Enterprise software maker Databricks found itself in the catbird seat when much of corporate America scrambled to make artificial intelligence a bigger part of their business strategies. But competition has also intensified with rival Snowflake and Palantir Technologies as they help companies make their proprietary data actionable in new ways.
Torrid growth has put San Francisco-based Databricks, a maker of data analytics and data management software, on Wall Street's radar. But the software maker hasn't announced plans for an initial public offering yet and it might not happen in 2025.
What's clear is that Databricks, founded by seven data scientists from the University of California, Berkeley, is on a roll, with revenue growing at 60% year-over-year.
At UC Berkeley, Databricks founders created Apache Spark — open source software for data storage and analysis. Started in 2013, Databricks quickly won over computer engineers with its software, optimized for cloud computing and more advanced than traditional relational databases.
Already growing fast, the company moved quickly to capitalize on the emergence of "generative"AI and ChatGPT-type conversational co-pilots. Many companies now want to search their proprietary data to gain new business insights and a competitive edge.
Custom AI Model Training Solutions
But first, corporate data needs to be organized and made reliable so it can used for training AI systems. Databricks offers tools for companies to customize off-the-shelf, large language models (LLMs) with their proprietary data.
"We started in the machine learning, data science world. We were doing AI before AI was cool," Databricks Chief Revenue Officer Ron Gabrisko told IBD in an interview. He added: "We have several hundreds of millions of dollars (of revenue) in the AI, machine learning, data science space."
Databricks continues to forge partnerships. On Feb. 13, it announced a pact with business software maker SAP. Germany-based SAP will integrate Databrick's AI data management tools into its new business data cloud.
Databricks expected its annualized revenue run rate to cross $3 billion in the December quarter. In the September quarter, its new data warehousing product achieved a $600 million revenue run rate, growing 150%.
Gabrisko rattled off a few examples of Databricks customers in the AI space. Biotech firm Regeneron discovered a new liver cancer cure using genome data and Databricks tools, he said, while large retail customers predict inventory needs.
Other Databricks customers using AI include Mastercard, electric vehicle maker Rivian and Shell, he added.
"We call it data intelligence. AI isn't really valuable without your data, at least in the enterprise," Gabrisko added. "The new trend with LLMs is making it easier to interact with data sets. We think the market is going to consolidate around this concept of data intelligence where you have all these AI use cases."
Databricks IPO May Wait
Wall Street is paying attention.
"Databricks has become the key partner for Fortune 500 customers in their gen AI journeys and is one of the hottest private companies in tech," said a Wolfe Research report.
In December, Databricks raised $10 billion in a funding round that valued it at $62 billion. It also recently added $5.25 billion in debt financing. So there's no pressing need for an IPO.
In addition to Snowflake, an emerging rival in business intelligence is Palantir. Palantir stock has surged.
Meanwhile, Snowflake stock has been pressured by views Databricks is gaining market share.
"The gap between Databricks and Snowflake is narrowing, with Databricks announcing a $3 billion revenue 2025 run rate versus Snowflake at $3.8 billion," said a Jefferies report.
Snowflake Rivalry Intensifies
In 2020, Snowflake pulled off the largest IPO by a software company. The Snowflake IPO in 2020 raised $3.4 billion. But Databricks might top that when it goes public.
Backed by venture capital firm Andreessen Horowitz, Databricks and Snowflake came to market from different starting points. But the two companies have converged.
Snowflake's data warehouse technology focused on transactional "structured" data from spreadsheets, financial records or customer information.
Databricks builds "data lake" technology. It's geared for structured data as well as "unstructured" data such as images, social media posts, emails, audio files and sensor data.
For Snowflake, the problem is that amid the emergence of generative AI the market is clearly shifting to data lakes from data warehouses.
"Many trends (data lakes, open formats, unstructured data) appear to favor Databricks," said a UBS report.
AI Competition Intensifies
Still, Snowflake seems to be waging a counterattack under a new chief executive. Snowflake stock has gained 20% in 2025.
Gabrisko pulled no punches in talking about the rivalry.
"I would say Snowflake has a good cloud data warehouse," he said. "They kind of missed the boat with data lakes with AI and machine learning. I would say we're at least five years ahead of them in terms of innovation. Most of our customers would consider snowflake legacy technology. We've migrated over 200 customers from Snowflake to Databricks including eight of their top 10 customers."
Meanwhile, Databricks acquired AI startup Mosaic for $1.3 billion in 2023. The purchase enabled Databricks to build its own LLM called DBRX.
Generative AI relies on LLMs, which require massive amounts of data to be trained. Large language models allow users to interact with AI systems without the need to write algorithms.
In the enterprise market, organizations have been developing small language models, or SLMs, using company data. The rise of free open-source models along with improved data science is making small language models more economical.
Also, Databricks has shelved plans to build more LLMs and is working closely with Facebook-parent Meta Platforms, which is a leading developer of open-source models. Meanwhile, Palantir is working with Anthropic and Elon Musk's xAI.
Databricks Acquisition Spree, Financials
Aside from Mosiac, well-funded Databricks has made other acquisitions. They included data management startup Tabular for reportedly $1 billion in June 2024.
Meanwhile, other deals included Arcion, Einblick, Lilac, Rubicon, Okera, and bit.io.
With the deals and internal hiring, Databricks has grown to more than 9,700 employees. Ahead of a possible IPO, it generates positive free cash flow, a key financial metric.
What's more, Databricks has over 12,000 customers, including 500 that generate over $1 million in annual recurring revenue. While Databricks works with all the big cloud platforms, Microsoft is a key partner.
Further, Gabrisko says it's early days in enterprise AI adoption.
"We have 12,000 customers. But we're only touching a smallish percentage of their overall data estate," Gabrisko said.
He added: "There's lots of data still on premise or in other systems. So I think the market's massive, and the amount of data increases every single day. Databricks is the innovation leader in data intelligence and we're growing faster than any other software company. And again, I think a big reason is the (market's) future is a lake house and data intelligence with AI — making forward looking predictions on any data specific to your enterprise."
Follow Reinhardt Krause on Twitter @reinhardtk_tech for updates on artificial intelligence, cybersecurity and cloud computing.
