Microsoft Fabric and Azure Databricks Announce Bi-Directional OneLake Interoperability

Summary

Microsoft and Databricks have announced a major integration delivering true bi-directional interoperability between Microsoft Fabric and Azure Databricks through OneLake. Organizations can now seamlessly read and write data in Delta Lake and Parquet formats across both platforms without data movement, duplication, or complex ETL pipelines. This establishes OneLake as a unified data foundation for multi-platform data estates.

What happened?

Microsoft and Databricks are bridging their respective lakehouse architectures with native integration features:

• OneLake Catalog Federation: Azure Databricks compute engines (SQL and notebooks) can now discover and query data stored in Microsoft Fabric’s OneLake directly as a foreign catalog using Unity Catalog federation.
• Unity Catalog Mirroring: In return, Azure Databricks Unity Catalog tables can be mirrored natively into Microsoft Fabric, making them queryable in Fabric workloads (such as Power BI, Data Factory, and Synapse Data Engineering) without copying underlying data.
• Native OneLake Storage: Azure Databricks can now store Unity Catalog managed tables directly in OneLake, treating OneLake as a primary storage layer.
• Direct Path-Based Access: Developers can access OneLake paths directly using the Spark ABFS (Azure Blob Filesystem) driver or the MSAL (Microsoft Authentication Library) in serverless environments.

Why it matters

Previously, enterprises using both Azure Databricks (for complex data engineering and ML) and Microsoft Fabric (for reporting and SaaS integrations) had to build and maintain custom pipelines to copy data back and forth. This resulted in:

• High operational overhead for data synchronization and governance.
• Increased storage costs and potential egress fees.
• Fragmented reporting due to stale copies. This bi-directional interoperability eliminates these issues, allowing organizations to maintain a single copy of data while choosing the best engine for each task.

Evidence

• Microsoft Fabric Community Updates Blog: The announcement article “Unlocking Microsoft OneLake as the data foundation for Azure Databricks customers” details the bi-directional setup.
• Azure Databricks Release Notes (June 2026): Technical release notes describing OneLake catalog federation configuration.
• YouTube Tutorials: Product team demos highlighting live cross-platform queries.

Analysis

This collaboration highlights the growing importance of open data formats like Delta Lake. By integrating OneLake and Unity Catalog, Microsoft and Databricks are acknowledging that most enterprise data estates are multi-platform. Rather than forcing clients into a single silo, this native interoperability allows them to protect their existing Databricks engineering pipelines while leveraging Fabric’s SaaS capabilities, particularly Power BI’s Direct Lake mode.

Practical Takeaways

1. Eliminate Power BI ETL: Use Unity Catalog mirroring to expose Databricks tables directly to Fabric, enabling real-time Power BI reports without complex data refresh pipelines.
2. Target OneLake Directly: For new data pipelines in Databricks, evaluate storing tables directly in the OneLake ABFS path (abfss://<workspace>@onelake.dfs.fabric.microsoft.com/...).
3. Align Security Policies: Consolidate access policies across both platforms as federated connections respect security rules defined in the source environments.

Open Questions

• Egress & Transaction Costs: What are the exact costs and potential internal data charges when querying large-scale datasets across the Fabric and Databricks boundary?
• Query Performance: What is the performance overhead of federated queries via Unity Catalog compared to native local storage access on high-concurrency clusters?
• Global Availability: When will these integration capabilities roll out across all global Azure regions?

Sources

1. Microsoft Fabric Community Blog: Unlocking Microsoft OneLake for Azure Databricks
2. YouTube: Microsoft Fabric + Databricks Better Together
3. Niracore: Microsoft Fabric vs. Databricks for Enterprises
4. Microsoft Learn: Azure Databricks Product Release Notes (June 2026)
5. Databricks Community: Why Alternatives Fall Short for Databricks