Databricks Soars to $188 Billion Valuation as AI Momentum Builds
Four funding rounds in eighteen months have taken Databricks from big-data stalwart to AI's most aggressively valued enterprise platform.
Key points
- Databricks announced a $188 billion valuation in a new funding round led by Coatue.
- The company raised $5 billion in February 2026 at a $134 billion valuation, five months before this announcement.
- Databricks champions Z.ai's GLM 5.2 as a cost-efficient open-weight model for coding tasks.
- Internal benchmarking across 3,000 software engineers found the open-source harness Pi among the lowest-cost options without quality loss.
Databricks has announced a new funding round, led by Coatue, valuing the company at $188 billion. The exact amount hasn't been disclosed yet. The round hasn't closed; Databricks expects that to happen later this summer, and other outlets have put the figure at roughly $3 billion.
It's an unusual move to announce before the money arrives, but a venture capitalist told TechCrunch the deal is secure, with so much investor demand that there was little reason to stay quiet.
How did Databricks get here so fast?
Four rounds in roughly eighteen months tell the story. Founded in 2013, Databricks built its early reputation helping enterprises store and query large volumes of cloud data. Then came the AI wave. In December 2024 it raised $10 billion at a $62 billion valuation. By September 2025 that had climbed to $100 billion on a $1 billion raise. February 2026 brought a $5 billion Series L at $134 billion. Now $188 billion.
This is our sixth Databricks story in the last 90 days at AI2Day, and the trajectory has been hard to ignore.
The acceleration isn't just narrative. Databricks already held vast stores of enterprise data, which put it in a strong position when corporate clients began demanding AI with the governance and security they expect from conventional software. The company has shipped products to match: Lakebase, a database built for AI agents; Unity, an AI gateway; and Omnigent, a meta-harness that coordinates multiple agents simultaneously.
Should you care which AI models Databricks uses?
Yes, because its model choices are shaping how enterprises think about AI costs. Databricks has become a prominent adopter of Chinese-based open-weight models, specifically Z.ai's GLM 5.2 for coding work. Open-weight means the underlying model weights are published so anyone can run or adapt them.
Last week CEO Ali Ghodsi published benchmarking results from tests run across his 3,000 software engineers. GLM 5.2 handled the most difficult coding tasks at lower total cost than proprietary models from Anthropic and OpenAI. The bigger surprise was what else the data showed: the choice of harness, meaning the agentic coding tool that wraps around a model and manages its context window and instructions, shifted costs just as much as the model itself. The open-source harness Pi came out among the best on cost without sacrificing output quality.
"The lesson here isn't that one harness is always cheaper or that native harnesses are worse. Instead, model choice is only one piece of the puzzle," Ghodsi's post concluded.
The honest read on this valuation: Databricks earns its AI credentials through shipping real products and publishing real benchmarks. Watch whether the round closes at the rumoured $3 billion figure and whether GLM 5.2 adoption spreads beyond Databricks' own engineers to its enterprise clients.
