Databricks on Thursday introduced a brand new spherical of funding that values the corporate at $188 billion. The spherical was led by Coatue.
Databricks didn’t disclose precisely how a lot it raised; it mentioned the cash isn’t in its fingers but and that the spherical will shut later on this summer time. (Different shops have since reported the elevate is roughly $3 billion.) Whereas it’s uncommon for a corporation to announce earlier than it will get the cash, a VC tells TechCrunch that the deal is strong, with so many corporations wanting in that the corporate had no motive to maintain its shiny new valuation a secret.
Actually, Databricks has been on a year-and-a-half fundraising tear because it efficiently transitioned its picture into an AI supplier and never only a yesteryear SaaS sensation. Yesteryear being again within the BC occasions (Earlier than ChatGPT).
Solely 5 months in the past, in February, Databricks closed a $5 billion Collection L elevate at a $134 billion valuation. 5 months earlier than that, in September 2025, it raised $1B at $100 billion valuation. And roughly 9 months earlier than that, in December 2024, it raised what was a record-breaking spherical on the time of $10 billion at a $62 billion valuation.
Databricks has raised so many rounds over time that this newest one grew to become the topic of memes about working out of letters of the alphabet. “Turning on alerts for after we get a Collection AA,” one particular person posted.
However its picture reconstruction has been legit. Based in 2013, it initially grew to success again within the massive knowledge period, with software program that enabled enterprises to retailer huge quantities of information within the cloud, but produce speedy analytics.
As a result of it already sat on troves of enterprise knowledge, Databricks was then well-positioned to reply as firms began wanting AI with the identical safety and governance they anticipate from conventional enterprise software program.
The corporate started rolling out one AI product after one other, like Lakebase, its database constructed for AI brokers, and Unity, its AI gateway, together with a “meta-harness” referred to as Omnigent that manages a number of brokers.
Databricks additionally more and more grew to become recognized as one of many massive examples of enterprises adopting extra reasonably priced Chinese language-based open-weight fashions (fashions whose underlying code is revealed for anybody to make use of and modify) for value management, one of many massive traits of 2026. It’s a specific champion of Z.ai’s GLM 5.2 as a mannequin for coding.
Final week Databricks CEO Ali Ghodsi shared the outcomes of some inner benchmarking executed to handle his personal AI prices for his 3,000 software program engineers.
The corporate in contrast AI fashions on the precise duties its programmers do. Not surprisingly, within the weblog put up revealing the outcomes, Databricks shared that “open fashions, and GLM 5.2 particularly, are actually in a position to deal with even the best degree of activity issue” in coding, and at a complete decrease value than proprietary fashions from Anthropic and OpenAI.
Nevertheless it did shock folks by discovering that the selection of harness — the agentic coding device, like Codex or Claude Code, that wraps round a mannequin and manages its context and directions — equally impacted prices. It discovered that open-source harness Pi to be among the finest at managing context surrounding every immediate, and subsequently one of many lowest prices selections with out sacrificing high quality.
“The lesson right here isn’t that one harness is at all times cheaper or that native harnesses are worse,” the put up declared. “As an alternative, mannequin alternative is just one piece of the puzzle.”
All of this has added to Databricks picture as an AI firm, even when it wasn’t based as an AI lab. This, in flip, has granted it the AI-halo for elevating cash and leaping its valuation. As we beforehand reported, the AI impact is so sturdy lately, that even sandwich store Jersey Mike’s talked about AI 22 occasions in its S-1 paperwork.
Whenever you buy by way of hyperlinks in our articles, we could earn a small fee. This doesn’t have an effect on our editorial independence.