Microsoft Fabric has fundamentally changed how organisations consume analytics in the Microsoft ecosystem. But its licensing model — built around Capacity Units (CUs) and F-SKUs — can be confusing. This guide breaks it down.
Fabric is Microsoft’s unified analytics platform that brings together Data Factory, Synapse Data Engineering, Synapse Data Warehouse, Power BI, and Real-Time Intelligence under a single SaaS experience with a single billing model.
Everything in Fabric runs on Capacity Units. Unlike the old model where you paid separately for Power BI Premium, Synapse compute, and Data Factory pipelines, Fabric pools all compute into a shared CU bucket.
Key insight: 1 CU = approximately the compute of 2 Power BI v-cores. An F64 capacity gives you 64 CUs — enough for a mid-size enterprise running dashboards, data pipelines, and a lakehouse simultaneously.
| Feature | F-SKU (Fabric) | P-SKU (Legacy Premium) |
|---|---|---|
| Starting tier | F2 (trial), F64 (production) | P1 (8 v-cores) |
| Workloads included | All Fabric workloads | Power BI only |
| Autoscale | Yes (smoothing + burst) | No |
| Pay-as-you-go | Available via Azure | Not available |
| Future investment | Active development | Deprecated path |
Fabric offers two commercial models:
The most common mistake is over-provisioning. Start with an F64 for production and use Fabric’s built-in monitoring to track CU consumption across workloads. Key metrics to watch:
Sceniuz recommendation: Run a 30-day Fabric POC on PAYG before committing to reserved capacity. This gives you real consumption data to right-size.
Our team has helped multiple enterprises migrate from Power BI Premium (P-SKUs) to Fabric (F-SKUs), right-size their capacity, and implement FinOps practices for ongoing cost governance. We bring hands-on experience across industries including pharma, retail, and financial services.