Databricks has surpassed $5.4 billion in annualized revenue growing at 65% year over year, is eyeing a valuation of $165 to $175 billion in its next funding round, and has watched its data warehousing product hit a revenue run rate that has more than doubled in a year, setting up what could be the defining enterprise software IPO of 2026.
The simplest way to understand what's happening at Databricks is to look at one number: $600 million. That was Databricks SQL's annualized revenue run rate a year ago, the company's data warehousing product that competes head-on with Snowflake. As of early 2026, according to Bloomberg, it had crossed $1 billion and was growing at more than 150% year over year. That's not catching up. That's a different trajectory entirely.
For context, Snowflake is a public company. Its fiscal year 2026 revenue landed at roughly $4.7 billion, growing at 29% annually. Databricks, still private, is at $5.4 billion in run rate and accelerating. The valuation gap already reflects this: Databricks is in talks for a round that would value it at $165 to $175 billion, nearly double Snowflake's market capitalization of around $83 billion. Four months ago, Databricks closed its Series L at $134 billion. The re-rate is happening fast.
Snowflake built its dominance on cloud data warehousing: fast SQL queries over structured data, priced by compute. It's a model that served enterprises well for years. But AI workloads don't fit neatly into that box. Training, inference, fine-tuning, vector search, and agentic systems want to sit close to raw data in varied formats, not sanitized tables loaded into a warehouse. That's the bet Databricks made years ago with its lakehouse architecture, and enterprises are now paying for it at scale. More than 60% of the Fortune 500 are Databricks customers, and over 800 of them are spending more than $1 million a year on the platform.
AI product revenue alone hit $1.4 billion annualized, which is a number most enterprise software companies would be proud to call their entire business. But for Databricks it's running alongside a data infrastructure business that is also growing at a pace Snowflake can't match. Gartner estimated that more than 50% of enterprises would adopt a lakehouse architecture as their analytics and AI foundation by 2026, up from less than 15% in 2022. That shift is now showing up in Databricks' revenue line.
The product roadmap is also moving beyond the lakehouse. In February 2026, Databricks launched Lakebase, a fully managed PostgreSQL database built for AI apps and agentic workloads, following its roughly $1 billion acquisition of Neon, a serverless Postgres startup. The pitch is operational and analytical data converging on a single platform: one place for the app database, the analytics warehouse, and the AI model serving layer. If that architecture gains traction with developers building AI agents, it opens a spending category that neither Snowflake nor the hyperscalers have cleanly captured yet.
The IPO math and what it means for enterprise AI valuations
Databricks hasn't filed an S-1 and CEO Ali Ghodsi has been measured about timing, with some reporting suggesting a filing in the second half of 2026 and others pointing to 2027. But the pre-IPO funding round currently in discussion is itself a signal. Companies don't raise at $175 billion unless they're either preparing to go public soon or they're comfortable with investors pricing in that outcome.
The stakes here go beyond Databricks itself. It is the most valuable private enterprise software company in the world. How it prices, how the market receives it, and whether public investors accept the AI infrastructure growth premium that private backers have already paid will set the tone for the backlog of late-stage AI companies watching from the sideline. Companies like Anthropic, xAI, and Scale AI are all sitting on private valuations built partly on the assumption that public markets will eventually absorb them at similar multiples. Databricks going out at $150 billion or higher validates that assumption. A stumble reprices the whole cohort.
Snowflake's current position makes the comparison unavoidable. Trading at roughly half of Databricks' private valuation while growing at less than half the pace, it's exactly the kind of incumbent contrast that IPO bankers build roadshow narratives around. Databricks doesn't need to say it directly. The numbers say it for them.
What actually matters to enterprises making infrastructure decisions right now is simpler than the valuation story. They're spending on AI and they need somewhere to put the data those systems run on. Databricks, with 20,000 customers, a warehousing product that more than doubled in a year, and a new operational database aimed at agentic workloads, is positioned as the answer to that question. Whether the IPO lands in 2026 or slips into 2027, the commercial momentum is already real.
Also read: SpaceX pays $60 billion in stock for Cursor, betting the AI coding toolchain is worth owning outright • Asia's chip stocks are beating Nvidia because the real AI bottleneck was never the GPU • France drops Palantir for a local rival and hands every civil servant a Mistral AI assistant
