Ali Ghodsi, co-founder and CEO of Databricks, speaks on the firm’s Knowledge and AI Summit in San Francisco on June 16, 2026.
Grant Terzakis Media | Databricks
Databricks has a novel function within the AI increase. Income continues to soar as companies swarm to the corporate’s information analytics instruments. However as purchasers depend on extra AI brokers to scrub up information and ask questions, Databricks is enduring greater prices, resulting in decrease margins.
“It is the consumption-based enterprise mannequin, agentic AI coming,” Databricks CEO Ali Ghodsi informed CNBC in an interview on Tuesday on the firm’s Knowledge and AI Summit in San Francisco. “The brokers are producing far more queries. We’ve all these brokers, the brokers and agent platform we’ve got additionally generates income, so it simply will increase the consumption of the whole lot throughout.”
Databricks informed analysts on the convention that annualized income jumped over 80% from a yr earlier and now sits at $6.9 billion, up from a determine of $5.4 billion within the fiscal fourth quarter.
With a personal market valuation of $134 billion, Databricks is price greater than rival Snowflake, which went public in 2020 and now has a market cap of about $83 billion. Snowflake’s annualized income is round $5.6 billion, primarily based on its newest quarterly outcomes printed final month.
Databricks continues to take a seat on the sidelines of the general public market whilst its fellow extremely valued friends line up for IPOs. SpaceX held the most important debut on document final week, topping a $2 trillion market cap in its first day of buying and selling on Friday. And AI mannequin builders OpenAI and Anthropic have filed for confidential choices.
Whereas Databricks is commonly grouped in with the AI mannequin corporations, it has a really completely different function available in the market. Databricks’ Genie can reply questions from enterprise customers about company information, and its Agent Bricks permits builders to construct customized AI apps. As these merchandise get extra in style, Databricks should spend extra on underlying fashions.
Ghodsi declined to offer Databricks’ present gross margin, however he mentioned it would go decrease. His firm now will get $1.7 billion in annual income from AI merchandise, up from $1.4 billion in February.
One of many huge traits in AI in the present day is that corporations are cracking down on runaway spending from token utilization. Databricks’s Unity AI Gateway can notify folks as they get near utilizing up their AI budgets.
Firms have stopped tokenmaxxing, or encouraging staff to make use of as many tokens as potential, and at the moment are “value-maxxing” to optimize effectivity whereas nonetheless drawing on AI’s powers, Ghodsi mentioned.
Giant corporations “need to have the ability to completely use the frontier, smartest fashions,” he mentioned, highlighting Anthropic’s Mythos. “They’re focused on that, however not for the whole lot, proper? And for the mundane duties, they completely wish to curb the fee and use easy open-source fashions.”
Chinese language fashions are extraordinarily in style amongst Databricks prospects, Ghodsi mentioned.
“The purchasers are actually demanding the selection,” he mentioned.
Databricks is on the lookout for progress by promoting instruments refined for particular industries. The corporate introduced its entry into the cybersecurity market in March with the introduction of the Lakewatch software program. On Tuesday, Databricks mentioned it might purchase Panther, a safety startup valued at $1.4 billion in 2021, and it unveiled CustomerLake software program for managing advertising and marketing information.
WATCH: Databricks CEO Ali Ghodsi: AI would not have an intelligence drawback. It has a context drawback
