Fabric Copilot Capacity (FCC): Copilot Now Available for All Capacities

Microsoft has announced an update that redefines access to artificial intelligence: Copilot and other AI resources are now available for all Microsoft Fabric capabilities, from the most basic level (Pro, Trial, PPU, P1, F2) for organizations with an F64 or P1 capability where Copilot Capacity has been enabled.

Microsoft has introduced the concept of “Fabric Copilot Capacity (FCC)” to expand access to GenAI within Fabric, as well as to make it easier to manage and leverage AI resources within the organization.

This means that you can use Copilot throughout your company for just over 50€, as long as your company has an F64 license.

A major step towards the democratization of enterprise artificial intelligence, in line with Microsoft's commitment to make AI capabilities available to all Fabric users.

In this article we explore in detail what subscription/licensing modalities or levels enable the use of Copilot, how AI capabilities are enabled, what requirements to consider, what are the possible limitations, how it affects pricing, and what impact it has on related services such as Power BI Embedded.

Microsoft's progressive release of AI capabilities through Fabric Copilot Capacity (FCC), announced to be deployed over “the next few weeks” is in direct response to requests from the user community.

What is Fabric Copilot Capacity (FCC)?

Fabric Copilot Capacity (FCC) is a configurable option in Microsoft Fabric capabilities —F64 or P1 needed— to allow users to access Copilot from workspaces associated with lower capacities and licenses.

Instead of each workspace using the resources of its primary capability to run Copilot, companies can now redirect all AI-related loads to a particular Copilot capability, thus optimizing resource consumption and centralizing the execution of these intelligent processes.

This has several advantages:

• It centralizes AI usage in a single location.
• Enables more efficient management of budget and consumption.
• You can assign specific groups or users to that capability.

Download the Complete Guide on Microsoft Fabric here

Why Use a Fabric Copilot Capacity (FCC)??

In large organizations with multiple departments or users, there may be a wide variety of workspaces without dedicated capacity. If all these users start using Copilot, consumption can become unpredictable.

Designating a Fabric Copilot Capacity (FCC) allows you to:

• Consolidate AI-related usage into a capacity of your choice
• Easily monitor how much Copilot is being used
• Ensure better performance by assigning a more powerful capacity for AI tasks—even if the data resides in smaller workspaces

How to Set Up Fabric Copilot Capacity (FCC)?

1. To begin, a Microsoft Fabric administrator must enable Copilot at organization level. Doing so will allow users across the company to access AI features within the platform.

2. Next, the Fabric admin needs to authorize capacity admins to assign a specific capacity as a Fabric Copilot Capacity (FCC). This step is essential to centralize Copilot usage in one managed environment.

3. Once the FCC is designated, the capacity admin can then assign specific users or groups to use that capacity for Copilot operations — no matter which workspace they're working in.

Example: Let’s say your team is working in a workspace that runs on an F2 capacity, but you don’t want their Copilot activity to consume F2 resources. In this case, you can:

• Set up a more powerful F64 capacity to serve as your Copilot Capacity
• Assign the team to the F64 (FCC)
• From that point on, all Copilot prompts from that team will be executed in the F64, even if their data remains in the F2 workspace

Operational Advantages:

• Improves traceability of AI usage
• Allows you to separate AI usage from other analytics operations
• Simplifies resource planning and scaling when Copilot is heavily used
• Prevents overloading smaller capacities with AI-intensive workloads

Fabric Copilot Capacity (FCC) from Pro, Trial, PPU and F2: Democratizing AI in Fabric

Until recently, Copilot was only available in select capacities and required high-tier licenses such as Power BI Premium P1 or Fabric F64 (64 capacity units), excluding smaller SKUs and capacities used exclusively for embedding.

However, with the latest update, Fabric Copilot Capacity (FCC) is now supported starting from F2, the most basic tier among Microsoft Fabric’s paid capacities.

This means that organizations with multiple capacities — and at least one with Copilot enabled in Power BI — can now use Copilot in any Power BI report, provided it’s linked to a capacity with Copilot support:

• Any Power BI report associated to a higher capacity, like F64, can run Copilot—even if the corresponding workspace is hosted in a lower-tier capacity.

Now, all Power BI users can access Copilot through the most basic Fabric capacity (F2), as long as an admin enables a Fabric Copilot Capacity (FCC) and gives them access.

As Microsoft stated in their official announcement:

“Customers with F2 and above can now use Copilot and AI capabilities to automate workflows, generate insights, and make impactful decisions.”

This marks a major shift in how organizations access generative AI within the platform. It enables broader adoption of GenAI tools across teams, empowering more users to create, explore, and analyze data using natural language—through Copilot and its integrations with tools like Power BI, Data Factory, and Data Science.

Fabric Copilot Capacity (FCC): Important Considerations

• FCC is not available in free or trial environments. Evaluation capacities do not support Copilot.
• To activate a Copilot Capacity, you must have at least an F64 or P1 capacity.
• Licensing alone is not enough — the capacity itself must be explicitly enabled for Copilot.
• This requires admins to activate Copilot within the specific capacity from the admin portal.
• No new license or price increase is required. Copilot features are included as part of Fabric capacity benefits.

Fabric Copilot Capacity (FCC): Licensing and Pricing

The expanded availability of Fabric Copilot Capacity (FCC) does not change Microsoft Fabric’s pricing model, but it significantly lowers the barrier to entry for enterprise-grade artificial intelligence.

Fabric Copilot Capacity (FCC): Capacity Pricing Table (EUR)

The following table outlines Fabric's main capabilities, their size, estimated monthly cost, estimated monthly cost adapted to working hours and the availability of Copilot and AI capabilities:

*Notes: The column “Estimated monthly price (EUR)” reflects the estimated monthly cost under the pay-as-you-go model in euros, considering the continuous use of the service 24 hours a day, 7 days a week, throughout the month. This is the full cost for permanent availability of capacity (CUs). In contrast, the column “Adjusted monthly price (8h/day, 5d/week)” shows the estimated monthly cost if the Microsoft Fabric capacity is used only during typical working days (8h/day, 5 days/week), which is equivalent to about 173 hours of use per month (8h × 5 days × 4.33 weeks). Given that a full month has 720 hours (24h × 30 days), this usage pattern represents approximately 23.8% of the total monthly time.

Fabric Copilot Capacity (FCC): Capacity Pricing Table (USD)

*Notes: The “Estimated monthly price (USD)” column reflects the estimated monthly cost under the pay-as-you-go model in dollars, considering the continuous use of the service 24 hours a day, 7 days a week, throughout the month. This is the full cost for permanent availability of capacity (CUs). In contrast, the column “Adjusted monthly price (8h/day, 5d/week)” shows the estimated monthly cost if Microsoft Fabric capacity is used only during typical working days (8h/day, 5 days/week), which is equivalent to about 173 hours of usage per month (8h × 5 days × 4.33 weeks). Given that a full month has 720 hours (24h × 30 days), this usage pattern represents approximately 23.8% of the total monthly time.

Fabric Copilot Capacity (FCC) does not only save on licensing costs but also improves operational efficiency. It helps:

• Accelerate productivity for data teams
• Automate repetitive analytics tasks
• Reduce reliance on technical roles

This translates into a stronger ROI through faster, more data-driven decision-making.

A Capacity-Based Model, Not Per-User Licensing

Microsoft Fabric uses a capacity-based licensing model. You pay for the capacity tier (e.g., F2, F4, F64), regardless of the number of users or features accessed. Every paid SKU — including Pro, Trial, PPU, and F2 — includes Copilot at no additional cost.

Example:

• F2: 2 CUs → ~$262/month
• F64: 64 CUs → ~$8,400/month
• F2048: 2048 CUs → ~$269,000/month

Flexible Scaling

The pay-as-you-go model lets you scale capacity temporarily to meet demand. For example, you can increase capacity during hackathons, data-intensive campaigns, peak project phases and scale back afterward to control costs.

Copilot Capacity: Additional Costs & Storage Considerations

• OneLake includes baseline storage based on the SKU tier
• Additional storage is billed separately
• While Copilot itself doesn’t store data, its use may generate more content and activity, indirectly impacting storage usage

💡 Tip: Use the official Microsoft Fabric SKU Calculator to get accurate cost estimates tailored to your specific use case.

How to Activate a Fabric Copilot Capacity (FCC)?

Activating a Fabric Copilot Capacity (FCC) in Microsoft Fabric is a relatively straightforward process, but it does require meeting a few technical, licensing, and configuration prerequisites.

Below is a step-by-step explanation of what you need to get started with Copilot and generative AI capabilities in your environment.

1. Fabric Capacity in a Supported Region

To enable Fabric Copilot Capacity (FCC), your capacity must be hosted in a region where Copilot is available.

• Currently, commercial regions in the United States and the European Union are supported by default.
• In these regions, FCC is automatically available when you purchase a paid Fabric capacity.

What if you're outside the US or EU?

• If your tenant is located in another region (e.g., Asia-Pacific), Copilot may be disabled by default.
• In such cases, an admin must manually enable the feature from the Microsoft Fabric Admin Portal by activating the following option:

“Allow data sent to Azure OpenAI to be processed outside of the capacity's geographic region.”

This requirement exists because Copilot relies on Azure OpenAI, which may process data outside the tenant’s region.

⚠️ Important: FCC is not available in sovereign clouds such as Azure Government or Azure China 21Vianet, due to limited GPU availability and compliance restrictions.

2. An Active Paid License Is Required

Fabric Copilot Capacity (FCC) is only available in environments with a paid Microsoft Fabric capacity.

• Free or trial accounts are not supported.
• If you're using an evaluation version, you’ll need to upgrade to a paid subscription, even at the most basic tier (e.g., Pro, Trial, PPU, or F2).

3. Admin Configuration in the Fabric Admin Portal

FCC access can be managed from the Microsoft Fabric Admin Portal, under the tenant settings.

Admins have the ability to:

• Enable or disable Copilot organization-wide
• Restrict usage to specific user groups

This allows for granular control over how AI resources are distributed across the organization.

4. Required User Permissions

• To use Copilot in Power BI, users need at least the Contributor role or higher.
• Users with Read-only access can also view Copilot-generated responses, as long as the content resides in a compatible capacity.
• No special AI role is required — standard Power BI and Microsoft Fabric roles are sufficient.

5. Gradual Rollout and Wait Times

• For instance, if you upgrade your capacity from F1 to F2, FCC features may take up to 24 hours to become fully available in the portal.
• This is particularly relevant in Power BI, where Copilot integration depends on backend provisioning tied to capacity.

Summary: Steps to Activate Fabric Copilot Capacity (FCC)

1. Have an active paid capacity (F2 or higher)

2. Make sure your region supports Copilot

3. Review and adjust privacy settings in the admin portal

4. Assign the right user permissions

5. Allow up to 24 hours for changes in capacity to take effect

AI Capabilities Included with Fabric Copilot Capacity (FCC)

The AI features enabled through Fabric Copilot Capacity (FCC) span multiple workloads and are designed to accelerate the productivity of analysts, data scientists, data engineers, and business users — without requiring advanced programming skills.

Below, we explore the main AI capabilities included with Copilot in Microsoft Fabric.

1. Copilot in Power BI: Intelligence at the service of the business

Power BI Copilot is one of the most visible and impactful use cases of Fabric Copilot Capacity (FCC). It enables a wide range of AI-driven capabilities, including:

• Report generation from natural language prompts
  Example: “Create a sales report by product and region for the last quarter.”
• Automatic report page summaries
• Suggestions for titles, descriptions, and narratives for visualizations
• Enhanced Q&A experiences with synonyms and business context
• DAX measure generation from plain text instructions
  Example: “Create a measure to calculate year-over-year revenue change.”

These features democratize data analysis, empowering non-technical users to build clear, functional dashboards in just minutes.

2. Copilot in Data Engineering and Data Science: Intelligent Notebooks

• Autocompletion and code suggestions for Python, PySpark, or SQL
• Code generation from natural language instructions
• Automatic code explanations to improve understanding and learning
• Prebuilt templates for common data tasks

3. Copilot in Data Factory: ETL Process Automation

Copilot also simplifies the design and execution of data transformation workflows in Data Factory:

• Automatic generation of ETL steps
• Explanations of complex transformation logic
• Optimization suggestions to improve data flow performance

Example: “Extract the day from a timestamp column and filter records from the last month.”

4. Copilot in Data Warehouse: Conversational SQL

For SQL environments and tabular analysis, Copilot offers a natural and intuitive experience:

• Translates natural language into SQL queries
• Autocomplete and recommendations for complex SQL logic
• Query generation via voice or text prompts
• Automatic explanations of queries and results

Example: “Show total sales by region for the last quarter.”

5. Copilot in Real-Time Analytics: Query Streaming Data in Real Time

Copilot integrates with Kusto Query Language (KQL) to query streaming and telemetry data:

• Natural language to KQL query translation
• Assisted exploration of logs, telemetry, and time series
• Real-time data analysis without writing a single line of code

Ideal for operations, monitoring, and rapid response to live events.

6. Fabric Data Agents: Conversational Interface for Your Data

An advanced capability that complements FCC:

• AI-powered data agents connected to OneLake and internal data sources
• Ability to understand structure, meaning, and relationships within your datasets
• Provides context-aware answers to complex business questions

Example: “What were the three largest expenses this year and how do they compare to last year?”

These agents leverage Azure OpenAI and Azure AI Foundry, enabling natural and powerful interaction with enterprise data.

7. Frictionless Automation and Productivity

Beyond specific workloads, Fabric Copilot Capacity (FCC) empowers teams to:

• Automate complex tasks using natural language
• Generate alerts, triggers, and conditional logic with Data Activator
• Instantly create visualizations, reports, and KPIs on demand

The philosophy behind FCC is simple: you no longer need to code to analyze and understand your data—you just need to ask.

Fabric Copilot Capacity (FCC): Limitations, Warnings, and Key Considerations

The expansion of Fabric Copilot Capacity (FCC) marks a major advancement for the Microsoft Fabric ecosystem. However, before moving toward widespread adoption, it’s essential to understand its technical, regulatory, and operational limitations.

Regional Limitations and Regulatory Compliance

• Fabric Copilot relies on Azure OpenAI, which means that certain data—such as prompts, schema information, or metadata—may be processed outside your tenant’s geographic region.
• Microsoft has enabled Copilot by default in commercial regions such as the United States and the European Union.
• If your capacity is hosted in other regions (e.g., Latin America or Asia-Pacific), you must manually enable the setting that allows data processing outside the local region.
• FCC is not available in sovereign clouds such as Azure Government or Azure China 21Vianet, due to compliance constraints and limited GPU availability.

Recommendation: Before enabling FCC, review your organization's privacy, data protection and compliance policy.

Preview Status

All Copilot features in Microsoft Fabric are currently in public preview. This means:

• They are not covered by a Service Level Agreement (SLA)
• They may exhibit bugs or unstable behavior
• Features are subject to change before the official release

Important: While Copilot provides powerful assistance, it's essential to manually review and validate any generated code, summaries, or results before using them in production environments.

Supported Languages

• Copilot currently delivers its best performance in English.
• Support for other languages, including Spanish, is limited or still under development.

Recommendation: In multilingual environments, it is advisable to prioritize English usage to ensure greater accuracy and consistency in results.

Performance Impact and Capacity Usage

• Using Copilot consumes capacity units (CUs) within your active Microsoft Fabric subscription.
• While there is no additional cost during the public preview, Copilot does contribute to overall capacity consumption
• Heavy usage by multiple users can potentially impact the performance of critical workloads

Recommendation: Use Copilot Capacity to route AI workloads to a dedicated instance, minimizing interference with essential data processes.

Costs and Licensing

• Currently, Copilot has no additional cost — its usage is included within your existing Fabric capacity.
• Microsoft has indicated that future billing models may be introduced based on actual AI feature usageFor SKUs below.
• F64, users still require Power BI Pro licenses to collaborate in shared environments.

Tip: Make sure your capacity is properly sized to handle the anticipated demand for AI across your organization.

Exceptions: Power BI Embedded (A and EM SKUs)

• SKUs designed exclusively for external embedding—such as A1, A2, EM1, etc.—are not compatible with Copilot.
• Only F-type (Fabric) and P-type (Premium) capacities include access to Fabric Copilot Capacity (FCC).

Important: If your organization uses Power BI Embedded to serve external clients, you will need to purchase an additional F capacity if your internal teams wish to leverage Copilot functionalities.

Governance and Responsible Use

abric administrators have several governance tools to manage Copilot usage:

• Enable or disable Copilot at the tenant level
• Restrict access to specific users or groups
• Assign a dedicated Copilot Capacity to centralize AI workload execution

Microsoft enforces its Responsible AI policy, which ensures:

• Your data is not used to train AI models
• Prompts are not stored once the session ends
• Clear guidelines are provided on ethical AI use

Recommendation: Establish an internal policy outlining who can use FCC, how it should be used, and under what conditions — especially if your organization operates in a regulated industry.

Conclusion: Fabric Copilot Capacity (FCC) as a Catalyst for Democratizing AI in the Enterprise

The introduction of Fabric Copilot Capacity (FCC) marks a turning point in the accessibility of artificial intelligence within the Microsoft Fabric ecosystem. Thanks to this innovation, advanced capabilities like Copilot—previously reserved for high-tier capacities—are now democratized, enabling even organizations with modest setups such as F2 to leverage the power of generative AI.

Beyond its technical significance, FCC presents a strategic opportunity: to centralize Copilot usage, optimize resource consumption, and unlock new natural language–driven analytics scenarios—without requiring advanced programming skills. All of this comes at no additional cost, with a flexible configuration that adapts to the needs of each organization.

In today’s business landscape, where agility, automation, and intelligent analytics are key competitive advantages, enabling a Fabric Copilot Capacity (FCC) becomes a critical decision to boost productivity, reduce operational burden, and foster a more inclusive, data-driven culture.