most popular data warehouses, Google BigQuery, Jordan is one of the few people who understand the market deeply and realize the potential of DuckDB. Jordan has assembled an experienced, second-to-none team to tackle this problem. They have built highly performant, scalable data processing systems at Snowflake, Databricks, AWS, Meta, Elastic, and Firebolt, and are very well-respected in the data community. Combining industry know-how with the dedicated research of DuckDB Labs (spun out of CWI), MotherDuck is on a mission to make DuckDB one the most popular and performant databases in the world.
We’re incredibly excited and honored to partner with the MotherDuck team and lead their Series A, alongside great partners from Redpoint, Altimeter, Amplify, and Madrona. Together with the rest of the A16z data portfolio, we’re investing in a future of data analytics that will be much more composable, accessible, and fun.
***
- Investing in doxxnet Joel de la Garza and Martin Casado
- AI Can Write Code. Why Isn’t Software Better? Diogo Almeida, Ben Horowitz, and Martin Casado
- Aaron Levie, Steven Sinofsky & Martin Casado: How Do You Secure a World of AI Agents? Aaron Levie, Martin Casado, Steven Sinofsky, and Erik Torenberg
- Databricks CEO on AI Pacing, Cyber Risk, and the Enterprise Ali Ghodsi, Martin Casado, and Sarah Wang
- Fei Fei Li: The Race to Build World Models For AI Fei-Fei Li, Justin Johnson, Ben Mildenhall, and Martin Casado
- Investing in Preference Model Jennifer Li
- The Next Frontier of AI Video Is Control Gorkem Yurtseven, Batuhan Taskaya, and Jennifer Li
- Who Grades the AI Models? | Ben Horowitz & Rayan Krishnan Rayan Krishnan, Ben Horowitz, Jennifer Li, and Erik Torenberg
- Investing in Vals Jennifer Li, Yoko Li, Raghu Raghuram, and Shangda Xu
- Investing in Exa Sarah Wang, Jennifer Li, Stephenie Zhang, and Jason Cui