The startup behind open source tool Polars raises $21M from Accel

Polars, the Amsterdam-based company behind the popular open source project of the same name, has raised €18 million (about $21 million) in a Series A round led by Accel, with participation from Bain Capital Partners and angel investors. But while raising this kind of money is the dream of many developers, its creator Ritchie Vink didn’t set out to do so.

It all started as a pet project during Covid. Frustrated with the limitations of Pandas, a tool for organizing and working with data tables, Vink decided to build a better query engine in Rust. Fast-forward five years, and Polars is today widely used by data scientists and teams for its ability to process data much faster.

This combination of performance and popularity was what caught venture investors’ attention, but the Series A was spurred by Polars’ roadmap for becoming a scalable business. Two years after launching as a company, the company in February launched Polars Cloud, a managed data platform that lets users run queries in the cloud at scale.

“In the open-source community, the joke is that you can rewrite anything in Rust and it becomes better,” said Accel partner Zhenya Loginov, who led the funding round. “The reason that it’s a joke is that it is not a real sustainable advantage, and you need to do a lot more.”

For Viks and his co-founder, former Xomnia CTO Chiel Peters, doing more means building products around the tool, like Polars Cloud and Polars Distributed. The latter is a distributed engine that will support use cases involving petabytes of data, rather than small datasets, and is currently available in public beta. Building the new feature is what most of the funding will go towards, Vink said.

With Polars Distributed, the startup aims to challenge Apache Spark, whose creators founded Databricks.

For Polars, going after Pandas’ market share was enough to secure a $4 million seed round led by Bain Capital in 2023. But Pandas remains open source without a dedicated commercial platform, and even though Polars has surpassed 24 million downloads, the path to returns would be unclear if it weren’t for Polars Distributed and the ability to scale up from a single machine to a managed cluster.

While Loginov referred to this as a “more interesting” second market, he agreed with Vink that the value of Polars is in closing the scale gap between Pandas and Spark. “If you go into being able to process data sets of any size and complexity, you’re solving a lot of challenges for a lot of enterprises. So we felt that the ultimate market is potentially extremely large,” Loginov said.

Polars claims the core product is already used in production across finance, life sciences, and logistics. Still, Polars Cloud and Polars Distributed are opening a new chapter for the company. 

For other founders hoping to turn open-source projects into commercial ventures, Loginov points to a key lesson from Vink’s journey. “Polars became successful because it addressed a really big problem — he found a niche where the technology available today is just outdated by a mile. So I would suggest finding a large problem.”