This year’s Microsoft Fabric Community Conference was bigger than ever. FabCon 2025 in Vegas hosted over 6,200 attendees, including data professionals, business leaders, Microsoft MVPs, and partners. Across six energizing days, the event featured 220+ sessions, 20 hands-on workshops, and contributions from 70+ sponsors, making it a true hub of learning and collaboration. The goal of Microsoft Fabric is to unify all data services required by an organization under a unified, extensible, and AI-ready platform, eliminating the need to juggle between disconnected tools and siloed operations. Fabric has already been adopted by over 21,000 organizations, which include 70% of the Fortune 500 companies. With more than 50% of Fabric customers using at least three workloads, Fabric is evolving as the nerve center for enterprise data needs.
Let’s catch up on everything new that FabCon 2025 announced for Fabric users, including cutting-edge new features to enhance ease of use and strengthen security. Here are some of the strategic takeaways and key highlights from Microsoft Fabric Conference 2025:
- Agentic AI Enhancements
Microsoft has rebranded Data Agents (previously AI Skills) and made them available across Fabric and Azure AI Foundry. These agents can connect to enterprise data and understand context to take actions. In practice, they can query OneLake to analyze patterns and trigger workflows on their own. This moves Fabric from being a reporting tool to a decision-making platform. For leaders, the key shift is to prepare teams to govern and measure the impact of these AI-driven agents. - Advancing Real-Time Intelligence in Microsoft Fabric
Real-time is now a core Fabric capability, empowering businesses to act on data as it arrives. Microsoft added new eventstream connectors for Weather, Confluent, and MQTT to bring more live data into OneLake. Be it fraud alerts, supply chain updates, or customer signals, businesses can now act quickly without compromising on security. - Security Updates for OneLake
OneLake is being pitched as the shared foundation for organizations’ data estates and is the new control panel for access. The “define once, enforce everywhere” enables fine-grained access controls at row and column levels in such a way that the same authorizations propagate to SQL, Spark, and Power BI.
It also highlighted Microsoft Purview DLP expansions and Data Observability to detect risky AI usage and sensitive data exposure. In short, Fabric’s security evolution aims to reduce policy duplication and give security teams practical, enforceable points.
- Practical Innovations for Data Teams
- Spark Workload Optimization using Autoscale Billing: Spark jobs can now be offloaded to a serverless, pay-as-you-go billing mode to avoid consuming shared capacity. This Autoscale Billing for Apache Spark helps separate steady capacity from sudden heavy workloads.
- AI for Every SKU (Copilot availability for all SKUs in Fabric): Copilot and Fabric AI features have been made available on all F2 and above paid capacities, lowering the entry bar for AI usage. This allows small teams to benefit from AI-driven insights and automation.
- Advancements in Mirroring for Microsoft Fabric: Mirroring from external databases now supports more sources and provides higher flexibility in configuration. Mirroring operations have been integrated with workspace monitoring, allowing users to track replication, latency, and failures. It also preserves source schema hierarchy, ensuring data consistency across the data estate.
- Notebook Innovations: Updates to the Fabric notebook are primarily focused on Copilot integration and AI functions for data preparation on Pandas, without having to use Spark.
- Apache Airflow Job Support: Apache Airflow is now generally available, supporting jobs directly inside Fabric, eliminating the need to set up separate infrastructure. This means teams can orchestrate complex data workflows in a managed, cloud-native environment.
- Fabric CLI (Preview): A new Command Line Interface (CLI) has been added, beneficial for automation, CI/CD scripting, and integrating with DevOps workflows. This helps maintain consistent environments across teams and enables faster rollouts.
- Fabric User Data Functions (Preview): User Data Functions were formally introduced in FabCon2025. It enables teams to embed reusable Python logic inside Fabric and removes dependency on external services like Azure Serverless Functions. This was highlighted as a key update in write-back enhancements, allowing users to update data directly within Fabric. It is particularly beneficial in scenarios such as financial planning and real-time adjustments within Fabric.
- Query Insights: This is a built-in query monitoring feature that provides centralized insights on long-running queries and frequency, giving teams visibility into query performance bottlenecks.
- Terraform Provider: Terraform integration is now supported in SQL database, enabling CRUD operations through Infrastructure-as-Code (IaC), for consistent and scalable deployment. The General Availability release also enables full Infrastructure-as-Code (IAC) for managing Fabric environments, supporting key DevOps practices for Data teams.
- Simplified Data Ingestion
- Direct Lake Semantic Models in Power BI Desktop:
The integration of Direct Lake semantic models into Power BI Desktop allows users to create and edit these models directly by connecting to multiple Fabric data sources. This simplifies data modeling from various artifacts and eliminates manual data refreshes. - Excel Integration with OneLake:
Microsoft announced a major update, enabling direct integration of OneLake data with Excel for Windows. Users can retrieve data from OneLake through the modern Get Data experience in Excel, empowering Excel-heavy operations to tap into the power of OneLake. - Synapse to Fabric migration:
Migrating from Azure Synapse to Microsoft Fabric has been made easier through an in-built Migration Assistant. It comes with smart assessments and Copilot assistance for seamless data and code transfers. It automatically converts Synapse schemas into Fabric format and reduces complexities. This accelerates the adoption of Fabric as the unified platform for an organization’s data requirements. - Copy Job (General Availability):
It is now generally available on Fabric, offering a quicker way to move data into Fabric from various sources. It now supports a broader range of connectors such as Oracle Database, Google BigQuery, and Amazon S3. Key updates include Copy Job Activities in data pipelines, incremental copy, and public API for integration with GitHub and Azure DevOps.
- Direct Lake Semantic Models in Power BI Desktop:
Staying Ahead with Microsoft Fabric
Microsoft Fabric is becoming the standard for enterprise data modernization and AI adoption. With its rapid influence across industries, including healthcare and banking and financial services, it is evident that organizations embracing Fabric will stay ahead of the curve, positioning themselves at the forefront of the AI revolution, while also securing their data in the process.
With a trusted Microsoft implementation partner, organizations can chart out a clear roadmap. From decoding pricing to optimizing workloads, our expert team at KANINI can guide and provide continuous support to Fabric adopters.
Talk to us today to learn more about how you can start leveraging these latest features from the Microsoft Fabric Conference 2025 in your enterprise.
Stay tuned for new features and the latest announcements from Microsoft events!
Frequently Asked Questions
FabCon 2025 is a Microsoft event dedicated to data and AI professionals to explore how Microsoft Fabric helps create unified, secure, and intelligent data environments. It brings together industry leaders, partners, and experts to share insights and new MS Fabric features that drive data modernization.
Microsoft Build or Ignite covers a wide range of technology topics, whereas FabCon is focused on data and AI solutions within Microsoft Fabric. It offers immersive and hands-on sessions designed for teams seeking to modernize their data platforms and integrate AI into their operations.
FabCon 2025 launched key innovations like upgraded security for OneLake, enhanced real-time analytics, Autoscale Billing for Spark, and several practical innovations for data teams to build faster and more secure data solutions. Read the full article for details.
Yes, recorded sessions from the conference are available on the Microsoft Blog and YouTube channel.
Author
Prashant Sharma
Prashant Sharma is the Data Analytics & AI Practice Lead at KANINI, with over two decades of experience in solution development and delivery. He leverages deep expertise in data analytics and GenAI to lead transformative projects, mentor agile teams, and drive data-driven innovation across industries.
