Microsoft Fabric Licensing: Capacity F SKU Pricing Guide 2026

What Is Microsoft Fabric and How Is It Licensed?

Microsoft Fabric is Microsoft's unified analytics platform, launched in 2023 and positioned as the successor to Azure Synapse Analytics and the natural evolution of Power BI Premium. Fabric combines data engineering (data pipelines, Apache Spark-based Lakehouse), data science (notebooks, MLflow), real-time analytics (KQL database, Eventstream), data warehousing (Warehouse), and Power BI business intelligence in a single SaaS-based platform hosted on Azure.

Fabric is licensed through capacity — a pool of Capacity Units (CUs) that powers all Fabric workloads within a defined capacity. The capacity is represented by the F SKU (F for Fabric), which replaces the legacy P SKU (Power BI Premium Per Capacity). Organisations purchase a Fabric capacity of a defined size (F2, F4, F8, F16, F32, F64, F128, F256, F512, F1024, or F2048), and all Fabric workloads run within that capacity's CU budget. Individual users who build or consume Fabric content are assigned to a workspace attached to the capacity.

The F SKU Pricing Tiers: F2 to F2048

The F SKU pricing range is more granular than the legacy P SKU, starting at very small capacities (F2) and scaling to enterprise-grade deployments (F2048). Pricing below reflects pay-as-you-go rates; reserved pricing delivers 30 to 40 percent savings.

F2 (2 CUs): approximately $263 per month pay-as-you-go. F4 (4 CUs): approximately $526 per month. F8 (8 CUs): approximately $1,053 per month. F16 (16 CUs): approximately $2,106 per month. F32 (32 CUs): approximately $4,212 per month. F64 (64 CUs): approximately $8,423 per month. F128 (128 CUs): approximately $16,846 per month. F256 (256 CUs): approximately $33,700 per month. F512 (512 CUs): approximately $67,400 per month. F1024 (1,024 CUs): approximately $134,800 per month. F2048 (2,048 CUs): approximately $269,600 per month.

The commercially relevant range for most enterprise Fabric deployments sits between F32 and F256. F32 to F64 covers mid-size BI deployments with moderate data engineering workloads. F64 to F128 covers large enterprise BI environments with active Lakehouse or Warehouse workloads. F256 and above addresses organisations with high-concurrency analytics requirements or significant real-time analytics workloads running alongside BI.

The Critical F64 Threshold

The most commercially important threshold in the F SKU range is F64. Below F64, every user who views published Power BI content embedded in the Fabric capacity requires an individual Power BI Pro licence at approximately $10 per user per month. At F64 and above, viewer access to published Power BI content is included in the capacity licence — individual Pro licences for viewers are not required.

This threshold fundamentally changes the TCO calculation for organisations with large numbers of Power BI report consumers. An organisation with 1,000 Power BI viewers currently paying $10 per user per month for Pro licences spends $10,000 per month on viewer licences alone. Upgrading to an F64 capacity at approximately $8,423 per month eliminates the per-user viewer cost and reduces total spend — while simultaneously enabling the full Fabric platform. The F64 break-even analysis is one of the most straightforward ROI calculations in the Microsoft licensing landscape, and it frequently justifies the F64 investment without needing to attribute value to any Fabric workloads beyond Power BI.

Feature Availability by F SKU Tier

Not all Fabric capabilities are available at all F SKU tiers. The F64 threshold matters not just for Power BI viewer access but also for a set of platform features that are gated below F64.

Below F64 (F2 through F32), the following capabilities are not available: unlimited viewer access to published Power BI content, XMLA endpoint read and write access (required for external tools like SQL Server Management Studio, Tabular Editor, and ALM Toolkit), paginated reports deployment, Power BI deployment pipelines, and the Lakehouse and Data Warehouse features at scale.

At F64 and above, all standard Power BI Premium capabilities are available, including XMLA endpoint (critical for enterprise BI development workflows), paginated reports, deployment pipelines for CI/CD, semantic model scale-out, and the full Fabric platform capability set. F64 is effectively the minimum viable Fabric tier for enterprise deployments — F2 through F32 are appropriate for development, testing, or limited pilot workloads but should not be treated as production enterprise analytics infrastructure.

F SKU vs P SKU: The Legacy Migration Decision

Many organisations that adopted Power BI Premium before Fabric's launch hold P SKU (Power BI Premium Per Capacity) licences. The P SKU range (P1, P2, P3, P4, P5) is the legacy Power BI Premium per-capacity offering. Microsoft has indicated that P SKUs will be retired in favour of F SKUs, though the transition timeline has not been definitively fixed. Organisations on P SKUs face a migration decision that requires commercial analysis.

P1 (8 CUs) maps approximately to F64 in terms of CU count and capability. P2 (16 CUs) maps to F128. P3 (32 CUs) maps to F256. The F SKU equivalents are generally priced at parity or slightly below P SKU list pricing when compared on a CU basis, and F SKUs are available on pay-as-you-go through Azure, which provides more flexibility than the annual EA-committed P SKU pricing.

The migration from P SKU to F SKU requires moving workspaces from Premium capacity to Fabric capacity and migrating any P SKU-specific configurations. Most organisations complete this migration with minimal disruption, but the migration should be planned and tested in a non-production environment before production cutover. From a commercial perspective, P SKU renewals should be delayed until F SKU pricing has been validated against current P SKU commitments — P SKU renewals at list pricing in 2026 are unlikely to be the most cost-effective option.

Fabric Capacity Within the Microsoft EA

Microsoft Fabric capacity can be procured through the Microsoft Enterprise Agreement or through direct Azure consumption (pay-as-you-go or Reserved Capacity). The EA route — typically through the Server and Cloud Enrollment (SCE) — provides the ability to negotiate Fabric capacity pricing as part of a broader Azure and Microsoft 365 deal.

Reserved Capacity: The 30–40% Savings Opportunity

Reserved Capacity commitments for Fabric (equivalent to Azure Reserved Instances for compute) lock in a defined CU tier for one or three years in exchange for 30 to 40 percent savings versus pay-as-you-go pricing. An F64 at pay-as-you-go costs approximately $8,423 per month. A one-year Reserved Capacity commitment brings this down to approximately $5,900 per month. A three-year commitment reduces it further to approximately $5,000 per month. Over a three-year period, the Reserved Capacity savings on a single F64 represent approximately $120,000 in avoided spend compared to pay-as-you-go.

Reserved Capacity is appropriate for stable, production workloads where the F SKU tier is predictable over the commitment period. For development or variable workloads, pay-as-you-go preserves the flexibility to pause capacity (and stop billing) during non-business hours, which can reduce effective monthly spend by 30 to 50 percent for workloads that do not run 24/7.

EA Negotiation for Fabric

When Fabric is included in an EA negotiation alongside M365, the combined contract value creates leverage for additional Fabric-specific discounts. Organisations negotiating an EA renewal that includes a Fabric Reserved Capacity commitment should present the Fabric spend as part of the total Azure MACC rather than as a standalone line item. Microsoft's field teams are more incentivised to discount when Fabric capacity is framed as a component of an Azure growth commitment rather than an isolated Fabric purchase. Current EA negotiated rates for Fabric Reserved Capacity typically achieve 20 to 40 percent below list price, depending on the size of the overall Azure commitment and the quality of the commercial relationship.

Capacity Autoscale and Cost Control

Microsoft Fabric supports a capacity autoscale feature that allows the capacity to temporarily burst above its committed CU tier to handle workload spikes. Autoscale is charged at a consumption rate above the committed capacity level. For production environments with variable workload patterns, autoscale prevents user-facing capacity throttling during peak periods — but it can generate significant unexpected cost if autoscale consumption is not monitored and governed.

Best practice for Fabric capacity management includes setting an autoscale budget cap in Azure Cost Management, monitoring capacity utilisation metrics in the Fabric Admin portal to identify workloads that consistently exceed capacity, sizing the base capacity tier to handle 80 to 90 percent of workload volume without autoscale, and using workload-level throttling policies to prevent low-priority workloads (e.g., background data pipelines) from consuming capacity headroom during peak BI usage periods.