What’s New in Microsoft Fabric

General Availability (GA) Highlights

1. Semantic Link

A shared semantic layer is now generally available across experiences. This lets BI developers, data engineers, and data scientists use the same semantic models in notebooks, Spark, SQL, and Power BI without separate model copies. That means fewer reworks between tools and consistent business logic across workloads.

2. Snowflake Key-Pair Authentication

The Snowflake connector now supports key-pair authentication across Power BI, pipelines, Dataflow Gen2, and mirroring. This enables secure, password-less connections to Snowflake and improves long-term automation credentials.

3. Default Domain Sensitivity Labels

New default sensitivity labels can be applied automatically to items based on their domain, reducing manual classification and helping keep data consistently protected from day one.

Data Engineering and Connectivity – Preview

4. Zero-copy Access to OneLake Data in Azure Databricks

Fabric now supports zero-copy access for Azure Databricks via OneLake catalog federation. That means Databricks can query the same OneLake data without copying it, preserving OneLake as the source of truth.

5. Microsoft ODBC Driver for Fabric Data Engineering

A new ODBC driver allows external tools (Python, .NET, and BI tools) to connect to Fabric’s Spark SQL engine with Azure AD authentication, session reuse, and async prefetch support.

6. Recent Data in Dataflow Gen2

Dataflow Gen2’s Get Data experience now surfaces recently used tables, folders, and databases so you can jump straight to what matters most.

Real-Time Intelligence and Event Processing – General Availability

7. Data Series Colors in Real-Time Dashboards

Users now have full control over data series colors (pie, line, bar, scatter, etc.) in real-time dashboards, improving consistency and business storytelling in operational monitoring views.

8. Processing CDC Streams Using Eventstreams SQL

Fabric Eventstreams SQL can now transform CDC events (e.g., from Azure SQL, PostgreSQL, MySQL) into business-ready streams using familiar SQL logic, centralising real-time change shaping upstream.

SQL and Analytics

9. SQL Database Collation Support

You can now define collation when creating SQL databases using CLI or REST API. This is important for regional sorting rules and compatibility.

DevOps and Platform

10. Official Support for fabric-cicd Tool

The fabric-cicd Python library is now officially supported, enabling structured CI/CD workflows across Dev, Test, and Prod with dependency management and parameterisation.