Microsoft Fabric Pricing Explained: F2, F4, F8, F64, and Capacity Sizing Guide
Microsoft Fabric is revolutionizing the way organizations handle data and analytics by providing a unified, AI-powered platform. However, navigating its pricing structure, capacity options, and governance implications can be a complex endeavor for many enterprises. Understanding how to correctly size your capacity and govern your environment is crucial for maximizing return on investment and avoiding unexpected costs. This comprehensive guide will demystify Microsoft Fabric pricing, explain the differences between capacity SKUs like F2, F4, F8, and F64, and explore the critical aspects of governance and monitoring.
Understanding Microsoft Fabric Pricing Models
Microsoft Fabric operates on a unified capacity-based pricing model, which simplifies purchasing by providing a single pool of compute for every workload. This means you purchase a specific amount of compute power, measured in Capacity Units (CUs), which is then shared across all Fabric workloads, including Data Engineering, Data Factory, Data Science, Data Warehouse, Real-Time Analytics, and Power BI .
There are two primary ways to purchase Microsoft Fabric capacity: Pay-As-You-Go and Reserved Capacity. The Pay-As-You-Go model offers flexibility, allowing you to scale up or down dynamically and only pay for what you use, billed by the minute. This is ideal for unpredictable workloads or short-term projects. On the other hand, committing to a one-year or three-year Reservation can yield significant savings—approximately 41% compared to the Pay-As-You-Go rates .
Pricing is based on US East region estimates and may vary. Storage costs are billed separately.
In addition to compute capacity, organizations must also account for OneLake storage costs. OneLake acts as the single, unified logical data lake for your entire organization. Storage is billed separately based on the amount of data stored and the tier (Hot, Cool, or Cold), starting at $0.023 per GB per month for Hot storage . Interestingly, Fabric offers free Mirroring storage for replicas up to a certain limit based on the purchased compute capacity SKU. For instance, purchasing an F64 capacity grants 64 terabytes of free Mirroring storage .
Microsoft Fabric Licensing Logic: Understanding the Foundation
Before diving into capacity sizing, it is essential to understand how Microsoft Fabric licensing works. The licensing model is fundamentally different from traditional per-user licensing and directly impacts your total cost of ownership.
Capacity vs. Per-User Licensing
Microsoft Fabric operates on a capacity-based licensing model combined with per-user licenses. The capacity license (F SKU) provides the compute resources, while per-user licenses determine who can create, edit, and view content .
Capacity Licenses (F SKUs): These are the foundation of your Fabric deployment. The F SKU you purchase determines the total compute power available to your organization. All Fabric workloads—whether it is data engineering, data warehousing, Power BI analytics, or real-time analytics—share this single pool of compute resources.
Per-User Licenses: Separate from capacity, individual users require per-user licenses to interact with Fabric. The license types include :
•Fabric Free License: Automatically assigned to all users on first sign-in. Allows creation and sharing of non-Power BI Fabric items (lakehouses, warehouses, notebooks) when the workspace runs on an F or Trial capacity. However, viewing Power BI content requires additional licensing on capacities smaller than F64.
•Power BI Pro License: Required for users who want to create, edit, or share Power BI items (reports, dashboards, semantic models) in workspaces other than their personal workspace.
•Power BI Premium Per-User (PPU) License: Provides most Power BI Premium features on a per-user basis, including advanced analytics, higher refresh rates (48 refreshes per day), and larger semantic models (>1GB).
Why Buy Fabric Capacity?
Organizations purchase Fabric capacity for several strategic reasons:
1. Unified Compute Pool: Unlike traditional Power BI Premium where you purchase per-capacity licenses, Fabric capacity provides a single, shared pool of compute that serves all workloads. This means your data engineering pipelines, data warehousing queries, and Power BI reports all draw from the same capacity, enabling better resource utilization.
2. Free Viewer Access at F64+: One of the most compelling reasons to upgrade to F64 or larger is the ability for users with a Fabric Free license to view Power BI content without requiring individual Pro licenses . For organizations with hundreds or thousands of report consumers, this can result in dramatic cost savings. For example, an organization with 500 report viewers would save approximately $7,000 per month by upgrading from F32 to F64 (the cost of 500 Pro licenses at $14/user/month).
3. Embedded Analytics: Fabric capacity enables Power BI Embedded scenarios, allowing you to embed analytics into customer-facing applications without requiring end-user licenses. This is particularly valuable for ISVs and SaaS providers.
4. Unified Data Lake: Fabric capacity provides access to OneLake, a single, unified data lake for your entire organization. This eliminates data silos and enables seamless integration across all Fabric workloads.
Pay-As-You-Go vs. Reserved Capacity: When to Choose Each
The decision between Pay-As-You-Go and Reserved Capacity is not merely about cost—it depends on your workload predictability and commitment level.
Pay-As-You-Go: Flexibility and Unpredictability
Best for: Organizations with unpredictable workloads, short-term projects, or those still evaluating Fabric.
•No long-term commitment required
•Scale up or down minute-by-minute based on actual usage
•Ideal for testing and proof-of-concept (POC) scenarios
•Pay only for what you use
•Higher per-unit cost (baseline pricing)
•Difficult to predict monthly expenses
•Not ideal for mission-critical, always-on workloads
Example: A consulting firm running ad-hoc analytics projects for different clients would benefit from Pay-As-You-Go, as capacity needs vary significantly by project.
Reserved Capacity: Cost Savings and Predictability
Best for: Organizations with predictable, sustained workloads and long-term Fabric commitments.
•~41% cost savings compared to Pay-As-You-Go
•Predictable monthly expenses for budgeting
•Ideal for production analytics environments
•Better ROI for sustained usage
•Requires 1 or 3-year commitment
•Less flexibility if workloads decrease
•Upfront financial commitment
Example: An enterprise with a production data warehouse and established Power BI reporting infrastructure should commit to Reserved Capacity, as their compute needs are stable and predictable.
Hybrid Approach: Combining Both Models
Many organizations use a hybrid strategy: purchase a Reserved Capacity for baseline, predictable workloads, then use Pay-As-You-Go for burst capacity during peak periods. This balances cost optimization with flexibility. Our Workspace & Capacity Management and Optimization services help organizations design this optimal mix.
Capacity Sizing Guide: Choosing Between F2, F4, F8, and F64
Selecting the right capacity SKU is perhaps the most critical decision when deploying Microsoft Fabric. Over-provisioning leads to wasted resources, while under-provisioning can result in throttling and poor performance. The capacity size dictates the compute power available for processing queries, running data pipelines, and rendering Power BI reports.
Entry-Level Capacities: F2, F4, and F8
The F2, F4, and F8 SKUs are considered entry-level capacities. They are highly suitable for small teams, proof-of-concept (POC) projects, or organizations just beginning their journey with Microsoft Fabric.
An F2 capacity provides 2 CUs, which translates to a specific amount of compute power evaluated every 30 seconds . While cost-effective, these smaller capacities have limitations. They are best utilized for lightweight data processing and reporting. If your organization relies heavily on complex DAX calculations or processes massive datasets, these SKUs might experience throttling during peak usage times. It is important to note that users interacting with content on these smaller capacities still require individual Power BI Pro licenses .
Enterprise Capacities: F64 and Beyond
The F64 capacity is a significant milestone in the Microsoft Fabric ecosystem. Providing 64 CUs, it is designed for enterprise-level workloads, robust data engineering pipelines, and large-scale reporting.
A crucial distinction of the F64 SKU (and larger) is its relationship with Power BI licensing. When you provision an F64 capacity, it unlocks the ability for users with a free Power BI license to view and interact with Power BI content hosted on that capacity . This is analogous to the traditional Power BI Premium P1 capacity. For organizations with a large base of report consumers, upgrading to an F64 capacity often becomes more cost-effective than purchasing individual Pro licenses for hundreds or thousands of users.
Power BI Embedded and Fabric
For organizations looking to embed analytics into their own applications or portals for external customers, Microsoft Fabric provides a compelling solution. The Fabric F SKUs can be utilized for Power BI Embedded scenarios (App Owns Data) . This allows you to deliver rich, interactive data visualizations to your customers without requiring them to have individual Power BI licenses.
When building customer-facing applications, leveraging a Fabric capacity for embedding can be a highly strategic move. If you are developing specialized portals or require advanced integration, platforms like Binexus.io can serve as excellent references for implementing scalable embedded analytics solutions. Our Fabric Migration & Modernization services can guide you through this process.
How to Estimate the Right Fabric Capacity: A Practical Framework
Capacity estimation is both an art and a science. Rather than guessing, organizations should evaluate several key factors that directly impact compute consumption. The Microsoft Fabric SKU Estimator provides a starting point, but understanding the underlying factors is crucial for accurate planning.
Factor 1: Number of Users and Concurrent Users
The number of users accessing your Fabric environment impacts capacity consumption, but not linearly. What matters most is concurrent users—how many users are simultaneously running queries, refreshing data, or interacting with reports.
Estimation Rule: Each concurrent interactive user typically consumes 0.5-2 CUs per query, depending on query complexity. Background operations (refreshes, ETL) consume significantly more.
Example: An organization with 100 total users but only 10 concurrent users during peak hours requires less capacity than one with 50 concurrent users, despite having fewer total users.
Factor 2: Refresh Frequency and Data Volume
Data refresh operations are among the most compute-intensive activities in Fabric. The frequency and volume of refreshes directly impact capacity requirements.
•Lightweight Refreshes (< 1 GB data, < 5 minutes): Consume approximately 50-200 CUs per refresh
•Medium Refreshes (1-10 GB data, 5-30 minutes): Consume approximately 500-2,000 CUs per refresh
•Heavy Refreshes (> 10 GB data, > 30 minutes): Consume 2,000+ CUs per refresh
•Organization runs 10 daily refreshes, each consuming 1,000 CUs
•Each refresh takes 10 minutes
•Daily CU consumption from refreshes: 10 × 1,000 = 10,000 CUs
•Distributed over 24 hours: 10,000 ÷ 1,440 minutes = ~7 CUs per minute baseline
Factor 3: ETL Workload and Data Engineering
ETL (Extract, Transform, Load) operations using Spark, Data Factory pipelines, and dataflows consume significant compute resources. The complexity of transformations and data volume directly impact consumption.
•Simple ETL (basic transformations, < 1 GB): 100-500 CUs per run
•Complex ETL (multiple joins, aggregations, 1-10 GB): 1,000-5,000 CUs per run
•Heavy ETL (complex ML transformations, > 10 GB): 5,000+ CUs per run
Example: An organization running three daily ETL pipelines:
•Pipeline 1 (simple): 300 CUs × 3 runs = 900 CUs/day
•Pipeline 2 (complex): 2,000 CUs × 2 runs = 4,000 CUs/day
•Pipeline 3 (heavy): 8,000 CUs × 1 run = 8,000 CUs/day
•Total daily ETL consumption: 12,900 CUs
Factor 4: Direct Lake Usage and Query Patterns
Direct Lake is a powerful storage mode that loads data directly from Delta tables in OneLake without requiring a separate import. However, Direct Lake has specific capacity implications .
•Queries use the VertiPaq engine, delivering Import-like performance
•Refresh operations are metadata-only (seconds instead of minutes)
•Minimizes data duplication and storage costs
•Direct Lake queries consume similar CUs to Import queries (efficient)
•Refresh operations consume minimal CUs (metadata-only)
•However, the underlying Delta table must fit in memory for optimal performance
•If data exceeds capacity memory limits, Direct Lake automatically falls back to DirectQuery mode, consuming more CUs
Estimation Rule: Direct Lake is most efficient when:
•Your Delta tables are < 50% of capacity memory (e.g., < 32 GB for F64)
•You have frequent queries but infrequent data updates
•You want to avoid managing separate Import refreshes
Example: An organization with a 100 GB Delta table should use at least F128 capacity (256 GB memory) to ensure Direct Lake operates efficiently without fallback to DirectQuery.
Putting It Together: A Complete Capacity Estimation Example
Scenario: Mid-sized financial services organization
•100 total users, 15 concurrent during peak hours
•5 daily semantic model refreshes (average 2 GB each, 15 minutes each)
•3 daily ETL pipelines (complex transformations, 5-10 GB each)
•Direct Lake semantic model (50 GB, frequently queried)
•Peak usage: 9 AM - 5 PM (8 hours/day)
•Interactive Queries: 15 concurrent users × 1.5 CUs per query = 22.5 CUs baseline
•Semantic Model Refreshes: 5 refreshes × 1,500 CUs = 7,500 CUs/day ÷ 1,440 minutes = 5.2 CUs/minute
•ETL Pipelines: 3 pipelines × 3,000 CUs = 9,000 CUs/day ÷ 1,440 minutes = 6.25 CUs/minute
•Peak Hour Consumption: 22.5 + 5.2 + 6.25 = ~34 CUs/minute during peak
•Off-Peak Consumption: ~10 CUs/minute (background operations only)
Recommendation: F64 capacity (64 CUs) provides comfortable headroom for this scenario. The organization could start with F32 (32 CUs) and monitor using the Capacity Metrics app, then upscale to F64 if throttling occurs.
•F64 Reserved (1-year): $5,002.67/month
•Alternative: 100 Power BI Pro licenses at $14/user/month = $1,400/month
•Total with F64: $5,002.67/month (includes all Fabric workloads + free viewer access)
•Total with F32 + Pro licenses: $2,501.33 + $1,400 = $3,901.33/month
In this case, F64 is justified by the ability to support all workloads plus free viewer access.
Strategic Capacity Planning and Upscaling
A key feature of Microsoft Fabric is its ability to handle sudden spikes in compute demand. Fabric employs a smoothing mechanism that averages compute usage over time, helping to prevent immediate throttling during short bursts of activity . However, for sustained high usage, organizations must employ a planned upscale strategy. By continuously monitoring the Capacity Metrics app, administrators can identify when utilization consistently approaches limits and proactively scale up the capacity (e.g., from F32 to F64) to prevent performance degradation and throttling . This dynamic scalability is a core advantage of the cloud-native Fabric architecture. Our Workspace & Capacity Management and Optimization services specialize in this strategic planning.
Automating Capacity Scaling with Azure Logic Apps
One of the most powerful strategies for managing Fabric capacity costs and performance is automating the scaling process. Rather than manually monitoring and adjusting capacity SKUs, organizations can leverage Azure Logic Apps to create intelligent workflows that automatically scale capacity up or down based on time-of-day, usage patterns, or other triggers.
Azure Logic Apps enables you to define recurrence-based rules that automatically upscale your Fabric capacity during peak business hours and downscale during off-peak times. For example, you can configure a Logic App to scale from F2 to F8 between 8 AM and 12 PM on weekdays, then automatically scale back down to F2 during evenings and weekends . This approach allows organizations to start with a minimal, cost-effective capacity and dynamically expand compute resources exactly when needed, without manual intervention.
The automation workflow typically involves three key components: a recurrence trigger that runs on a defined schedule, a condition action that evaluates the current time or usage metrics, and Azure Resource Manager actions that invoke the Fabric capacity scaling APIs . By using the Azure Resource Manager connector in Logic Apps, you can programmatically change the SKU of your Fabric capacity, enabling true "always upscale" scenarios where capacity grows with demand.
This approach is particularly valuable for organizations with predictable workload patterns. For instance, a retail organization might scale up capacity before running end-of-day reporting processes, then scale down afterward. A financial services firm might maintain higher capacity during market hours and reduce it after trading closes. The flexibility of Logic Apps automation means you can tailor scaling behavior to match your organization's unique operational rhythm.
One important consideration: the capacity must be running (not paused) for the scaling operation to succeed. Organizations using the PATCH API method (rather than PUT) can scale even when capacity is paused, providing additional flexibility for cost optimization . When implementing automation, it is recommended to use service principals or managed identities for authentication rather than user credentials, ensuring more secure and reliable long-term operation of your scaling workflows. Our Pipeline & Dataflow Automation expertise can help you design and implement these sophisticated automation scenarios.
Governance Implications in Microsoft Fabric
With great power comes great responsibility. The unified nature of Microsoft Fabric means that robust governance is no longer optional; it is a necessity. Proper governance ensures data security, compliance, and discoverability across the enterprise.
Centralized Administration and Domains
Microsoft Fabric provides a centralized Admin Portal where tenant-level settings are configured. However, recognizing that a one-size-fits-all approach rarely works for large enterprises, Fabric introduces the concept of Domains .
Domains allow organizations to logically group data and workspaces by business unit, department, or project. This federated governance model enables central IT to define overarching policies while delegating specific administrative controls to domain owners. For example, the Finance department can manage its own workspaces and data access policies within the Finance Domain, ensuring compliance with strict financial regulations, while Marketing operates within its own distinct Domain . This is a core component of our Power BI Consulting & Governance services.
Security, Purview Integration, and Data Loss Prevention
Securing data at the source is paramount. Microsoft Fabric integrates deeply with Microsoft Purview to provide advanced security and compliance capabilities.
Through Purview Information Protection, organizations can apply sensitivity labels to Fabric items, ensuring that data remains protected even when exported . Furthermore, Data Loss Prevention (DLP) policies can be configured to automatically detect sensitive information (such as credit card numbers or personal identification) uploaded into Fabric, triggering alerts or blocking access to prevent unauthorized sharing .
Monitoring and the Capacity Metrics App
You cannot manage what you cannot measure. Continuous monitoring is essential for optimizing performance, managing costs, and ensuring the health of your Microsoft Fabric environment.
The cornerstone of Fabric monitoring is the Microsoft Fabric Capacity Metrics app. This powerful, pre-built Power BI application provides granular visibility into how your capacities are being utilized .
Key Features of the Capacity Metrics App
•Health Page: Provides a high-level overview of all capacities, immediately highlighting any capacities experiencing high utilization, throttling, or query rejections .
•Compute Page: Offers a detailed 14-day trend of compute performance, breaking down utilization by operation type (Interactive vs. Background) and identifying specific items consuming the most resources .
•Timepoint Analysis: Allows administrators to drill down into specific 30-second intervals to diagnose exactly which operations caused a spike in compute usage, enabling precise performance tuning .
Implementing Governance at Scale
As your Fabric environment grows through automated capacity scaling, governance becomes increasingly critical. With multiple teams potentially sharing capacity resources, clear policies around data access, cost allocation, and performance SLAs must be in place. This is where Momentum One's expertise becomes invaluable.
Our Power BI Consulting & Governance services help organizations establish governance frameworks that scale alongside their capacity. We can assist in designing domain architectures that isolate costs by business unit, implementing chargeback models using the Fabric Chargeback app, and establishing monitoring dashboards that provide real-time visibility into capacity utilization and costs. By combining automated scaling with robust governance, you create a self-optimizing analytics platform that grows efficiently with your organization.
Assess Your Fabric Readiness: The Momentum One Assessment Tool
While the capacity estimation framework provided above is comprehensive, every organization's situation is unique. Variables such as existing infrastructure, data quality, team expertise, and business requirements create a complex landscape that benefits from professional assessment.
Momentum One offers a Fabric Readiness Assessment tool specifically designed to evaluate your organization's preparedness for Microsoft Fabric adoption and provide accurate capacity estimates tailored to your specific environment .
What the Fabric Readiness Assessment Covers
1. Current State Analysis:
•Inventory of existing Power BI, SQL Server, and data warehouse infrastructure
•Assessment of data quality, governance maturity, and security posture
•Identification of technical debt and modernization opportunities
•Evaluation of team skills and training needs
2. Fabric Readiness Scoring:
•Organizational readiness across technical, governance, and operational dimensions
•Identification of gaps and risks
•Prioritized recommendations for addressing readiness gaps
•Timeline and effort estimates for addressing each gap
•Detailed analysis of your specific workloads (Power BI, Data Engineering, Data Warehouse, Real-Time Analytics)
•Accurate CU consumption projections based on your actual data volumes, refresh frequencies, and user patterns
•Scenario planning: what-if analysis for different growth scenarios
•Cost projections for Pay-As-You-Go vs. Reserved Capacity options
•ROI analysis comparing your current infrastructure costs to projected Fabric costs
•Phased migration strategy aligned with your business priorities
•Risk mitigation strategies
•Resource requirements and timeline
•Success metrics and KPIs
Why Professional Assessment Matters
While the estimation framework in this article provides a solid foundation, professional assessment offers several advantages:
Accuracy: Generic estimation frameworks assume standard workload patterns. Your organization likely has unique characteristics—complex ETL pipelines, specialized data science workloads, or unusual refresh patterns—that impact capacity requirements. A professional assessment captures these nuances.
Risk Mitigation: Underestimating capacity can lead to throttling and poor user experience. Overestimating wastes budget. Professional assessment helps you find the optimal balance.
Holistic Planning: Capacity is only one piece of the puzzle. A comprehensive readiness assessment identifies governance gaps, security requirements, and organizational readiness issues that could derail a Fabric implementation.
Cost Optimization: By understanding your true capacity needs and usage patterns, you can make informed decisions about Reserved vs. Pay-As-You-Go capacity, automation strategies, and workload optimization opportunities that can save thousands of dollars annually.
Competitive Advantage: Organizations that accurately assess their Fabric readiness and plan migrations strategically gain faster time-to-value and avoid costly missteps.
Getting Started with Your Assessment
1.Initial Questionnaire: Complete a detailed questionnaire about your current infrastructure, workloads, and business requirements (approximately 30-45 minutes)
2.Data Collection: Our team reviews your responses and may request additional information about specific workloads
3.Analysis: Momentum One's Fabric experts analyze your data and create customized recommendations
4.Readiness Report: Receive a comprehensive report with capacity estimates, readiness scores, migration roadmap, and actionable recommendations
5.Consultation: Schedule a consultation with our team to discuss findings and next steps
The assessment is designed to be thorough yet practical, providing insights you can act on immediately.
Accelerate Your Fabric Journey with Momentum One
Navigating the complexities of Microsoft Fabric pricing, sizing, and governance requires specialized expertise. A misstep in capacity planning or a gap in your governance strategy can lead to significant financial and operational setbacks.
At Momentum One, we are dedicated to helping organizations build scalable, secure, and high-performance analytics environments. Our comprehensive suite of services is designed to guide you through every stage of your data journey:
•Data Modeling & DAX Optimization: High compute consumption is often the result of inefficient data models. We optimize your data models and DAX calculations to ensure peak performance, reducing the load on your Fabric capacity and saving you money.
•Pipeline & Dataflow Automation: We design and implement sophisticated automation workflows that optimize your data ingestion and processing pipelines.
Ready to transform your data strategy with Microsoft Fabric? Start with the Fabric Readiness Assessment to understand your organization's readiness and get an accurate capacity estimate tailored to your specific environment. Then, visit our Services Page to learn how Momentum One can help you optimize your capacity, enforce robust governance, and maximize your analytics ROI. Contact us today to schedule a consultation.
References
Microsoft Fabric
17.06.2026