2025 Microsoft Fabric Community Conference Announcements

Problem

The second edition of the Microsoft Fabric Community Conference (nicknamed FabCon) was held from 31 March 2025 to 2 April 2025 in Las Vegas. As usual, the conference was packed with announcements of new features and improvements. The sessions themselves were not recorded. However, during the conference, Microsoft released a ton of information via the official Microsoft Fabric blog.

In this tip, we’ll make a valiant attempt to summarize the 2025 FabCon news.

Solution

This news overview is not a comprehensive list of everything announced or presented at FabCon, but a selection of the most important items (which might be subjective). If possible, referrals to external sources are made for additional reading.

Copilot

Probably the biggest news – at least according to Microsoft – is that the F64 capacity restriction Copilot and AI capabilities has been lifted. This means Copilot can start at an F2 capacity. (Note: This may not be the best idea since it can consume significant resources).  With the recently added Fabric Copilot Capacities (FCC) feature, or dedicated capacities creation, for Copilot usage, you can now truly democratize the use of AI at your company.

fabric copilot capacities

OneLake Security

The security model of OneLake will get an update. Called OneLake security, it will make it possible to define row- and column-level security directly in OneLake. These security roles will be followed when workloads like Spark, the SQL Analytics Endpoint, or Power BI Direct Lake read from the delta tables in OneLake. This means it will be easier to implement security, as there will be no need to apply rules in multiple engines. However, in some engines such as the warehouse, SQL DB, or Power BI in import mode, you may still need to configure security rules.

This feature is currently in private preview, but you can sign up for early access. More information can be found in the blog post The next evolution of OneLake security (Preview).

OneLake security

(source: The next evolution of OneLake security (Preview))

Fabric Data Factory

A long-awaited feature, support for Azure Key Vault, helps define connections in your pipelines, similar to Azure Data Factory or Synapse Pipelines. This will fetch secrets from a Key Vault, enhancing the security robustness of your solution.

Azure key vault

(source: Fabric Data Factory: What’s New and Latest Roadmap)

CI/CD support for Data Pipelines is now generally available. However, if you deploy a data pipeline using a deployment pipeline to another workspace, your connections still point to the original workspace. It’s not possible to override your connections using deployment rules, although possible for semantic models, for example. Consequently, you need to parameterize your connections (explained in this post: Dynamic Warehouse & Lakehouse Connections in Microsoft Fabric Data Pipelines), probably in combination with the new feature, variable library. (Variable library is still in preview, and at the moment, the limitations mention that it doesn’t support connection parameterization.)

Other updates include:

  • Service Principal Name (SPN) authentication: Now supported for the pipeline CRUD APIs.
  • Database mirroring: Updates include the ability to reach sources behind a firewall, mirroring for Azure Database for PostGreSQL Flexible Server, and CSV support for Open Mirroring. Find the detailed list: Mirroring in Fabric – What’s new.
  • Activities: You can now have up to 120 activities in a single pipeline (although it’s advised to build more modular pipelines).
  • Lift&Shift: Support for the lift&shift of SSIS packages is in private preview.

There are additional updates not mentioned here.  For more information, check out the blog post Fabric Data Factory: What’s New and Latest Roadmap for the full details.

Real-Time Intelligence (RTI)

There are a couple of announcements for RTI in Fabric:

  • New connectors for Eventstreams, such as MQTT, Solace PubSub+, Azure Data Explorer, real-time weather, and others.
  • Multiple schemas support.
  • Data Preview for Apache Kafka, Confluent Kafka, and Amazon Kinesis.
  • General Availability for Azure and Fabric Events (OneLake events, Azure Blob Storage Events, Job and Workspace events).
  • CI/CD support.
  • Preview for Managed Private Endpoint and Entra ID support for custom endpoints.
  • KQL Database Entity Diagram.

You can find more information in the blog post Unlock the power of Real-Time Intelligence in the Era of AI: why Fabric Real-Time Intelligence is a game-changer.

User-defined Functions in the Warehouse

The Fabric Warehouse now supports three types of user-defined functions:

  1. SQL-based functions, just like we know them from SQL Server. Previously, the Warehouse only supported inline table-valued functions, but the private preview of scalar-value functions in the Warehouse has been announced. Both types of functions are already available in the SQL DB.
  2. Fabric-based functions, which are now and are quite similar to (serverless) Azure Functions. For the moment, only Python is supported but other languages will follow. You can integrate those functions into notebooks or pipelines, but now you can also call them directly from SQL!
  3. AI-based functions, which leverage large language models (LLMs) to bring powerful AI capabilities to the Warehouse. There are built-in AI functions, such as ai_analyze_sentiment, but through the ai_generate_response function, you can also create user-defined AI functions.
ai_generate_response function

(source: Another dimension of Functions in Data Warehouse)

Additional Features

Next Steps

  • You can find all Microsoft Fabric tips in this overview.
  • Marc Reguara from Microsoft has shared a list with a full overview of all Fabric blog posts containing announcements on LinkedIn. Be sure to check it out, as some updates didn’t make it into this overview!

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *