According to Fortune, Databricks CEO Ali Ghodsi is laying out a three-part plan to grow his data intelligence startup into a trillion-dollar company. The firm, which is reportedly in talks for funding at a $134 billion valuation, is targeting the transactional database market with its Lakehouse product. Ghodsi revealed that over 80% of new databases on its platform are now launched by AI agents, not humans, a trend he calls a “huge growth factor.” The other key areas are its Agentbricks platform for building enterprise AI agents and fostering “vibe coded” applications on top of its infrastructure. Ghodsi didn’t give a timeframe but suggested an IPO, possibly next year, is part of the journey, though he called going public “not a really big deal.”
The three-legged trillion-dollar stool
So, Ghodsi’s “trifecta” is basically an attempt to own the entire modern data stack. First, going after Oracle’s old-school database turf with Lakehouse is a classic land grab. It’s a bet that companies are tired of legacy systems. Second, the AI-agent-driven growth is fascinating. If AI is auto-generating code and needs a place to put data, Databricks wants to be the default parking garage. That’s a powerful, almost passive, growth channel if it holds. And third, Agentbricks is about moving up the value chain from storing data to actively doing things with it. It’s not just a database anymore; it’s trying to be the brain of the corporation. The real trick is getting all three pieces to work together seamlessly.
The IPO and the math
Here’s the thing: a $1 trillion valuation is a monster leap from $134 billion. We’re talking about a roughly 7x increase. For context, that’s like growing from today’s Salesforce to beyond today’s Nvidia. It’s an audacious goal. Ghodsi’s casual attitude about the IPO is interesting—”not a really big deal.” It feels like he’s saying the company’s fate isn’t tied to the public markets, but let’s be real, an IPO is the liquidity event that would cement that valuation and provide the war chest for this kind of expansion. Saying “maybe” to a 2025 IPO keeps all options open while they wait for the perfect window.
The giant problem
But can they actually pull it off? Look, the ambition is clear, and Databricks is a formidable player. However, every single one of those three growth vectors has massive, entrenched competition. The database market? Oracle, Google, AWS, Microsoft. AI agents? Every big tech cloud provider is building their own stack. Even the “vibe coding” app layer is crowded. Their bet is that their unified approach is the winner. And maybe it is. But a trillion dollars? That puts them in the league of Apple and Microsoft. That requires not just execution, but essentially defining a new category and then dominating it completely. I think the path is clearer to a few hundred billion. A full trillion? That’s the realm of global tech ecosystems. It’s a hell of a goal, and you have to admire the confidence, but the road is ludicrously steep.
