On Monday, Databricks announced it reached a $5.4 billion revenue run rate, growing 65% year-over-year. More than $1.4 billion of that total was generated by its AI products. Co-founder and CEO Ali Ghodsi shared these growth numbers to counter widespread talk that AI will undermine SaaS businesses. He explained that for Databricks, AI is actually increasing product usage.
Ghodsi also aims to distance the company from the SaaS label, as private markets value it as an AI company. This announcement coincided with Databricks officially closing its previously announced massive $5 billion raise at a $134 billion valuation, along with securing a $2 billion loan facility.
The company operates in two spheres. While still best known as a cloud data warehouse provider for storing and analyzing enterprise data, it is seeing significant growth from AI. Ghodsi highlighted one AI product in particular, an LLM user interface named Genie, as a key driver of data warehouse usage.
Genie exemplifies how a SaaS business can replace a traditional user interface with natural language. For instance, Ghodsi uses it to ask why warehouse usage and revenue spike on certain days. Previously, such a request required specialized query languages or custom reports. Now, any product with an LLM interface can be used by anyone, a shift Ghodsi credits for boosting usage.
The real threat AI poses to SaaS is not that companies will rip out their core “systems of record,” like Salesforce or SAP, which store critical business data. Ghodsi notes that moving these systems is difficult, and model makers are not offering databases to replace them. Instead, AI aims to replace the user interface with natural language for humans, or APIs for automated agents.
According to Ghodsi, the danger for SaaS businesses is that people will no longer spend years mastering specific product interfaces. Once the interface becomes simple language, the products become invisible, like plumbing. This erodes the major moat these businesses built through user training.
SaaS companies that adopt new LLM interfaces can grow, as Databricks is demonstrating. However, this shift also creates opportunities for AI-native competitors to offer alternatives better suited for AI and agents. This is why Databricks created its Lakebase database, designed specifically for agents. Ghodsi reports early traction, noting that in its first eight months on the market, Lakebase generated twice the revenue the data warehouse did at the same stage.
With its recent funding round closed, Ghodsi stated the company is not immediately planning another raise or preparing for an IPO. He believes now is not an ideal time to go public. The substantial funding provides security, giving Databricks many years of runway and protection should market conditions decline as they did after the 2022 post-ZIRP crash.