Databricks has raised more than $4 billion in a Series L funding round that values the company at $134 billion, underscoring how aggressively investors are backing the “picks-and-shovels” layer of enterprise AI: governed data, analytics, and platforms for deploying AI applications inside large organizations. [1]
The new round lands as enterprises move past early-gen AI experimentation and start investing in production systems—especially AI agents and “data-intelligent applications” that can operate securely on proprietary corporate data. Databricks says it crossed a $4.8 billion revenue run rate in Q3 2025, up more than 55% year over year, while also delivering positive free cash flow over the last 12 months. [2]
What happened on December 17, 2025: the Databricks mega-round, in context
While Databricks announced the financing on Tuesday, December 16, the story dominated tech and finance coverage into Wednesday, December 17, 2025, because the deal is one of the largest late-stage fundraises in the AI infrastructure category—and because the valuation jump is striking so soon after its prior raise. [3]
Key facts from today’s coverage:
- Amount & valuation: More than $4B raised at a $134B valuation. [4]
- Leads:Insight Partners, Fidelity Management & Research Company, and J.P. Morgan Asset Management led the round. [5]
- Use of proceeds: R&D, go-to-market expansion, and talent retention (including employee liquidity via secondary share sales), plus continued AI investment. [6]
- Business performance:$4.8B revenue run rate in Q3 2025, >55% YoY growth, and positive free cash flow (last 12 months). [7]
Who invested—and why this round matters beyond the headline valuation
Databricks’ Series L wasn’t a quiet extension; it brought together a roster that signals broad institutional conviction in enterprise AI infrastructure.
Alongside the lead investors, Databricks listed additional participation from major names including Andreessen Horowitz, BlackRock, Blackstone, Coatue, GIC, MGX, NEA, Ontario Teachers’ Pension Plan, Robinhood Ventures, T. Rowe Price, Temasek, Thrive Capital, and Winslow Capital, among others. [8]
The strategic logic is straightforward: the next wave of AI value in big companies is expected to come less from commodity foundation models and more from the platforms that can securely connect models to enterprise data, enforce governance, and operationalize AI workflows across departments.
Reuters captured that competitive urgency directly, quoting CEO Ali Ghodsi describing the market as an investment arms race: “It’s a race, and everybody’s investing.” [9]
The numbers: Databricks’ growth engine (and the metrics investors keep circling)
Databricks’ funding story is tied to traction signals that look more like public-company KPIs than a typical private startup narrative.
From the company’s disclosures and reporting:
- $4.8B revenue run rate in Q3 2025, up more than 55% from a year earlier. [10]
- AI products exceeded a $1B revenue run rate, and data warehousing also exceeded a $1B revenue run rate. [11]
- Net retention rate sustained at >140%. [12]
- 700+ customers consuming at over $1M annual revenue run rate (a strong indicator of large-scale deployments). [13]
Databricks also positions itself as deeply embedded in large enterprises: it says more than 20,000 organizations rely on Databricks, including “over 60% of the Fortune 500,” and it is headquartered in San Francisco with “30+ offices around the globe.” [14]
From “data platform” to “AI application factory”: what Databricks says it’s building next
A major reason this round drew attention on December 17 is that Databricks isn’t pitching “more dashboards.” It’s pitching an application layer for AI—the idea that enterprises will build thousands of internal AI-driven workflows (agents, copilots, automated analytics, monitoring systems) that sit on top of governed data.
Reuters reports Ghodsi outlined a strategy focused on building “data intelligence apps,” including investment in a database tailored for AI agents and Agent Bricks for embedding intelligence into software—while letting customers use multiple underlying model providers (OpenAI, Anthropic, Google, and open-source options). [15]
The Databricks press release frames this moment as the convergence of generative AI with new “coding paradigms,” positioning the company as the foundation for “Data Intelligent Applications.” [16]
The three product bets Databricks is emphasizing
Databricks says the Series L capital will accelerate investment across three strategic products:
- Lakebase — described as a serverless Postgres database “purpose-built for the age of AI.” [17]
- Databricks Apps — positioned as an application layer to build and deploy data and AI applications. [18]
- Agent Bricks — an agent-building and scaling layer aimed at helping organizations deploy “high-quality agents” on their data. [19]
Databricks’ message is that these pieces combine into a stack where Lakebase becomes the system of record, Databricks Apps the user experience layer, and Agent Bricks the engine for multi-agent systems. [20]
Why Lakebase keeps coming up: the Neon acquisition and “database for AI agents”
If you’re seeing “Lakebase” everywhere in today’s write-ups, it’s because Databricks is treating the database layer as the battleground for agentic AI.
TechCrunch notes that Lakebase is based on Postgres and was enabled by Databricks’ approximately $1 billion acquisition of Neon, a serverless Postgres startup. [21]
Databricks’ own announcement earlier this year described Neon as a “developer-first, serverless Postgres company” built with an architecture separating compute and storage—capabilities it argued are especially well-suited for developers and AI agents that need to spin up databases quickly. [22]
This matters because agents don’t behave like traditional enterprise apps. They generate and test, branch and retry, often creating spiky workloads that can punish rigid infrastructure. A serverless database model (in theory) makes those workloads more economical and easier to operate—especially as companies experiment with internal agent-based tools.
Partnerships with model providers: Databricks’ bet on “choice” as LLMs commoditize
Databricks is leaning hard into the idea that enterprises will want optionality in models—using different providers depending on cost, performance, latency, or governance needs.
Reuters reports Ghodsi sees “commoditization happening with LLMs,” which increases the value of Databricks’ platform for customizing and deploying models securely. [23]
In practice, that “choice” strategy has shown up as partnerships:
- Databricks announced a multi-year partnership to make OpenAI models available within the Databricks Data Intelligence Platform and its Agent Bricks product. [24]
- Databricks also announced a multi-year partnership with Anthropic to bring Claude models to the platform for customers building AI agents over enterprise data. [25]
TechCrunch reported Databricks has struck deals “worth hundreds of millions” with Anthropic and OpenAI to offer their models within enterprise products. [26]
Where the $4B+ is going: R&D, go-to-market, acquisitions—and employee liquidity
For a raise of this size, investors and employees both watch the “use of proceeds” language closely.
Reuters reports Databricks plans to use the funds for research and development, to expand its go-to-market teams, and for talent retention, including providing liquidity to employees through secondary share sales. [27]
The company’s release similarly says the capital is expected to support growth investments, potential AI acquisitions, deeper AI research, and employee liquidity. [28]
That employee-liquidity emphasis is important in today’s private markets: when companies stay private longer, they must create ways for long-tenured staff to realize value without waiting for an IPO.
IPO timing: Databricks signals flexibility, not a countdown clock
On a day when the tech world continues to debate whether the IPO market is truly “back,” Databricks’ actions answer a different question: Why go public if you can raise like this privately?
TechCrunch framed it bluntly: IPOs used to be the mechanism to raise massive sums; now, companies like Databricks can raise “ungodly amounts” without stepping into public-market scrutiny. [29]
Reuters reports Databricks is not ruling out an IPO in 2026, but Ghodsi referenced the 2022 market downturn and layoffs as a scenario he hopes to avoid as a public company. [30]
Meanwhile, the Wall Street Journal reported Databricks intends to use the money to support product development and expansion—including hiring—while still not committing to an immediate IPO timeline. [31]
How Databricks got here: from a $62B valuation to $134B in one year
Today’s $134 billion valuation looks even more dramatic when set against Databricks’ recent funding timeline.
- In December 2024, Reuters reported Databricks raised $10 billion in an oversubscribed round that valued it at $62 billion. [32]
- In September 2025, Reuters reported Databricks closed a $1 billion round valuing it at $100 billion, projecting $4 billion in annualized revenue amid surging AI demand. [33]
- Now, Databricks says it has raised more than $4 billion at a $134 billion valuation, with a $4.8B run rate and >55% YoY growth. [34]
This sequence helps explain why Databricks is being discussed as one of the most consequential private enterprise software companies of the AI era: the capital inflows are tracking an expectation that data/AI platforms become long-term “operating systems” inside large companies.
The bigger story behind today’s headlines: AI spend is moving from models to systems
One of the clearest takeaways from December 17’s coverage is that Databricks is riding a shift in enterprise AI priorities:
- Early AI excitement centered on model capability.
- The next phase centers on deployment: clean data, governance, retrieval, monitoring, tool use, and agent orchestration inside secure corporate environments.
Databricks is explicitly positioning itself as a “neutral and secure option” where sensitive data can be used without leaving the customer’s cloud environment, according to Reuters. [35]
And while Databricks’ raise is the headline, it’s also part of a broader capital pattern on December 17: investors are funding adjacent categories like data security, as AI adoption increases both the value and risk profile of enterprise data. [36]
What to watch next
With fresh capital and a valuation that puts it in rare company, Databricks’ next milestones will likely shape how enterprises operationalize AI in 2026:
- Product adoption signals for Lakebase, Agent Bricks, and Databricks Apps (especially production deployments, not pilots). [37]
- Expansion of go-to-market capacity as Databricks pushes deeper into global enterprise accounts and regulated industries. [38]
- More partnerships and acquisitions aimed at rounding out the “AI application factory” stack. [39]
- IPO posture: whether “not ruling out 2026” evolves into a clearer timetable—or whether the private markets remain too attractive to leave. [40]
For now, Databricks’ Series L round sends a clear signal across the market: the enterprise AI race is increasingly being won not by whoever has the flashiest model demo, but by whoever can turn corporate data into secure, governed, scalable AI systems—and do it fast. [41]
References
1. www.reuters.com, 2. www.reuters.com, 3. www.reuters.com, 4. www.reuters.com, 5. www.reuters.com, 6. www.reuters.com, 7. www.reuters.com, 8. www.prnewswire.com, 9. www.reuters.com, 10. www.reuters.com, 11. www.reuters.com, 12. www.prnewswire.com, 13. www.prnewswire.com, 14. www.prnewswire.com, 15. www.reuters.com, 16. www.prnewswire.com, 17. www.prnewswire.com, 18. www.prnewswire.com, 19. www.prnewswire.com, 20. www.prnewswire.com, 21. techcrunch.com, 22. www.databricks.com, 23. www.reuters.com, 24. www.databricks.com, 25. www.databricks.com, 26. techcrunch.com, 27. www.reuters.com, 28. www.prnewswire.com, 29. techcrunch.com, 30. www.reuters.com, 31. www.wsj.com, 32. www.reuters.com, 33. www.reuters.com, 34. www.reuters.com, 35. www.reuters.com, 36. www.reuters.com, 37. www.prnewswire.com, 38. www.reuters.com, 39. www.prnewswire.com, 40. www.reuters.com, 41. www.reuters.com
