Microsoft Fabric Updates Blog

Empowering admins and developers with a Fabric platform ready for any project

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.

AI is rapidly changing how we create, use, and trust data. This shift places new, simultaneous demands on two critical groups. Developers are being asked to build production‑grade, data‑powered applications with the same rigor, automation, and observability expected of modern software systems. Administrators, meanwhile, must scale governance, security, and capacity to support explosive growth in users, data assets, and AI‑driven workloads without slowing innovation.

Microsoft Fabric is designed to meet both needs together. By bringing development, operations, governance, and security into a single, integrated platform, Fabric enables both groups to move faster with confidence. In this blog post, we’ll explore the latest Fabric capabilities we are bringing to empower both developers and admins.

Empowering data developers to build mission-critical applications on Fabric

Modern applications are data products. Your AI features depend on trusted, reproducible data, and your stakeholders expect every change to move from pull request to production with traceability, telemetry, and clear SLAs. In that environment, every developer is becoming a data developer.

Yet traditional data platforms were not built for developer workflows. Fragmented tools, manual handoffs, limited observability, and governance bolted on at the end slow teams down. The impact is real: at least 30% of enterprise LLM projects are projected to be abandoned after proof of concept due to poor data integration and grounding challenges (Gartner, Enterprise GenAI & LLM Research).

This is what we are solving with Microsoft Fabric. Our goal is to make data development feel like software development: automation at every step, repeatable builds, standard CI/CD, and governance and observability built in from day one. Today, we’re introducing a broad set of new developer-focused capabilities across the platform.

Fabric Model Context Protocol (MCP): Local GA and Remote preview

The Model Context Protocol (MCP) is an open standard for connecting AI applications to external systems. We are evolving Fabric MCP with two major milestones that give developers unprecedented flexibility in how they work with Fabric through AI.

Fabric Local MCP (Generally Available), providing an open-source local server that connects AI assistants to Fabric like GitHub Copilot, Claude, and Cursor. Developers gain access to OpenAPI specifications, item definition schemas, best practices, and OneLake file operations, enabling both code generation scenarios and direct interaction with Fabric resources in local-to-cloud workflows.

In parallel, we are introducing Fabric Remote MCP (Preview). This cloud-hosted MCP server enables AI agents and automation tools to perform authenticated operations in Fabric through a standardized protocol. Agents can create workspaces, manage permissions, work with item definitions, and more—all within existing RBAC boundaries and with full audit logging.

Together, local and remote MCP enable new workflows: agents use Local MCP to learn Fabric APIs and generate code, then use Remote MCP to execute operations in your Fabric environment—all through a single AI conversation.

Bringing Fabric into your existing development lifecycle with Git, CLI, and more

We are simplifying Git-based development in Fabric with enhancements that reduce setup time and eliminate unnecessary complexity. Selective branching allows developers to branch for a specific feature and pull only the items required, rather than duplicating an entire workspace. Built-in change comparison makes it easy to review differences before syncing. Branched workspace now provides clear visual context between feature workspaces and their source, improving confidence when using branch-out workflows.

Figure: YouTube video showing the built-in Git integration in Fabric

We are also expanding the Fabric Command Line Interface (CLI) to strengthen code-first development. CI/CD library integration is now generally available, enabling full workspace deployments directly from the command line. The CLI has been enhanced to better support AI agents, allowing them to generate deployment scripts, automate repetitive actions, and interact intelligently with Fabric. Finally, CLI added support for semantic model refresh and report rebinding through CLI commands.

We’re introducing the Azure DevOps Pipelines extension in Fabric, bringing Fabric natively into existing Azure DevOps CI/CD workflows. With a built-in Fabric CLI pipeline task that provisions automatically at runtime, teams can execute deployment and management commands without installing tools or maintaining custom build agents. We’re also announcing the preview of the Import and Export Item Definitions Batch APIs for Fabric, enabling the programmatic export, import, update, and synchronization of Fabric items across workspaces at scale.

And finally, we are launching a connection reference item type in Variable Library, allowing teams to reference approved connections by ID instead of embedding connection strings. This allows secure, environment-specific configuration across dev, test, and prod.

Extending and customizing Fabric with the Fabric Extensibility Toolkit (Generally Available)

We are also announcing that the Fabric Extensibility Toolkit (FET) (Generally Available). This toolkit represents the next evolution of the Workload Development Kit (WDK) but is redesigned to help any developer bring their data apps to Fabric for their own organizations along with a simplified architecture and additional automation to drastically streamline development. Developers can now simply build their own Fabric items, and everything else like distribution, user interface, and security is taken care of for you. With this release, we are adding full CI/CD support with variables and remote notifications for workloads. Management of workloads has also been made easier for admins with a new experience in the Admin portal and support via APIs.

Figure: Screenshot of the new Fabric Extensibility Toolkit’s getting started page

Streamlining the developer experience with improved UI in Fabric

We are improving everyday experiences in Fabric based on your feedback. First, we’re introducing recovery for items, giving you the ability to configure a retention period at the tenant level and recover accidentally deleted items during period.

We’re also working to improve the multitasking experience released in preview in 2025 with horizontal tabs for open items, support for multiple active workspaces, and a new object explorer. Now we are releasing the tabbed multitasking experience into general availability and further improving the experience with the ability to pin multitasking tabs, customize their layout, open them in a new browser tab, and support monitoring pages in multitasking tabs. Finally, we are releasing the ability to have multiple task flows in a workspace.

Figure: GIF highlighting the new multitasking experiences in Fabric

Learn, experiment, and deliver value faster with Fabric Jumpstart

Finally, we are launching Fabric Jumpstart, a new open-source initiative designed to accelerate adoption with ready-to-run, end-to-end solutions that deploy directly into your workspace. Fabric Jumpstart delivers a curated library of tested accelerators, demos, and tutorials that bundle datasets, notebooks, pipelines, reports, and supporting assets into a single self-install experience. With a single command, you can deploy real-world scenarios such as a real-time healthcare billing system, a stateful streaming lakehouse for industrial sales and logistics, or platform monitoring capabilities. With Fabric Jumpstart you are just minutes away from seeing best practice reference implementations, learning new concepts, and accelerating your Fabric journey.

Foster a secure and governed data culture with new tools designed for admins

AI is supercharging data cultures. Data professionals are building faster, and AI is helping business users find and apply insights. But as data culture accelerates, so do the demands of administrators. The volume of data items, access requests, and capacity requirements are all growing rapidly. And there is growing pressure to quickly enable AI initiatives, pushing the boundaries of acceptable risk.

The tension is clear. Gartner found that only 14% of security leaders can effectively secure data assets while also enabling the use of data to achieve business objectives. At the same time, 56% of nonexecutive directors believe organizations should increase their risk appetite in 2025 to drive shareholder value. Technical decision makers are being asked to move faster at a time when the risks have never been higher.

This is where Microsoft Fabric comes in. Fabric equips admins and data owners with industry-leading governance and security capabilities designed for the AI era, helping organizations scale innovation while maintaining full visibility, consistent policy enforcement, and control over their data.

Take control of your data estate in the OneLake catalog

Since its launch, the OneLake catalog has evolved into the central control plane for data across Fabric. It now provides a unified layer where discovery, governance, management, and security come together.

To enhance it further, we are releasing insights tailored for admins in the Govern tab into general availability. This new view provides immediate visibility into domains, capacities, workspace activity, protection coverage, and data curation. Admins can seamlessly move from high-level signals to detailed Power BI reports, act on personalized recommendations to remediate issues, or chat with Copilot to interpret trends and surface next steps. You can even build custom Power BI reports that fit your needs using the underlying semantic model.

Figure: GIF showing the new insights for Admins in the OneLake catalog Govern tab

The Govern tab also supports data owners, providing insights into the health of their data items and enabling them to take in-place actions to improve the curation, discoverability, and security of their data. Additionally, in the Explore tab, data owners can now apply tags to workspaces to improve discoverability for users and agents alike. Copilot can also now automatically generate descriptions for semantic models, helping item owners improve the discoverability of their models.

For developers, we are releasing public APIs for the OneLake catalog for search and discovery, providing programmatic access to catalog metadata. These APIs power relevancy-ranked search and are natively available through the Fabric MCP server, enabling AI agents to discover Fabric content.

Secure every step of the data journey with OneLake security and Purview data security

OneLake security will be generally available in the coming weeks, enabling data owners to define roles, enforce row- and column-level controls, and manage permissions through a single unified model. This security propagates automatically, so whether you query the data in SQL or open it in Excel, you can only see what has been authorized. We are also adding even more features which you can learn about in the OneLake @ FabCon & SQLCon announcement blog.

We’ve also expanded Purview Data Loss Prevention policies coverage to include data warehouses which is now generally available. Additionally, you can automatically restrict access to these sources as well as KQL and SQL data items, now in preview. If Purview detects sensitive data uploads, such as social security numbers, it will automatically notify your security admins and trigger any access restriction policies you have in place.

Additionally, Microsoft Purview Insider Risk Management integration with Fabric Lakehouse is now generally available. Admins can combine OneLake data access, DLP, sensitivity labeling, and audit signals to detect and investigate insider risk, reduce blind spots, and protect data. We are also adding support for quick policy creation for data theft scenarios in Purview IRM. Organizations now have more visibility into how much they are billed with the IRM PAYG usage report for Fabric, providing customers with an easy-to-use dashboard to track their consumption.

You’ll also be able to monitor Copilot in Fabric and Fabric data agents, using Purview DSPM, now in preview. You can use DSPM for AI to discover data risks such as sensitive data in user prompts and responses and identify and investigate risky AI usage.

Connect to your most sensitive data with end-to-end network security in Fabric

Since the launch of Microsoft Fabric, we’ve released a huge set of flexible security tools, from always-on security and encryption to fine-grained security and isolation for your most sensitive data. Now, we are taking another big step forward.

First, workspace-level IP firewall rules (Generally Available), this capability provides a way to control workspace access based on the incoming request’s IP address. This is especially useful for organizations that want to allow access to Fabric workspaces over public networks while maintaining controlled, secure entry points. By setting IP-based access rules, administrators can add a layer of network security to restrict inbound connections to only approved IP addresses.

We are also dramatically extending outbound access protection (OAP) to fifteen more items in Fabric. Outbound access protection can help you prevent exfiltration of data to unapproved locations outside of the organization. Now, OAP’s generally available coverage includes data integration items like dataflows and data pipelines, graph in Fabric, mirrored databases, OneLake shortcuts, and data warehouses. We are also releasing coverage for Power BI reports and semantic models, Fabric data agents, KQL, Cosmos DB, and more in Fabric in preview.

Customer managed keys (CMK) give you the ability to protect data using your own encryption keys. We are excited to announce you can now apply CMK to workspaces in a capacity which is already protected by Power BI’s bring your own key (BYOK), making it easier to adopt CMK. We are also releasing a new API for CMK to provide admins with a consolidated way to retrieve encryption status across all workspaces. Finally, CMK support for SQL databases in Fabric is generally available.

Lastly, we are extending Azure private links at the workspace level to cover SQL database and Activator items.

Gain more control and visibility into your Fabric capacity with new tooling

We’re introducing a new set of capacity management capabilities designed to give administrators greater control, flexibility, and insight. Surge protection for workspaces (Preview) allows you to define workspace usage limits, helping prevent non-critical workloads from consuming too many resources during unexpected spikes.

We’re also launching Fabric capacity overage (Preview) enabling admins to allow controlled, automatic over-consumption during peak demand and hence avoid throttling of critical workloads. Instead of provisioning for rare traffic surges, you can size capacity for typical workloads and enable overage only when needed.

To strengthen visibility, we are expanding monitoring capabilities in both the Capacity Metrics app and the Real-Time Hub. The updated Capacity Metrics app introduces enhanced health, history, and time-based analysis views, giving you deeper insight into workload trends.

Explore more Microsoft Fabric innovation

Alongside these announcements, we are introducing several innovations across Fabric. For a broad overview of all these announcements, read Arun Ulag’s hero blog “FabCon and SQLCon 2026: Unifying databases and Fabric on a single, complete platform.” If you want more details, check out the Fabric March 2026 Feature summary blog, the Power BI March 2026 feature summary blog, and the latest posts on the FabCon and SQLCon 2026: Unifying databases and Fabric on a single, complete platform, including the following blogs

Please enter your information to subscribe to the Microsoft Fabric Blog.