Just a few weeks after attending the European Microsoft Fabric Community Conference (known as FabCon Europe) in Vienna (Austria), I’m still reflecting on the many announcements and innovations unveiled there — and continuing to engage with customers eager to go deeper.

Today, I want to share why Microsoft Fabric is rapidly redefining the enterprise data intelligence platform landscape. With its unified architecture, deep AI integration, and continuous innovation, Microsoft Fabric is positioning itself as the go-to solution for organizations seeking scalable, intelligent, and real-time data capabilities. It's also a platform that welcome business users and non-advanced data engineers — which is super important to innovate at scale and increase adoption.

And don’t worry—I’ll skip Power BI this time! We all know Microsoft continues to lead the Business Intelligence (BI) space, being recognized by analysts like Gartner as a leader for the seventh consecutive years.

That said, here are the topics I’ll be covering today:

• Ecosystem Momentum and Community Energy

• Unified Data Foundation: OneLake (including mirroring and shortcuts)

• Real-Time Intelligence (RTI): Planet-Scale Streaming (including the Digital Twin Builder)

• Fabric MCP

• Performance That Scales with Value

Let’s get started!

Ecosystem Momentum and Community Energy

Back in March, FabCon Americas 2025 in Las Vegas was a tremendous success—bringing together over 6,100 customers and partners, making it the second-largest Microsoft conference of the year. The event featured 220+ sessions, 20 hands-on workshops, and a high-impact executive track at Allegiant Stadium.

Building on that momentum, FabCon Europe 2025 in Vienna showcased the rapid growth of the Fabric community, with 4,200+ attendees, 130+ expert-led sessions, and exclusive experiences such as Fabric Partner Day, roadmap deep dives, AMAs, and a vibrant expo hall—making it the largest Fabric-focused event to date in Europe.

Looking ahead, we invite you to join us at FabCon Americas 2026 in Atlanta, taking place March 16–20 at the Georgia World Congress Center. Expect deep technical workshops, inspiring keynotes at State Farm Arena, and a vibrant community experience—including the second annual Data Viz World Championship and an unforgettable attendee celebration at the Georgia Aquarium.

Unified Data Foundation: OneLake

At the core of Microsoft Fabric is OneLake, a multi-cloud, open-format data lake that powers every workload (or data engine) across the platform. Data is stored in open Delta Parquet—ensuring no vendor lock-in and full interoperability through industry-standard APIs.

OneLake provides organizations with a Data Mesh as a Service (DMaaS), allowing logical grouping of data by business domain and data products. Something I personally love is that every dataset in OneLake is automatically indexed for discovery, MIP labeling, lineage tracking, PII scanning, governance, and compliance—simplifying data management at scale.

OneLake seamlessly connects data across Power BI, Data Engineering, RTI and Data Enginering — eliminating silos and reducing latency.

Microsoft Fabric modernizes data movement with three flexible options:

• Shortcuts – Instantly access data across clouds and domains without duplication, reducing storage needs and latency.

• Mirroring – Achieve near real-time freshness by continuously syncing operational databases directly into OneLake, eliminating complex ETLs (Extraction, Transformation & Load).

• Open Mirroring – Enable developers and partners to stream data changes through APIs, using the open Delta format for system-wide consistency.

To learn more, please check my previous article: Unlocking Business Agility with Multi-Agent AI and Microsoft Fabric’s Data Mirroring

At FabCon Europe, we announced the General Availability of Databricks Mirroring, alongside new Oracle (including Exadata) and Google Big Query mirroring capabilities (both now in Public Preview), which extend Microsoft Fabric’s reach into legacy systems like SAP and Oracle—truly making OneLake a universal data foundation.

Another exciting announcement around OneLake is the preview of Table APIs — a major step forward in open and interoperable data management. These new APIs enable programmatic interaction with data tables via REST, built on open standards like Apache Iceberg. OneLake now automatically serves Delta Lake tables as Iceberg tables, enabling seamless access from platforms such as Apache Spark park and Snowflake, both of which natively support Iceberg. This innovation eliminates the need for complex ETL pipelines, embraces a zero-copy architecture, and empowers developers to build custom applications and analytics workflows across diverse ecosystems — all while reinforcing Microsoft Fabric’s commitment to openness and flexibility.

If you’re concerned about egress costs when using cross-cloud sources with shortcuts, it’s worth noting that the cross-network connectivity feature is now Generally Available (GA). This caching feature allows you to set a data retention period between 1 and 28 days, with the flexibility to reset the cache at any time through the UX or API.

Another exciting updates from FabCon was the announcement of Shortcut Transformations, a new capability in Microsoft Fabric that simplifies the process of converting raw files, starting with .CSV files, into Delta tables. This feature eliminates the need for traditional ETL pipelines, enabling users to transform data directly on top of files with minimal setup. In my opinion, this new feature will; help users to accelerate time-to-insight by skipping complex ETL pipelines, enable self-service data transformation for business users, and maintain governance with built-in monitoring and lineage tracking.

Together, these deliver on the promise of zero copy, zero ETL, and immediate, secure access to data—unlocking faster insights at scale.

Real-Time Intelligence (RTI): Planet-Scale Streaming

Microsoft Fabric Real-Time Intelligence (RTI) is a unified, event-driven analytics platform designed for data in motion. It enables ingestion, transformation, process, modeling, analyze, and visualization of streaming data with minimal latency—empowering industries like Manufacturing, Logistics, Oil & Gas, Mining, and Telecommunications to act in real time.

During the Executive Track at the FabCon Europe, Yitzhak Kesselman shared that 46% of Microsoft Fabric customers (over 25,000) are now using RTI, making it the fastest-growing workload in Microsoft Fabric with a 6x year-over-year (YoY) usage increase.

It its core, the Real-Time Hub centralizes streaming inputs from Azure Event Hubs, Apache Kafka, and Google Pub/Sub. Data flows into Eventstream for transformation, is stored in Eventhouse (a high-performance, KQL-based storage), and can be visualized through Power BI, Reflex, or Activator dashboards.

A highlight from Microsoft Ignite 2005 was the Digital Twin Builder (DTB), now in public preview. DTB allows organizations to model real-world entities and processes for advanced analytics and simulation.

In case you don't know, ontology provides a structured, semantic layer to model real-world entities and relationships for intelligent operations.

• Modeling: Allow us to define domain-specific entities (e.g., machines, sensors, processes). We can use Digital Twin Builder’s low-code/no-code interface for easy creation.

• Mapping: A way to connect real-time data streams (e.g., from Kafka, Azure IoT) to ontology concepts. Enables semantic alignment and consistent interpretation.

• Relationships: Provide a tool to set the model interactions between entities (e.g., pump → cooling system). Supports AI reasoning, anomaly detection, and contextual insights.

• Extensions: Allow as to add new entities, attributes, or relationships as business needs evolve. Ensures scalability and adaptability of the digital twin.

DTB's key capabilities include:

• Ontology modeling for standardized asset representation

• Data mapping from systems like SCADA, MES, and ERP

• Semantic relationships for contextual insights

• Low-code/no-code experiences for operators

• Integration with Power BI and Real-Time Dashboards

RTI also supports both low-code and pro-code users, offering schema discovery, natural language querying (NL2KQL), and AI agent integration through the Model Context Protocol (MCP)—enabling tools like GitHub Copilot to interact with Fabric data.

By the way, the MCP for Microsoft Fabric was also announced in Preview. Fabric MCP is a developer-focused framework that unifies Fabric’s public APIs, item definitions, and best-practice guidance into a single contextual layer designed for AI-assisted development. It includes a local MCP server that packages the full OpenAPI specifications for Fabric’s public APIs, JSON schemas for all item types (such as Lakehouses, pipelines, semantic models, notebooks, and Real-Time Analytics workloads), along with built-in guidance on pagination, error handling, and other key development best practices. You can learn more here.

We also announced the public preview of Graph, Maps and Anomaly Detection in Microsoft Fabric.

• Graph in Fabric enables organizations to visualize and analyze complex relationships that influence business outcomes.

• Maps in Fabric brings geospatial intelligence to AI-driven operations, transforming vast amounts of location-based data into interactive, real-time visualizations that enhance situational awareness and empower smarter, location-aware decisions.

• Anomaly Detection which automatically identifies unusual patterns and outliers in our Eventhouse tables. It provides recommended models and enables continuous monitoring with automated actions to help organizations respond proactively to emerging issues.

To explore these capabilities in more detail, check out Yitzhak Kesselman’s blog, “The Foundation for Powering AI-Driven Operations: Fabric Real-Time Intelligence”.

Performance That Scales with Value

Having these capabilities is impressive, but customers also demand industry-leading performance—and I’m excited to share that Microsoft Fabric delivers incredible performance at a fraction of the cost compared to legacy platforms.

• The Native Execution Engine provides up to 4x faster performance on Spark workloads than open-source Spark, with internal benchmarks showing up to 6x end-to-end gains. Automated Table Statistics further enhance query efficiency, improving complex workload performance by up to 45% by enabling Spark’s Cost-Based Optimizer to make smarter decisions.

• During FabCon Europe, Bogdan Crivat shared that Fabric Data Warehousing has introduced over 40 significant enhancements since its 2025 launch, resulting in a 35% performance improvement year over year.

• When compared to Snowflake, Microsoft Fabric offers up to 3x lower costs for large-scale workloads (e.g., 100 TB under heavy compute).

• Similarly, Fabric Dataflow Gen2 improved runtime from 4,672 seconds in Gen1 to 2,944 seconds, and with the new modern query evaluator, execution time dropped further to 1,601 seconds—and when parallelized, it’s reduced to an astonishing 480 seconds.

• During FabCon Europe, Microsoft also announced pricing improvements that translate into major cost savings:a up to 25% savings for the first 10 minutes (equivalent to 12 CUs) and up to 90% savings beyond 10 minutes (equivalent to 1.5 CUs).

• Additionally, customers can achieve up to 41% savings with reserved capacity SKUs (F2–F2048), all with transparent pricing.

These metrics are just a few examples that clearly underscore Fabric’s ability to deliver high-performance analytics and real-time intelligence with predictable, optimized cost structures.

Final Thoughts

Microsoft Fabric is far more than the next evolution of Azure Synapse or Power BI—it’s a strategic enabler for AI transformation, real-time decision-making, and enterprise-scale analytics.

With its unified data architecture, governance-first design, and AI-driven intelligence, Microsoft Fabric is redefining how organizations modernize their data estates and unlock the full potential of their data.

Last and not least, I would like to invite you to learn more about all the highlights from the Fabcon Europe 2025 by reading this post and also an Arun Ulag's post: Build data-rich agents on an enterprise-ready foundation