Pola.rs €18M Series A from Accel

2 hours ago 2

At Polars we have big plans. There is much about data processing that must be easier, more intuitive, faster and frictionless. To make this happen, we have raised an investment round of €18 million, led by Accel with participation of BCV which followed up on their seed investment.

What we will do with the funds

With our seed investment, we have grown adoption from 250k to over 23M monthly users. Polars’ capabilities have increased tremendously with faster data-processing, streaming execution and much better cloud interopability. We also launched Polars Cloud, the data platfrom that will bring low-latency, distributed data processing, fully integrated in Polars. This is just a fraction of what we have implemented in the last 2 years. Our seed unlocked a great amount of what we envision for Polars. With this follow up investment we expand on this vision. We want to:

  • Make Polars OSS fully streaming ensuring that single-node queries utilize hardware to their maximum capacity.
  • Continue our development of a state of the art distributed engine, able to run all Polars queries in the cloud and on-premises. Delivering one intuitive DataFrame API for all scales.
  • Build a one stop data platform that provides the best Polars experience available. This includes managed hardware, autoscaling, query insights, profiling and much more.

Get started with Polars Cloud and Polars Distributed

With Polars, we cracked the code if single-node data processing. If you can scale vertically, you should use Polars. We have designed Polars to leverage the resources at hand to their maximum potential. With Polars Cloud, we think bigger. You should not be rewriting DataFrame API’s depending on the scale you want to process. Polars will offer a single API that will be optimal for all scales. We recognize that often you don’t need a tool like PySpark. We want to ensure, you never need it. The code you’ve written once, will scale with your needs, only a remote call away.

with pc.ComputeContext( ... ) as ctx: ( pl.scan_parquet("<my-dataset>", .group_by("keys") .agg( count_order=pl.len() ) .remote(ctx) .execute() .show() )

Above is a simple code example to illustrate how Polars queries can be seamlessly transtioned into distributed queries. Polars Cloud is currently live on AWS, use the link below or sign up for on-premise.

Our commitment to Open Source

Polars is an open source company and the original goals of OSS Polars are very important to me. When I started Polars, the goal was to build a faster DataFrame library that could replace pandas and improve on many of the footguns I had experienced. Soon after, that expanded into building a state-of-the art single-node query engine specialized for DataFrames.

The Polars company is dedicated to preserving those roots while extending them further. We are now bringing the power of OSS Polars to managed clusters and distributed computing. By leveraging the OSS Polars streaming engine on the worker nodes, we ensure that OSS Polars continues to improve, and that progress directly strengthens both Polars OSS and the company’s offering.

Read Entire Article