Databricks hits $188B valuation, extending its run as AI’s favorite second act

Databricks announced a new round of financing on Thursday valuing the company $188 billion. The lap was led by Coatue.

Databricks has not disclosed exactly how much it raised; it said the money is not yet in hand and the round will close later this summer. (Other outlets have since reported the increase is roughly $3 billion.) While it’s unusual for a company to announce something before it gets the money, one VC tells TechCrunch the deal is solid, and with so many companies participating that the company had no reason to keep its shiny new valuation a secret.

In fact, Databricks has already been raising money for a year and a half as it successfully transformed its image into an AI provider and not just a SaaS sensation of yesteryear. I used to be back in the BC times (before ChatGPT).

Just five months ago, in February, Databricks closed a $5 billion Series L raise at a $134 billion valuation. Five months earlier, in September 2025, it raised $1 billion at a valuation of $100 billion. And about nine months earlier, in December 2024, it raised what was then a record round of $10 billion at a valuation of $62 billion.

Databricks has organized so many rounds over the years that this last one became the topic memes about running out of letters of the alphabet. “Enable alerts for when we get a Series AA,” one person wrote.

But its image reconstruction has been legitimate. Founded in 2013, it initially rose to success in the big data era, with software that allowed enterprises to store vast amounts of data in the cloud while still producing rapid analytics.

Because it already had large amounts of enterprise data, Databricks was then well positioned to respond when companies wanted AI with the same security and management they expect from traditional enterprise software.

The company began rolling out one AI product after another, such as Lakebase, the database built for AI agents, and Unity, the AI ​​gateway, along with a “meta-harness” called Omnigent that manages multiple agents.

Databricks also increasingly became known as one of the big examples of companies adopting more affordable China-based open-weight models (models whose underlying code can be published and modified by anyone) for cost containment, one of the big trends of 2026. It is a particular proponent of Z.ai’s GLM 5.2 as a model for encryption.

Last week Databricks CEO Ali Ghodsi shared the results of a number of internal benchmarks conducted to manage its own AI costs for its 3,000 software engineers.

The company compared AI models to the actual tasks their programmers perform. Not surprisingly, in the blog post revealing the resultsDatabricks shared that “open models, and GLM 5.2 in particular, can now handle even the highest levels of task difficulty” in coding, and at an overall lower cost than proprietary models from Anthropic and OpenAI.

But it did surprise people by discovering that the choice of harness – the agentic coding tool, such as Codex or Claude Code, that wraps a model and manages its context and instructions – equally affected costs. It turned out that open-source Pi is one of the best at managing the context around each prompt, and is therefore one of the cheapest choices without sacrificing quality.

“The lesson here is not that one harness is always cheaper or that native harnesses are worse,” says the message explained. “Instead, model choice is just one piece of the puzzle.”

All this has contributed to Databrick’s image as an AI company, even though it was not founded as an AI lab. This in turn has given it the AI ​​halo for raising money and increasing its valuation. As we previously reported, the AI ​​effect is so strong these days that even the Jersey Mike sandwich shop mentions AI 22 times in its S-1 filings.