Firebolt Analytics Inc., a startup led by former executives from analytics unicorn SiSense Inc., today exited stealth mode with a cloud data warehouse that it says can provide more than 100 times better performance than rivals.

Firebolt, based in Tel Aviv, is backed by $37 million from investors that include Bessemer Venture Partners, Zeev Ventures, TLV Partners and Angular Ventures.

The company provides a cloud data warehouse designed to enable rapid analysis of large data sets. The platform, the startup claims, is the first in the industry that provides subsecond performance for queries carried out against terabytes or petabytes of information. Firebolt says the fast response times are the result of its platform’s ability to crunch data at speeds up to 182 times higher than competing data warehouses.

The kind of near-real-time query completion the startup promises has appeal for enterprises. Analytics teams working with large data sets can often run their queries against only a subset of the available information because of performance constraints, which limits the insights that can be gleaned. Increasing performance makes it possible to increase the amount of information it’s practical to analyze at once.

“While companies can store massive amounts of data, most organizations are only able to analyze a fraction of that big data, and often find themselves looking at stale data that does not reflect the current state of their business,” said Firebolt Chief Executive Officer Eldad Farkash, who had previously co-founded SiSense. That startup, which provides business intelligence software, was valued at more than $1 billion after a funding round earlier this year. 

Firebolt says that its data warehouse’s performance is the result of several internally developed technologies implemented behind the scenes. One is a custom cost-based optimizer that tweaks the SQL queries entered by users to improve their speed. Additionally, Firebolt’s data warehouse employs an approach called sparse indexing to lower the amount of unnecessary data that users’ queries retrieve, which improves performance by reducing superfluous computations.  

The startup provides its data warehouse as a software-as-a-service offering running on Amazon Web Services. Companies are charged the base cost of the compute, storage and networking resources they use, plus an annual subscription that isn’t dependent on how much data they analyze 

The startup claims that its pricing model makes its platform more cost-efficient than competing data warehouses. One “large marketing tech company,” Firebolt claims on its website, managed to perform a data ingestion task that cost $1,800 on a rival platform for just $50. 

Image: Firebolt