Databricks announced a new funding round on Thursday that values the company at $188 billion. The round was led by Coatue.
Databricks did not disclose exactly how much was raised. No money is in hand yet and the round is expected to close later this summer, he said. (Other outlets have since reported the raise to be roughly $3 billion.) It’s unusual for a company to make an announcement before receiving any money, but one VC told TechCrunch that a deal is solid. Because too many companies want there to be no reason to keep their shiny new value a secret.
In fact, Databricks has been fundraising for a year and a half, successfully transforming its image as an AI provider as well as a former SaaS sensation. Yesterday we were back in the BC era (before ChatGPT).
Just five months ago, in February, Databricks closed its $5 billion Series L fundraising at a valuation of $134 billion. Five months earlier, in September 2025, it raised $1 billion at a valuation of $100 billion. And about nine months ago, in December 2024, it raised a then-record funding of $10 billion at a valuation of $62 billion.
Databricks has made so many rounds over the years that this latest round became the subject of a meme about its lack of alphabet letters. “Turn on notifications when I get Series AA,” one person posted.
However, the image reconstruction was legal. Founded in 2013, the company initially found success in the big data era with software that helped businesses generate fast analytics while storing massive amounts of data in the cloud.
Because Databricks already holds so much enterprise data, it was well-positioned to respond as enterprises began to want AI with the same security and governance they expect from traditional enterprise software.
The company began rolling out AI products one after another, such as Lakebase, a database built for AI agents, Unity, an AI gateway, and a “metaharness” called Omnigent that manages multiple agents.
Databricks has also become increasingly known as one of the big examples of companies adopting a cheaper, China-based open model (where the underlying code is published for anyone to use and modify) to manage costs, one of the big trends for 2026. This is a special champion of Z.ai’s GLM 5.2 as a coding model.
Last week, Databricks CEO Ali Ghodsi shared the results of internal benchmarking the company conducted to manage its own AI costs for its 3,000 software engineers.
The company compared its AI models against actual tasks performed by programmers. Naturally, in a blog post revealing the results, Databricks shared that “open models, especially GLM 5.2, can now handle even the highest levels of task difficulty in coding” and have a lower total cost than proprietary models from Anthropic and OpenAI.
But we were surprised to discover that the choice of harness, an agent coding tool like Codex or Claude Code that wraps the model and manages its context and instructions, has an equal impact on cost. We’ve found that the open source Harness Pi is one of the best at managing the context around each prompt and is therefore one of the lowest cost choices without sacrificing quality.
“The lesson here is not that one harness is always cheaper or that the stock harness is worse,” the post declared. “Instead, model selection is only one piece of the puzzle.”
All of this adds to Databricks’ image as an AI company, even though it was not founded as an AI lab. This in turn has given AI a halo beyond its funding and valuation. As we previously reported, the AI effect is so strong these days that even the sandwich shop Jersey Mike’s mentioned AI 22 times in its S-1 document.
If you purchase through links in our articles, we may receive a small commission. This does not affect our editorial independence.
