Fabric Highlights from Microsoft Build 2025

At Microsoft Build 2025, Microsoft’s annual event for developers and businesses, Microsoft has just released a wave of exciting updates and Fabric plays an important role in the data scene. As Microsoft’s all-in-one platform for data and AI, Fabric is making it easier than ever for companies to turn messy data into smart decisions, automate workflows and stay ahead in a world where AI is changing everything. In this blog, we’ll walk through the biggest Fabric news from Microsoft Build 2025.

Materialized Lake Views: Declarative Pipelines, Built-in Quality and Smarter Refreshes

Materialized Lake Views (MLV) in Fabric empower developers to define data transformations, such as aggregations, projections (selecting specific columns), or filters, using a simple “CREATE MATERIALIZED LAKE VIEW” SQL statement. These declarative pipelines automate common scenarios for refining data across layers, such as those in the medallion architecture. Once created, MLVs behave like standard Delta Lake tables: they can be queried by Power BI via Direct Lake (for high-performance analytics without data duplication) or by Spark for further processing. Behind the scenes, Fabric ensures the view stays up to date by refreshing it automatically based on a predefined schedule. Extra features include automatic lineage management and the ability to easily add data quality rules. A monitoring dashboard is included to keep an eye on your data quality. Databricks already has something similar called ‘Delta Live Tables’ (DLT). We are excited to see a similar feature making its entry into Microsoft Fabric.

Open Mirroring: Unify Data with Zero ETL

Now generally available with Parquet support, Open Mirroring simplifies ingesting data from any source (legacy apps, databases, custom solutions) into Fabric’s OneLake. No complex pipelines needed—just use the APIs or Python SDK to replicate data as Parquet/CSV into a landing zone. Fabric then auto-converts it to Delta format, optimised for AI/BI workloads. This feature enables building connectors for niche systems using the Open Mirroring Python SDK. Open Mirroring partnerships enable pre-built integrations with Oracle, SAP, MongoDB, Salesforce and more via partners like Oracle Golden Gate and Theobald.

Mirroring for SQL Server: Real-Time Simplicity

Microsoft Fabric now offers preview support for mirroring across all SQL Server versions (2016–2022), including the upcoming SQL Server 2025. With zero ETL, Mirroring syncs transactional databases to OneLake in near-real time: Add your SQL Server in the Fabric Portal, select tables, and let Fabric handle the rest. An initial snapshot populates data in Delta format in OneLake, with continuous updates streaming every transaction, from schema changes to row updates, ensuring data in OneLake stays up to date. This solution provides a bridge between operational databases and analytics.

Fabric CLI Hits GA

Fabric command line interface, first unveiled at FabCon Las Vegas, is now generally available. Thousands of developers already leverage its file-system-like interface for scripting and local workflows; now enterprises can confidently embed it into CI/CD pipelines, deployment and automation frameworks.

Copy Job’s CDC Support (Preview)

Copy Job now natively supports Change Data Capture (CDC) in preview, automating replication of inserts, updates, and deletes from sources like Azure SQL DB, SQL Server, and Azure SQL Managed Instance. CDC ensures destinations stay synced with minimal source load: it processes only changed data, eliminating manual intervention. For non-CDC sources, Copy Job defaults to watermark-based incremental updates (using timestamps/IDs) to merge changes. With intelligent detection of CDC-enabled tables and hybrid flexibility, teams optimise pipelines for real-time accuracy or batch efficiency—all while Fabric handles the heavy lifting.

Fabric new security features (Preview)

Microsoft unveiled three critical security features which are now in preview, empowering enterprises to safeguard sensitive data with precision while maintaining compliance:

1. Workspace-Level Azure Private Link (Preview)

Secure inbound traffic to specific workspaces (e.g., isolating production from development) via private networks, bypassing the public internet. Launching in early summer, this granular control ensures compliance without hindering collaboration in less-sensitive environments.

2. Outbound Access Protection for Spark (Preview)

Block accidental or malicious data exfiltration by restricting Spark workloads to approved destinations. Initially targeting Spark, this feature (preview in early summer) will soon extend to Data Factory pipelines, dataflows and Copy Jobs.

3. Customer-Managed Keys (CMK) for Encryption (Preview)

Add a second encryption layer using your own keys, stored in Azure Key Vault, to meet strict regulatory demands. Now available in select regions, CMK complements Fabric’s default encryption, putting control of data-at-rest security in your hands.

Warehouse Snapshots (Preview)

Microsoft Fabric’s Warehouse Snapshots (now in preview) deliver read-only, point-in-time views of your data warehouse, frozen to avoid conflicts with live updates or ETL processes. Retained for 30 days, snapshots let analysts run consistent queries for reporting or historical analysis, while engineers seamlessly “roll forward” snapshots to the latest state via atomic operations, ensuring third-party tools always access fresh data through unchanged connection strings.

Composite Models Evolve: Direct Lake + Import Mode

Microsoft Fabric now supports hybrid semantic models combining Direct Lake (querying OneLake directly) and Import mode (full Power BI feature support) in a single solution. This breakthrough lets you blend massive-scale analytics with advanced modelling. Hybrid models merge speed and flexibility—keep large fact tables in Direct Lake for performance and enrich dimensions with Import mode’s advanced features. Developers can build these today using XMLA endpoints (SSMS, Tabular Editor, Fabric Notebooks) or wait for web modelling (coming soon). In addition, hierarchies and calculated columns in Import tables now work seamlessly with Analyze in Excel.

Translytical Task Flows (Preview): Power BI Reports That Act

Microsoft Fabric’s Translytical task flows (in preview) bring automation to Power BI reports, letting users trigger actions directly from dashboards. Powered by Fabric User Data Functions, these flows bridge analytics and operations: update CRM records, annotate datasets, or send dynamic notifications with a click. For example, a sales report could auto-alert managers when deals stall, or a supply chain dashboard could adjust inventory levels in real time. By embedding actions into reports, Fabric turns insights into outcomes, all while integrating with external systems like Teams or Azure Logic Apps. Translytical task flows collapse the gap between analysis and action, letting teams automate workflows where decisions happen.

What’s next?

Microsoft Build 2025 has redefined what’s possible with Microsoft Fabric, delivering innovations that bridge data, AI, and actionable insights—from Materialized Lake Views simplifying medallion architecture to Translytical task flows turning reports into workflows.

To stay ahead of what’s next, explore the newly redesigned Microsoft Fabric Roadmap. This interactive hub lets you track upcoming features. Whether you’re excited about data engineering, real-time analytics, or AI integration, the roadmap offers transparency into Fabric’s evolution.

As Fabric continues to evolve, one thing remains clear: the future of data isn’t just about managing information—it’s about empowering every team to act on it. Dive into the roadmap, experiment with today’s previews, and join us in building a smarter, more connected data ecosystem.

Curious about how Microsoft Fabric can benefit your organisation? Don’t hesitate to reach out—we’d be happy to discuss how it can create value for you!