SingleStore raises $80M for distributed SQL database
Raj Verma, CEO of SingleStore, explains why the vendor rebranded from MemSQL and how its platform is more than just an in-memory database, but is a general-purpose database.
Database vendor SingleStore said on Dec. 8 that it has raised a Series E round of funding, bringing in $80 million.
The San Francisco-based vendor was founded in 2011 and until Oct. 27 of this year was known as MemSQL. The core technology behind SingleStore is a distributed SQL database platform that can be used to enable analytics workloads and applications.
In October 2019, while it was still known as MemSQL, the vendor introduced a capability it referred to as SingleStore, which brings together both rowstore and columnstore tables in an effort to help enable faster database queries.
Leading SingleStore is CEO Raj Verma, who joined the company in July 2019 as co-CEO, becoming the sole CEO in September 2020.
In this Q&A, Verma talks about why SingleStore is growing, the impact of the coronavirus pandemic and where the vendor is headed in the future.
Why raise $80 million now and is it the last step on a path to an IPO?
Raj Verma: There are a few reasons. One is that the data market has evolved tremendously over the course of the last few years and specifically over the last year or so, in the pandemic era. Organizations want insights that are a lot deeper, because that's probably the only way they have to make informed decisions.
Analytics is becoming a huge area of excitement and demand in the tech marketplace, and that's why you see some very lofty valuations for vendors like Snowflake. From our perspective, we are seeing unprecedented interest in SingleStore and we feel that if we make the right investments right now, we could capture a bigger part of the market and capture that quicker.
SingleStore CEO Raj Verma
I don't think we've got a plan that's set in stone which says that by a certain period, we would go public, I don't think many companies do that. In any case, we have to sort of explore how we proceed, but we're in no real hurry. I guess, if we're adding value to our customers and building great technology, then good things will happen. If that is an IPO, then it is an IPO and we will do it at the right time.
What changes have you seen in the way organizations use and think about data in the pandemic era?
In the COVID-19 environment, the analytics reporting dashboard insights that companies seek are remarkably different.
Raj VermaCEO, SingleStore
Verma: In the pre-COVID-19 era, data was more of a tail light, so it was all right to take a look at what happened yesterday and learn from those strengths.
In the COVID-19 environment, the analytics reporting dashboard insights that companies seek are remarkably different. Now, instead of running 10 or 20 queries to generate a report, organizations are using data as the headlights to try and see through the fog and the mist. To do that you're sometimes running between 250 to 400 queries to generate a report. That report is then viewed and taken into consideration by executive management. If there are 10 seconds of latency per query, you have your CEO sitting staring at an empty screen for hours and that's not acceptable. So organizations want faster insights.
What was the rationale behind the rebranding of MemSQL to SingleStore?
Verma: With the MemSQL name we were telling customers that we were in an in-memory SQL database. We started as an in-memory database, but we are no longer only that. If we were only an in-memory database, our revenues would be probably 10% of what they are right now. So we are a lot more now; we're a general-purpose database.
Our grand vision was to get rid of data silos and have organizations and their data strategists view data as a single source of insight, rather than siloed data. With SingleStore the idea is one store for all your data needs.
Are there any particular lessons learned from your experience in the Hadoop big data market that can inform SingleStore and distributed SQL?
Verma: I think there are enormous lessons to be learned from that marketplace. Fundamentally, the premise of Hadoop was to be the storage layer for very large amounts of data. Hadoop enabled users to be able to work with very large data sets, but Hadoop wasn't built for low latency.
Over the last five years or so Hadoop has had a tough run and the reason for that is that people don't only want a dumb layer to store data, they actually want very quick and deep insights.
SingleStore has bet on the fact that the need for low latency and high concurrency, with the ability to query using SQL, is the way forward.
What's next for SingleStore and its distributed SQL database vision?
Verma: We are very much looking forward to our second release of cloud product with SingleStore Managed Service version 2.0.
I don't want to give away too much, but we are working on changing the way analysts look and work with data, giving them the ability to switch on and off data types to get various insights.
