If you haven’t already, check out Arun Ulag’s hero blog “FabCon and SQLCon 2026: Unifying databases and Fabric on a single, complete platform” for a complete look at all of our FabCon and SQLCon announcements across both Fabric and our database offerings.
Overview
Following the SQL database in Microsoft Fabric (Generally Available) at Microsoft Ignite in November 2025, customer adoption has accelerated. Organizations are modernizing SQL workloads, reducing operational overhead, and bringing operational data closer to analytics and AI. This next wave of capabilities builds on the preexisting pillars—Simplified, Autonomous & Secure, and Optimized for AI—driven directly by customer feedback.
Modernize with confidence – Migration Assistant (Preview)
The new Migration Assistant for SQL databases simplify moving SQL Server and Azure SQL workloads into Fabric. Designed for SQL developers, it imports schema via DACPACs, identifies compatibility issues, and provides clear, actionable guidance before migration. Built-in assessment and data copy workflows help teams move from evaluation to cutover with less manual effort, preserving existing SQL skills while accelerating time to value on Fabric’s unified analytics platform. Ready to simplify your SQL migration journey? We will begin rolling this out in the coming weeks, and it will soon be accessible through the Fabric portal.
Figure: Migration options for SQL databases in Microsoft Fabric.
Figure: Migration assistant showing overall status and progress across migration steps.
Autonomous by default, configurable by design (Preview)
SQL database in Fabric remains SaaS-first, removing the burden of day-to-day database management while giving customers more control where it matters. You can now configure database-level compute limits to cap maximum vCore usage, helping manage scaling behavior and cost on shared Fabric capacities, with safe defaults that preserves existing behavior. This opt-in control is especially useful where predictability matters more than peak throughput.
Learn more: SQL database in Fabric compute limits.
Figure: Database settings page for configuring maximum vCore capacity limits.
Expanded support for database compatibility levels, additional T-SQL capabilities, full-text search, and expanded support for relevant ALTER DATABASE SET options make it easier to bring existing applications to SQL databases in Fabric without code changes and enabling continual parity with Azure SQL database.
Full collation support (Preview)
You can now use all Azure SQL database collations when creating databases. This enhancement gives SQL database in Fabric users greater flexibility and compatibility for global data scenarios, reporting, and app development, no matter your language or regional requirements.
Check out the YouTube demo and explore sample Git repo to learn more about creating a new SQL database in Microsoft Fabric with a specific collation using the Fabric REST API. The Fabric SQL Database collation does not affect the collation of the replicate data in the SQL analytics endpoint. For details on how to change the collation of the SQL analytics endpoint, explore this documentation, refer to: Data Warehouse collation - Microsoft Fabric.
Restore deleted databases
In Fabric, when a database is deleted, it goes into a soft-deleted state into the Fabric Workspace’s Recycle Bin Tab. Depending on the retention configured, the deleted database can be recovered from the Recycle bin while in retention. In addition to this Recycle bin experience, the Fabric SQL database also has the backup retention period configurable from 1-35 days. When the database is hard deleted from the Recycle bin, the backups are still available for the configured backup retention period. This new improvement allows you to restore the backup into a new database to any point in time within the restorable period.
Learn more: Automatic backups in SQL database - Microsoft Fabric | Microsoft Learn
Pre-deployment and post-deployment scripts
The SQL database source control integration and deployment pipelines in Fabric now support executing additional T-SQL content as part of the database definition. Enhance your database branching and provisioning with static data management or other customization by adding a Shared Query to the database and designating it as a pre-deployment or post-deployment script. The SQL projects database definition continues to be compatible with the full ecosystem of database CI/CD tooling, ensuring that development of SQL database in Fabric is both streamlined and flexible.
Learn more information on source control and deployment pipelines for SQL database in Fabric.
Manage built-in Mirroring (Preview)
Built‑in database mirroring to OneLake makes operational data immediately available for analytics and AI, with zero ETL. You asked for more control over which tables are mirrored and we delivered. You can now selectively manage the tables mirrored into Fabric OneLake, explore feature documentation. We also have the capability thru APIs to Start & Restart the Mirroring capability using the REST API.
Set up and use workspace-level private links.
Enterprise readiness
Enterprise features continue to expand with support for over 5,000 Azure SQL Database collations, SQL auditing, customer-managed keys (CMK), and expanded availability zone support.
Customer‑Managed Keys (CMK) (Generally Available)
Fabric already secures all data at rest with Microsoft Managed keys. For organizations with strict
governance or regulatory needs, CMK adds an extra layer of control. With CMK, you can encrypt SQL database in Fabric data using your own keys stored in Azure Key Vault, giving:
- Full ownership and rotation of encryption keys.
- Granular access control.
- End to end auditability of key usage.
- Support for industry specific compliance requirements.
Check out the How to encrypt Fabric SQL Database with Customer Managed Keys video and learn more in the data encryption in SQL database in Fabric documentation.
Auditing (Generally Available)
This critical security and compliance capability enables organizations to track and log database
activities, providing clear visibility into who accessed what data, when, and how. By maintaining an
immutable audit trail, auditing helps customers meet compliance requirements, investigate suspicious
activity, and support forensic analysis with confidence.
Learn more: Auditing for Fabric SQL Database - Microsoft Fabric | Microsoft Learn
Dynamic Data Masking (DDM) (Generally Available)
Dynamic Data Masking enables customers to reduce exposure to sensitive data by defining masking functions directly on selected table columns. When these masked columns are queried by users without appropriate privileges, the data is returned in a masked format, while the underlying data remains unchanged. This approach allows teams to protect sensitive information such as personal or financial data during development, analytics, and support scenarios, while maintaining full control over which columns are masked and how masking is applied—without requiring application changes.
Learn more: Dynamic Data Masking - SQL Server
Optimized for AI—Vector Index Improvements in SQL Database in Fabric
We have introduced significant enhancements to DiskANN Vector Indexes that dramatically improve performance, flexibility, and functionality for vector search operations. These enhancements deliver substantial improvements across multiple areas.
Learn more: VECTOR_SEARCH (Transact-SQL) - SQL Server | Microsoft Learn
Advanced vector quantization techniques have been integrated to provide better storage efficiency and faster query performance, with these optimizations being transparent to users and requiring no code changes.
Full DML support removes the previous limitation that made vector-indexed tables read-only after index creation. You can now perform full INSERT, UPDATE, DELETE, and MERGE operations while maintaining vector index functionality with automatic, real-time index maintenance.
Additionally, iterative filtering fundamentally improves vector search by applying predicates during the search itself rather than as a post-filtering step. This eliminates the need to over-fetch vectors and ensures consistent result counts when matching data exists.
These enhancements work seamlessly with the existing query optimizer that allows you to switch
between exact KNN searches and approximate ANN searches, providing a more robust, performant, and flexible vector search experience that supports dynamic applications with frequently changing data.
What’s next
These updates represent the next step in our commitment to making SQL database in Fabric the simplest, most autonomous, and most AI‑ready SQL platform for modern applications. We’ll continue listening closely to customer feedback as we expand capabilities across all three pillars of simplified; autonomous and secure; and optimized for AI.
To learn more, explore the product overview and tutorials and join the conversation by sharing your feedback on the Fabric Community Forums.