Databricks CEO Ali Ghodsi is playing the long game to unseat Snowflake as the darling of the data world.
Snowflake made a name for itself helping companies use stored information to drive deeper analytics. Tackling that market, which could be worth $35 billion by 2025, helped propel the company to a historic IPO in September. But its stock has drifted downward since.
Databricks, which is plotting its own public offering, is wading deeper into Snowflake's territory with a new product that lets customers query data for more basic statistical analyses using SQL, a programming language that Snowflake also relies on.
Meanwhile, Snowflake is laying the groundwork to let its users tap the same analytical data to power artificial intelligence-backed algorithms.
That means the two, which until now were able to largely play nice together in the enterprise-data sandbox, may soon find themselves flinging their toys at each other. But while Ghodsi acknowledged that Snowflake has the advantage right now, he argued that will soon change.
"Operational AI is going to be a much bigger market," Ghodsi told Protocol. "It's not a bigger market right now than data warehousing, but over the next 10 years it will be."
In Ghodsi's view, it's much easier to take a system that can support advanced analytics and infuse the ability to do basic statistical work than trying to tackle it from the opposite direction, which he argued is Snowflake's struggle right now.
Snowflake does offer users several features, like data pipelining, which are important foundations for advanced algorithms. It also recently added support for Java and Python, two of the most popular AI programming languages, in an attempt to bring more data scientists on board. Currently, however, Snowflake's tools are more suited for data analysts — and will likely remain that way for a while.
Snowflake readily admits it isn't trying to replicate Databricks's model. Instead, the company is relying on integrations with AI platforms from Google, Microsoft and AWS, according to SVP Christian Kleinerman.
"Do you have algorithms that are natively hosted by us? No. But it's not because we can't or because we don't know how to. It's because we know the space is very fluid," he told Protocol. "Our entire initiative in AI and ML has been to build extensibility into Snowflake so you can interface with your tool of choice."
Snowflake was able to capitalize on the rush among enterprises to empower more of their employees to access data and conduct more advanced statistical analyses. But building AI models is much more difficult, as it involves training algorithms to begin to draw future predictions or discover unknown correlations from those vast data sets. Regeneron, for example, used Databricks to find a genome for chronic liver disease.
And there's clearly a lot of enthusiasm for Ghodsi's vision. Databricks has raised $1.8 billion — $1 billion of that in February — and is valued at an astonishing $28 billion. (Ghodsi even believes Databricks is undervalued at this point.) The company is also backed by Salesforce, AWS, Microsoft and CapitalG, a venture fund under the Alphabet umbrella alongside Google. Some of those giants have their own AI engines — AWS has SageMaker and Microsoft has Azure ML, for example — so backing Databricks is a good proof point of just how powerful its platform is. It's also an indicator that, while the major cloud providers partner with Snowflake, they may see longer-term value in a deeper relationship with Databricks.
Still, with a market cap of $57 billion, Snowflake is a much larger company. While Databricks's value is likely to skyrocket when it IPOs, Snowflake definitely has the first-mover advantage. And given the time it will take to establish operational AI as a full-fledged market, the company has some runway. There's also a plethora of AI startups Snowflake can pick from and, with its equity and a war chest of $4 billion, it's not short on stock or cash it can use to acquire tech to help support the pivot. The company, for example, recently invested an undisclosed sum in Dataiku and has an equity stake in DataRobot.
The market is ultimately going to be large enough to support both. But most software vendors are never content with second place, which means the fights between Snowflake and Databricks are likely just beginning.