Follow Datanami:
January 6, 2021

Starburst Rides Presto to a $1.2B Valuation


Starburst today announced the completion of a $100 million Series C round of funding led by Andreessen Horowitz. With a $1.2 billion valuation and $164 million in total funding, Starburst, which is a big backer of the open source Presto query engine and sells an enterprise version of it, now has additional resources to meet the surging demand for cloud data analytics.

As we wrote yesterday, Presto is on the cusp of breaking out in a big way, largely driven by the changing nature in how companies build analytic systems. Cloud-based data warehouses have been taking share from on-prem systems for years. But in 2020, what was a steady stream of movement to the cloud suddenly became a waterfall.

Through its support for Presto, Starburst is well positioned to capitalize on the mass migration of data to the cloud, where data is stored in a variety of different stores. As a federated query engine, Presto essentially is a data warehouse without its own a database, which means users can analyze data sitting across databases, data lakes, distributed file systems, and streaming data systems.

Presto has come a long way from 2012, when it created at Facebook as the follow-on to Apache Hive. Back in those days, Facebook used Presto to query data stored its massive Hadoop systems. While the social media giant has moved away from Hadoop, it still utilizes a version of HDFS to store petabytes of data (it also uses Alluxio in an interesting way to accelerate Presto queries).

Today, Starburst employs the four original creators of Presto, including Martin Traverso, Dain Sundstrom, David Phillips, and Eric Hwang. In a recent interview, Traverso told Datanami that ability to query data wherever it sits is Presto’s main advantage versus cloud data warehouses, which typically require an ETL step to move and prepare data for analysis.

“There’s a center of gravity that’s shifting, but I don’t think we’ll ever get to a point where data will be centralized. It just doesn’t make sense,” Traverso said. “There are different use cases, different reasons why you might need to store data in different places.”

For example, you wouldn’t store data in a cloud object store if you needed to perform point lookups or modify data on a very granular basis, he says. “There are better technologies for that, like Cassandra, or SQL databases,” he said. “Some data will have to have to stay in those systems because it just makes sense.

“What’s important is being able to have access to all that data easily without having to go through complex migration processes and transformation processes,” he continued. “Being able to access them where they are is important.”

Companies that are adopting a multi-cloud strategy will also find advantages in being able to query data wherever it sits. Essentially, wherever there is need for flexibility and a desire to push back against centralizations of data in a single cloud–or any single data store, database, file system, etc.–then Presto will bring benefits to the user.

Starburst advocates the “analytics anywhere” benefits of Presto

“We think that customers need to have options,” Traverso said. “If you try to force in a single direction or a single technology, that is going to be a risk for customers at some point. They may want to have options around where they put the data, how they manage it, and to support different use cases. There’s no single solution that satisfies everything.”

Presto offers connectors for cloud-based object stores, as well as NoSQL databases. The software supports the capability to join data from multiple sources as part of the query, which is another useful feature. According to Traverso, Presto can also query data that is being streamed through Apache Kafka and Amazon Kinesis, which just adds to the tool’s usefulness.

“That’s one of the unique features of Presto, you can query from Kafka and at the same time query from some other source, and join them in the same query,” Traverso said. “You can read from Kafka and join against a MySQL database, for instance.”

Starburst officially was created in 2017, when it was spun out of Teradata. The San Diego data warehouse giant nurtured Presto for several years following its 2014 acquisition of a company called Hadapt, of which Borgman was the CEO. But it ultimately decided that Presto was not a good fit for the large, Fortune 500 type companies that it caters to. Ironically, Starburst has found success among those types of companies, counting Comcast, Walmart, and the Financial Industry Regulatory Authority (FINRA) among its clients.

Starburst eschewed venture capital in the past, as its executives preferred instead to concentrate on building a profitable business without outside assistance. “When you don’t take VC money, you look smaller than you are,” Borgman told Datanami in 2019, when the company had about 30 employees and 50 paying customers.

Starburst CEO Justin Borgman

As the company has grown, that reluctance to take VC money has dissipated. In June 2020, Starburst completed a Series B round of funding worth $42 million. Around the same time, Starburst got some competition in the Presto space in the form of Ahana, which launched with a seed round of $2.25 million from GV (formerly Google Ventures). Ahana also was a backer of the new Presto Foundation, the organization dedicated to managing the open source PrestoDB project.

The Presto distribution that Starburst backs, PrestoSQL, is a fork of the original PrestoDB strain that emerged from Facebook. Starburst also backed the Presto Software Foundation to provide governance to the PrestoSQL distribution. In late December, following Facebook’s decision to copyright the name Presto, the backers of PrestoSQL changed the name of the distribution to Trino, and changed the name of the foundation to the Trino Software Foundation.

With $100 million in the bank, courtesy of Andreessen Horowitz, one of the preeminent venture capitalists in the big data space, Starburst is primed to grow its Presto (ah, Trino) business to new heights.

“To succeed in today’s digital economy, organizations need to break through the limits previously experienced when accessing their most critical asset–their data,” David George, a general partner at Andreessen Horowitz, said in a press release. “With its disciplined approach, long-term vision, and workhorse mentality, Starburst is building the technology required to unlock the value of all data. By enabling organizations to access data wherever it resides, Starburst is providing the foundation for the future of the data-driven enterprise.”

Related Items:

Presto Poised for a Breakout Year as Data Explosion Continues

Will the Presto Community Ever Be United Again?

Building Presto Business No Magic Trick for Starburst