Microsoft Fabric pricing explained

Everything you need to know

Microsoft Fabric brings data integration, data engineering, warehousing and business intelligence together in a single platform. This integrated approach also comes with a new pricing model, which can be confusing if you are used to individual Azure services. In this article, we explain the pricing structure of Microsoft Fabric, including capacity SKUs, payment options, storage costs, and licenses — so that you can make well-informed decisions.

The Microsoft Fabric Capacity Model

Unlike Azure, where each service has its own rate, Microsoft Fabric works with a capacity-based pricing model. You purchase a so-called Fabric Capacity (an F SKU) — a pool of computing power shared by all Fabric services, such as Data Factory, Spark Data Engineering, Data Warehouse, and Power BI.

Each capacity SKU has a specific number of Compute Units (CUs). An F2 has 2 CUs, an F4 has 4 CUs, and so on up to F2048. All workloads within Fabric draw from that same pool, meaning you have one compute bill instead of separate costs per service.

Pay-As-You-Go vs. Reserved Capacity

There are two ways to purchase Fabric capacity:

Pay-As-You-Go (PAYG) is a flexible option where you pay an hourly rate (billed per minute, with a minimum of one minute), but only as long as the capacity is active. You can scale up or down at any time, or pause the capacity when you do not need it.

Reserved capacity means that you commit to a fixed F SKU for one year in exchange for a significantly lower rate — about 40% cheaper than PAYG. The downside: you pay continuously, regardless of actual usage. Reserved capacity cannot be paused.

A handy rule of thumb: if you need the capacity more than ~60% of the time, a one-year reservation pays off. For sporadic or business-hours-only use, PAYG is often more cost-effective.

Overview of Fabric SKU Rates

Prices are indicative for a US region. From F64, individual Power BI viewer licenses are waived.

Additional Cost Components

Storage in OneLake

The F SKU only covers compute power. All data in Microsoft Fabric is stored in OneLake, priced similarly to Azure Data Lake Storage: approximately $0.023 per GB per month (~$23 per TB per month). With 10 TB, you reach about $230 per month in storage costs.

Optional features like KQL cache (~$0.246/GB/month) and backup/BCDR storage (~$0.041/GB/month) can be added on top.

Network Costs

Data transfer within the same Azure region is generally negligible. If you regularly export large datasets to other regions or on-premises systems, bandwidth costs ($0.09/GB) can quickly add up. With 100 GB of daily export, that can amount to around $270 per month.

Power BI Licenses

For capacities smaller than F64, users require a Power BI Pro license (~$10/user/month) or a Premium Per User license (~$20/user/month). With 50 report users on an F32, this quickly adds up to an extra $500 per month.

As soon as you use an F64 or larger, the costs for viewer licenses are waived. Calculate the tipping point: when is a larger SKU cheaper than paying for Pro licenses for your entire team?

Total Cost of Ownership: Two Scenarios Compared

Scenario A: PAYG (F32, 16 hours per day)

A company with peak load up to 32 CUs opts for a PAYG F32, active for 16 hours per day. They have 10 TB of data and 50 Power BI users.

Scenario B: Reserved (F16, 24/7)

The same company chooses a reserved F16 capacity, active 24/7 for one year.

Scenario B is clearly cheaper on an annual basis, but only provides 16 CUs instead of 32. A hybrid approach — a reserved baseline capacity supplemented with PAYG for peak loads — often offers the best balance.

Cost Optimization: Practical Tips

1. Adjust capacity based on actual usage

Do not pay for compute power you do not use. Monitor CU usage via the Fabric Capacity Metrics app and scale down in a timely manner. Consolidate workloads to get more out of a single capacity. Correct right-sizing typically saves 20-30% or more.

2. Manage peak load

Provision for average usage, not absolute peaks. Scale up temporarily during occasional peak loads. Working elastically can reduce monthly compute costs by nearly half.

3. Optimize licensing costs

Calculate the tipping point: when is upgrading to F64 cheaper than Pro licenses for all users? Also regularly check if all licenses are still actively being used.

4. Keep storage under control

Implement retention policies, avoid unnecessary data copies (use OneLake shortcuts), and design compact data models. Data grows automatically — a conscious policy prevents unexpected bills.

5. Minimize data transfer

Keep data within the same Azure region and process data as much as possible within OneLake itself. Compress and filter data that must be sent externally.

6. Optimize workloads

Efficient Spark notebooks, good SQL indexing, and smartly scheduled batch work directly reduce CU consumption — and thus costs.

How QUIKK Helps You Get the Most Out of Microsoft Fabric

Microsoft Fabric is a powerful platform, but its value lies in how you use it. QUIKK helps organizations set up, migrate, and further develop their data environment — from data warehouse to Power BI dashboard — so that you can focus on insights, not the technology.

Greenfield implementation or further development

Whether you are still working with separate Excel files or already have an existing data warehouse: QUIKK builds a Microsoft Fabric environment that fits your situation. We set everything up from scratch, or build further on what you already have — including new data sources, better performance, and a future-proof structure.

Migration from Qlik, Tableau, or other tools

Are you still using Qlik Sense, QlikView, Tableau, or SSRS? QUIKK migrates your existing reports to Power BI within Microsoft Fabric, without data loss and without months of downtime. Everything you already had, but scalable and ready for the future.

Power BI dashboards your team understands instantly

QUIKK develops Power BI dashboards that translate complex data into clear insights — directly usable for your team. From financial overviews to operational management: we deliver insights within 3 to 4 weeks. Already built more than 100 dashboards for various organizations.

Data Engineering: one reliable source

Data scattered across separate systems is one of the biggest cost drivers in a Fabric environment. QUIKK sets up a data warehouse that merges all sources into a single reliable source. This not only lowers your storage costs by avoiding duplication but also ensures faster queries and lower CU consumption.

Training: your team independently in control

Prefer to teach your own people to work with Power BI and Fabric? QUIKK offers practical training at every level, from basic user to advanced developer. Hands-on, using your own datasets, so you can immediately apply what you learn.

Conclusion

Microsoft Fabric offers a predictable, scalable, and flexible pricing model. But that flexibility also requires discipline: those who do not plan well will pay unnecessarily much. The key lies in understanding the capacity model, monitoring usage, and smartly combining reserved baseline capacity with PAYG for peaks.

With QUIKK as your partner, you ensure that your Fabric environment is not only technically sound but also delivers immediate value for your organization — from day one.