Databricks just pulled off something that would have seemed absurd two years ago. The company announced a fresh funding round on Thursday that values it at $188 billion, led by Coatue. It hasn’t disclosed the exact amount — the cash won’t actually land until later this summer — but other outlets are pegging it around $3 billion. Announcing a round before the money’s in the bank isn’t standard practice, but when enough firms are clamoring to get in, you don’t sit on a number like that.
What’s remarkable is the pace. Databricks closed a $5 billion Series L in February at a $134 billion valuation. Five months before that, it raised $1 billion at $100 billion. Go back to December 2024 and it was a $10 billion raise at $62 billion. That’s four rounds in about a year and a half, a cadence so relentless it’s spawned jokes about what happens when you run out of alphabet letters.
The company that started in 2013 selling big data analytics software has successfully repositioned itself as an AI provider. It helps that Databricks already sat on massive enterprise data stores when companies started demanding AI with real governance. The product rollout has been steady: Lakebase for AI agents, Unity as an AI gateway, and Omnigent, a meta-harness that wrangles multiple agents at once.
The company has also become a loud advocate for open-weight models, particularly Z.ai’s GLM 5.2 for coding. CEO Ali Ghodsi recently published internal benchmarks from his 3,000 engineers showing that model choice is only half the cost equation. The harness — the tool wrapping around the model — matters just as much. Open-source harness Pi emerged as a standout for keeping context costs low. The finding caught people off guard and reinforced Databricks’ credibility in a market where even sandwich chains are stuffing AI mentions into IPO filings.
