Denys Rudyi - Fotolia
Speed is the central focus of Firebolt, a new cloud data warehouse vendor that has emerged from stealth after securing $37 million in Series A financing.
Zeev Ventures led the funding round with participation from TLV Partners, Bessemer Venture Partners and Angular Ventures.
Firebolt's launch on Dec. 9 came three months after Snowflake, a cloud data warehouse vendor founded in 2012, showed the demand for cloud data warehousing technology when it broke records with its initial public stock offering. Snowflake raised $3.4 billion with its IPO -- easily surpassing the previous record for tech companies held by VMWare, which just under $1 billion in 2007.
The need for speed
Founded in 2019 and based in Tel Aviv, Israel, Firebolt is a cloud data warehouse designed to handle terabytes to petabytes of data to enable customers to work with their data in less than a second.
It's that desire for speed that drove co-founders Eldad Farkash and Saar Bitner, both of whom previously helped build BI and analytics vendor Sisense, to develop Firebolt.
Farkash, Firebolt's CEO, said that when he spoke to data engineering teams, they constantly asked how they could be enabled to build data features and get them ready for use within their organizations more quickly.
"Firebolt is about speed and efficiency with the elasticity you would only get from a big vendor," Farkash said. "Firebolt is doing the same thing from an elasticity perspective, but the speed and efficiency is new and different."
Eldad FarkashCo-founder and CEO, Firebolt
According to Farkash, the big cloud data warehousing platforms he referred to such as Amazon Redshift, Microsoft Azure, Google BigQuery and Snowflake were developed while the cloud migration trend was just getting started and were thus built to solve problems users faced during that period. Among them, he said, were formatting, replacing files and dealing with scale.
Firebolt, meanwhile, was developed to address problems data teams face in the current environment, many of which revolve around getting data ready for consumption as fast as possible so organizations can make data-driven decisions as close to real time as possible, Farkash said.
That, he said, is what differentiates the startup from existing cloud data warehouse vendors.
"In the past, data-driven features were very limited and we would massage the feature as much as we could so the data would be properly served, but things have changed dramatically," Farkash said. "Users are expecting so much more, and companies that are data-driven are basically depending on the data. It's companies that are selling to traditional marketplaces but were born in the cloud and are using data in completely different ways."
Differentiation for startup vendors, meanwhile, comes down to three potential factors, according to Dave Menninger, research director of data and analytics research at Ventana Research.
"To bring a new product to market, it needs to be faster, better or cheaper," he said. "Typically you can pick two out of three. It sounds like Firebolt has chosen faster and cheaper."
The investor's perspective
While speed is what Farkash claims differentiates Firebolt from established cloud data warehouse vendors, it's also what attracted investors.
Gil Dibner is founder and general partner of Angular Ventures.
His firm, he said, receives more than 3,000 pitches from startups each year yet only invests in six -- less than one in every 500. The ones Angular Ventures invests in, he continued, generally have an experienced founding team with deep expertise in their domain, no significant barrier to entry in their market and are taking advantage of a significant trend.
"Firebolt has all three of those things," Dibner, who had a preexisting relationship with Farkash dating back to Farkash's work at Sisense, said, "They are positioning themselves at the intersection of a bunch of trends. I was going against the grain when I backed him the first time, but I think now it's a different story and they've built a lot of interest in this investment."
That interest, Dibner added, is ultimately based on Firebolt's speed, which he called a breakthrough.
"When you talk about a technical breakthrough, people talk about order of magnitude," Dibner said. "With Firebolt, they literally are talking about one or even two orders of magnitude on a number of these axes."
While not the lead investor, Dibner said that Angular Ventures' investment in Firebolt is its largest investment to date, and that should Firebolt establish a foothold in the cloud data warehouse market and attract further funding, Angular Ventures will continue to participate.
"We're super excited to continue to support Firebolt on its journey," he said. "We're all waiting to see the response of people when they start using it, but the early responses I've heard have been extremely positive. That's what triggered the round."
When it comes to Firebolt's key differentiator, one of the ways the platform is faster and more efficient that existing cloud data warehouses, according to Farkash, is that Firebolt forces users to sort their data when they ingest it into Firebolt. With the data sorted upon ingestion, Firebolt is able to automatically apply new coding to the data that compresses the files.
The result is that when users run queries, the number of terabytes scanned is smaller than it would be without the files being compressed, and that results in greater speed.
"The sorting, the indexing, the compression is all hidden," Farkash said. "Customers don't need to deal with it. We bring brutal simplicity to the game."
In addition to speed, Firebolt touts a significant pricing advantage over other cloud data warehouse vendors.
In fact, according to Menninger, pricing is an area in which there may be real opportunity for startup vendors.
"I have heard complaints about Snowflake costs as workloads scale," he said. "Snowflake is taking steps to address this, but it may also create an opportunity for Firebolt."
Firebolt offers a consumption-based pricing structure with customers paying only for the amount of storage they require and the number of queries they run on their data. In addition, for larger customers whose pay-per-use costs have the potential to be significant, Firebolt offers subscription-based pricing.
Exact cost details, however, are not listed on Firebolt's website, and Farkash declined to give specifics when asked.
"We've been expanding very quickly over the past few months," Farkash said. "The market is hungry. Because of COVID -- it started before that, but companies today look at data that is something that is about surviving."