Databricks Acquires Tabular for $2B to Unify Its Data Lakehouse
Databricks has acquired Tabular, the company behind Apache Iceberg's commercial platform, for approximately $2 billion, aiming to build a unified open data and AI lakehouse stack that directly challenges Snowflake.
Original sourceDatabricks has completed its acquisition of Tabular, the startup founded by the original creators of Apache Iceberg, for roughly $2 billion. Tabular built the commercial platform around Iceberg, the open table format that has become the de facto standard for large-scale analytical workloads — and a direct competitor to Databricks' own Delta Lake format. The deal consolidates two of the most important open table format ecosystems under one roof.
The strategic logic is clear: Databricks gets both the team that invented Iceberg and the commercial customer base Tabular built around it. Rather than fighting a format war, Databricks is absorbing the opposition and positioning itself as format-agnostic infrastructure for enterprise data and AI workloads. The stated goal is a unified platform where data lakehouse storage, governance, and AI tooling coexist without forcing customers to bet on a single table format.
For Snowflake, this acquisition tightens the competitive pressure considerably. Snowflake has leaned heavily into Iceberg support as a way to attract customers who want openness without committing to Databricks' Delta Lake. With Tabular now inside Databricks, the 'Iceberg-neutral' positioning Snowflake relied on becomes harder to sustain. The deal also signals that Databricks is willing to pay acquisition prices that reflect the strategic value of open-source stewardship, not just revenue multiples.
Tabular's co-founders — Ryan Blue, Daniel Weeks, and Jason Reid — built Iceberg while at Netflix and later formalized it as an Apache project. Their return to the ecosystem under Databricks' umbrella will shape how Iceberg evolves going forward, raising legitimate questions about whether a project stewarded by a commercial acquirer can maintain its community-first governance.
Panel Takes
The Builder
Developer Perspective
“The primitive here is table format ownership: Databricks just bought the team that defined how Iceberg's metadata layer actually works, which means they now control the reference implementation and the governance roadmap. That's not a marketing move — that's acquiring the people who can merge or reject your PR. The DX bet is format neutrality as a first-class API contract, which is the right call if they execute, but the moment Databricks subtly favors Delta Lake in query planner optimizations, the open ecosystem trust evaporates overnight.”
The Skeptic
Reality Check
“This is a $2 billion solution to a format war Databricks was losing on vibes — Iceberg had the mindshare, Delta Lake had the home-field advantage, and Tabular was winning enterprise conversations by being the 'not Databricks' Iceberg option. Now that option is gone, which helps Databricks but also hands Snowflake a clean narrative: 'the open format is now owned by your other vendor.' The thing that kills this deal's thesis in 12 months isn't integration failure — it's Apache governance drama the moment Databricks tries to steer an Iceberg spec change that benefits their commercial product.”
The Futurist
Big Picture
“The thesis Databricks is betting on: by 2027, enterprise AI infrastructure converges on open table formats as the durable storage layer, and whoever controls the format spec controls the data gravity. That's a falsifiable claim with a specific dependency — it requires Iceberg adoption to keep accelerating faster than proprietary formats can re-entrench. The second-order effect nobody is talking about is what this does to the catalog and governance layer: if Databricks owns the format and Unity Catalog, every metadata API, access policy, and lineage graph routes through them, which is a leverage point that makes the $2B look cheap.”
The Founder
Business & Market
“The buyer here is the enterprise data platform budget — the same CFO line item that Snowflake has been defending for five years, and Databricks has been attacking. At $2 billion for a company with meaningful ARR but not at traditional SaaS multiples, Databricks paid a strategic premium, not a revenue multiple, which tells you exactly how existential the format war was for their long-term margin story. The moat they're building is real: owning the spec, the commercial platform, and the AI runtime on top of it creates switching costs that go three layers deep — format, governance, and compute — which is the kind of lock-in that survives even a 10x drop in underlying model costs.”