🔄 Power BI Premium vs Fabric SKU Mapping
- gowheya
- Sep 12, 2025
- 2 min read
Updated: Sep 15, 2025

This article aim to provide a high level summary of your current Power BI Premium licences and map it with a corresponding Fabric SKU. It also highlights key migration benefits.
Current SKU | vCores / Capacity | Approx. Fabric Equivalent (F-SKU) | Notes / Considerations |
Premium Per User (PPU) | 1 user with Premium features | ⚠️ No direct equivalent – requires Fabric capacity (minimum F64) | PPU features (AI, paginated, large models) are included in Fabric, but per-user licensing is not supported. Migration requires moving to shared Fabric capacity. |
Premium P1 | 8 vCores | F64 | Closest baseline for migration. Offers comparable performance and feature set. Fabric allows elastic scaling, unlike fixed P1. |
Premium P2 | 16 vCores | F128 | Equivalent scale. Better concurrency and refresh throughput under Fabric. |
Premium P3 | 32 vCores | F256 | High-capacity workloads (large semantic models, high refresh). Fabric provides additional workloads (Data Warehouse, Real-Time Analytics). |
Premium EM SKUs (EM1/EM2/EM3) | Embedded workloads only | Fabric F-SKU required (starting F64) | Fabric unifies embedded and enterprise workloads—no EM equivalent. Use Fabric for embedding scenarios. |
🌟 Migration Benefits (SKU Perspective)
Elastic Scaling: Fabric SKUs (F-series) scale up/down easily; Premium was fixed.
Unified Workloads: One SKU covers BI, Lakehouse, Data Warehouse, Real-Time Analytics, AI.
Simplified Licensing: No split between Premium Capacity vs PPU.
Performance Gains: DirectLake storage mode bypasses dataset import limits for very large models.
⚠️ Migration Limitations / Gotchas
No PPU Equivalent: Users on PPU must be migrated to Fabric capacity, potentially increasing cost if only a few users need Premium features.
Capacity Learning Curve: Fabric admins must monitor and prioritize workloads (BI, Dataflows, AI jobs).
Minimum Entry Point: Fabric starts at F64, which is a bigger jump than individual PPU licenses.
Embedded Workloads: No direct EM SKU; embedded scenarios require Fabric F-capacity.
Refer to Microsoft documentation here: https://azure.microsoft.com/en-au/pricing/details/microsoft-fabric/



Comments