September 20, 2023

MotherDuck: helping distributed data fly

In the 2010s, “Big Data” engendered a lot of apprehension amongst large engineering teams. There was a flood of data coming; you had to keep every single piece, and if you didn’t have a plan, you would be overwhelmed. 

This idea of needing the ability to query all of your data at any point is a relic of a bygone era. A better way exists. In this case, analysts need a database that lets you work with data fast and efficiently regardless of where it’s stored—locally or in the cloud. A database that dynamically flexes up and down to accommodate the amount of data being stored and queried. At Felicis, we believe that DuckDB is the database that was promised, and we believe the company best positioned to commercialize this incredible technology is MotherDuck, whose Series B we have recently led and are announcing today.

DuckDB is well equipped to solve many data challenges due to its architecture–it is fundamentally a low-overhead, in-process database with no external dependencies. It intelligently uses the processing power of a local device to quickly and seamlessly process data for various use cases. The technical infrastructure to support DuckDB has emerged in the last few years with the advent of the M1 chips on most Macbooks and network capabilities required to support fast querying at the edge. 

MotherDuck’s hybrid data architecture (edge + cloud) is compelling for enterprises that want to implement DuckDB locally and store a portion of their data in the cloud–a truly modern and novel distributed system for the serverless world. We believe this approach will help realize tremendous cost savings and enable faster performance for data teams.         

Led by ex-BigQuery veterans (including a tremendous supporting cast from AWS, Databricks, Elastic, Facebook, Firebolt, Neo4j, SingleStore, Snowflake, and more), the MotherDuck team is intimately familiar with the large-scale databases. The CEO, Jordan Tigani, wrote in his piece, Big Data is Dead, that his job for several years was to tout the efficacy of storing and querying more and more data. This experience led to a realization that (1) most customers don’t need the amount of data they store, and (2) most queries just aren’t that big (as Figure 1 below shows).

Source: Big Data is Dead, Jordan Tigani.

We’re incredibly excited to work with Jordan, Tino, Ryan, and the MotherDuck team, as well as investors a16z, Redpoint, Madrona, Amplify, and Altimeter in helping usher in a new age of on-demand analytics available locally or in the cloud. MotherDuck is now open to everyone who works with data, you can read all about this momentous event on their site and more about the announcement in TechCrunch. The era of serverless data analytics for the masses is here.