Microsoft Fabric Pricing UK: What You Need to Know (2026)

Microsoft Fabric pricing in the UK starts with capacity rather than user licences. In the September 2026 UK South pricing reference used in this guide, an F2 works out at about £225.72 per month on pay-as-you-go, while an F64 works out at about £7,222.91 when running for 730 hours.

Those figures are only the starting point. Your final Fabric bill also depends on the Azure region, active usage, purchasing option, OneLake storage, Power BI licensing, and how efficiently the capacity is sized.

Microsoft Fabric uses a capacity-based model measured in Capacity Units (CUs). The same shared capacity can support data engineering, data warehousing, analytics, real-time intelligence and Power BI workloads.

This guide explains Microsoft Fabric pricing UK businesses need to consider in 2026, including F2–F256 capacity costs, F64 pricing, pay-as-you-go versus reservations, OneLake storage, capacity overage, Spark autoscale billing and Power BI licensing.

Key Takeaways

• Microsoft Fabric uses a shared capacity model measured in Capacity Units.

• Fabric capacity SKUs range from F2 to F8192.

• The UK South PAYG pricing reference used in this article is approximately £0.1546 per CU-hour.

• An F2 running for 730 hours works out at approximately £225.72 per month.

• An F64 running for 730 hours works out at approximately £7,222.91 per month.

• Microsoft advertises up to 40.5% savings over standard PAYG pricing for eligible Fabric reservations.

• OneLake storage has its own pricing model and should be budgeted separately from Fabric capacity compute.

• F64 or larger capacities have important Power BI consumption licensing implications.

• Fabric capacity overage can help prevent throttling during periods of excess consumption.

• Spark workloads also have an optional autoscale billing model for dynamic or unpredictable workloads.

• Capacity should be selected according to workload demand and concurrency rather than user count alone.

How Microsoft Fabric Pricing Works

Microsoft Fabric uses a shared pool of compute capacity rather than separate compute for every workload.

A single capacity can support multiple Fabric experiences, including:

• Data Factory

• Data Engineering

• Data Science

• Data Warehouse

• Real-Time Intelligence

• Power BI

Capacity is measured in Capacity Units (CUs). An F2 provides 2 CUs, F64 provides 64 CUs and F128 provides 128 CUs.

The SKU determines the compute capacity available to your workloads. It does not, by itself, determine your total monthly bill.

The amount of time the capacity is active, the Azure region, purchasing option, storage requirements and workload demand all affect the overall cost.

Microsoft's current pricing page describes Fabric capacity as a shared pool and states that capacity has a one-minute minimum. It also allows customers to scale capacity up or down and pause it when it is not needed.

Microsoft Fabric Capacity Pricing in the UK

The following table uses the UK South PAYG reference rate of approximately £0.1546 per CU-hour used for this 2026 pricing guide.

The monthly figures are calculated using 730 active hours per month. They are therefore estimates for an always-running capacity rather than a separate fixed monthly Microsoft list price.

Pricing note: The £0.1546/CU-hour figure is a point-in-time UK South pricing reference used for this article. Microsoft states that Azure prices are estimates and that actual pricing can vary according to the Microsoft agreement, purchase date and currency exchange rate. Microsoft also converts displayed prices using applicable exchange rates. Always check the current Microsoft pricing page before making a procurement decision.

What does F64 cost per hour?

Using the UK South reference rate:

64 CUs × £0.1546 = £9.8944 per hour

Rounded to two decimal places, that is approximately:

£9.89 per hour

For a continuously running 24-hour period:

£9.8944 × 24 = £237.47

And for a 730-hour month:

£9.8944 × 730 = £7,222.91

This makes the difference between smaller development capacities and an always-on production F64 significant.

UK South vs UK West

Microsoft Fabric pricing varies by Azure region, so UK organisations should select the appropriate region before using a figure for procurement.

The Microsoft pricing page provides region and currency selectors, allowing customers to review the applicable pricing for UK South or UK West in GBP.

The UK South figures above are therefore best treated as a regional reference rather than a universal UK price.

If your Fabric environment is being deployed in UK West, check the live Microsoft pricing table with:

Region: UK West
Currency: GBP (£)

Do not assume that the UK South figure is automatically the UK West price.

Microsoft specifically notes that pricing can vary by region and that the live pricing page should be used for current pricing.

Microsoft Fabric Capacity SKUs Explained

Microsoft currently offers Fabric capacity from F2 through F8192.

The smaller SKUs are useful for lighter workloads, development environments and organisations starting with Fabric. Larger capacities provide more compute for heavier processing and concurrent workloads.

The important point is that capacity should follow actual consumption.

An organisation with ten users running occasional reports could require less compute than a ten-user environment running frequent data pipelines, Spark jobs, warehouse queries and semantic model refreshes at the same time.

F2, F4 and F8: Smaller Fabric Capacities

F2, F4 and F8 sit at the lower end of the Fabric capacity range.

Based on the UK South reference rate used in this article:

• F2: approximately £225.72/month at 730 active hours

• F4: approximately £451.43/month

• F8: approximately £902.86/month

These capacities can make sense for organisations that are testing Fabric, developing solutions or running relatively light workloads.

The advantage of starting smaller is not simply lower cost. It also gives the organisation an opportunity to observe actual Fabric consumption before committing to a larger capacity.

If monitoring shows that the capacity is regularly approaching its limits, the organisation has evidence for scaling up.

F64 Capacity and Power BI

F64 provides 64 Capacity Units and is an important threshold for organisations using Fabric with Power BI.

The reason is partly technical and partly licensing-related.

Microsoft's current licensing guidance states that users with a Microsoft Fabric Free licence can view Power BI content when the capacity is F64 or larger and the user has the Viewer role in the workspace.

This does not mean that everyone using Power BI on F64 receives unrestricted free access.

The distinction is important:

Viewer is not the same as publisher or editor.

Users creating, publishing, sharing or collaborating on Power BI content can still require an appropriate Pro or PPU licence, depending on the scenario.

For an organisation with a large reporting audience, this difference can materially affect the overall business case for F64.

However, F64 should not be purchased purely to reach the Power BI licensing threshold.

The capacity still needs to provide enough compute for the organisation's Fabric workloads.

What Else Goes Into the Total Fabric Cost?

Fabric capacity is only one part of the overall budget.

Depending on the implementation, businesses also need to consider:

• OneLake storage

• Power BI licensing

• Spark autoscale usage

• Capacity overage

• Other applicable Azure charges

• Implementation and configuration

• Administration and support

The most important point is to avoid treating Fabric as a simple per-user licence.

The capacity, storage and licensing components need to be assessed together.

OneLake Storage Pricing

OneLake storage has its own pricing model, separate from Fabric capacity compute.

Microsoft's current pricing page lists separate OneLake storage categories including hot, cool, cold and cache storage, along with Business Continuity and Disaster Recovery storage options.

The cost depends on factors such as:

• Amount of data stored

• Storage tier

• Data retention

• Access patterns

• BCDR requirements

• Cache usage

There is also an important Power BI distinction.

Microsoft states that Power BI native storage is separate from OneLake storage, while data stored in OneLake for Power BI import semantic models is included in the Power BI licensing price.

That means businesses should not simply multiply every GB in their Power BI environment by a OneLake storage rate.

The storage architecture needs to be understood first.

Mirroring storage

Microsoft also provides free Mirroring storage for replica data based on the purchased capacity SKU.

For example, an F64 provides up to 64 TB of free Mirroring storage for eligible replica data.

This allowance is specifically for Mirroring replica data. It is not a general free OneLake storage allowance. Microsoft states that OneLake storage is billed when the free Mirroring limit is exceeded or when the provisioned capacity is paused.

Capacity Overage: An Important 2026 Cost Consideration

Capacity overage is another factor to include in a Fabric cost model.

Microsoft allows eligible Fabric capacities to enable Fabric capacity overage, which lets organisations pay for excess consumption instead of immediately experiencing capacity throttling.

Administrators can select which capacities have overage enabled and set a 24-hour paid-overage limit.

There is an important pricing distinction: Microsoft states that overage can apply to both PAYG and reservation capacities, but the overage is priced relative to the PAYG rate and does not receive the reservation discount.

That makes overage useful as a protection mechanism, but it should not become a substitute for proper capacity sizing.

If an F16 environment regularly requires substantial overage, it is worth investigating whether moving to F32 would provide a more predictable and economical operating model.

Spark Autoscale Billing

Spark workloads have a separate pricing option worth knowing about in 2026.

Microsoft offers Autoscale Billing for Spark as an opt-in PAYG model for dynamic, ad-hoc or unpredictable Spark workloads.

Under this model, Spark jobs run on dedicated serverless resources rather than consuming the shared Fabric capacity, helping reduce resource contention with other Fabric workloads.

The base Fabric capacity is still required for non-Spark workloads and OneLake.

Microsoft also states that standard regional PAYG CU-hour rates apply to Spark autoscale billing and that reservation discounts do not apply.

For organisations with highly variable Spark processing, this is worth evaluating separately from the core Fabric capacity requirement.

Pay-as-You-Go vs Reservation

Microsoft Fabric supports both pay-as-you-go and reservation purchasing options.

PAYG is useful when usage is still being measured or workloads change significantly.

It also provides the flexibility to scale capacity up or down and pause it when it is not required.

Reservations are better suited to predictable, sustained workloads.

Microsoft currently advertises savings of up to 40.5% over standard pay-as-you-go pricing for eligible Microsoft Fabric reservations. Its reservation guidance specifies that Fabric Capacity Units are available in one-year increments for this saving, excluding Power BI capacity products.

The important word is up to.

The saving should not be presented as a guaranteed 40.5% reduction in every commercial scenario.

When PAYG makes sense

PAYG is worth considering when:

• Fabric is still being evaluated.

• Development and test environments are used intermittently.

• Workloads are seasonal.

• Usage varies significantly throughout the year.

• You want to measure actual consumption before committing.

When a reservation makes sense

A reservation deserves consideration when:

• Fabric is already running as a production platform.

• Capacity usage is predictable.

• The organisation expects to maintain the same baseline capacity.

• The financial benefit of a longer commitment outweighs the flexibility of PAYG.

For example, an organisation might operate an F64 baseline throughout the year but temporarily scale higher during major reporting periods.

The right purchasing strategy could therefore involve a reserved baseline combined with PAYG capacity when additional compute is required.

How Fabric Data Warehouse Affects Your Costs

Fabric Data Warehouse runs on shared Fabric capacity rather than requiring a separate Fabric Data Warehouse capacity licence.

That means its cost is tied to the compute capacity required by the overall Fabric environment.

For example, one capacity could simultaneously support:

• Data Factory pipelines

• Data Warehouse queries

• Data Engineering notebooks

• Semantic model refreshes

• Power BI reports

• Real-Time Intelligence

If a warehouse executes demanding queries while Power BI models are refreshing and engineering jobs are running, the combined workload puts more pressure on the shared capacity.

This is why assessing each workload independently can produce the wrong answer.

The relevant question is:

How much capacity does the combined workload require at its busiest realistic periods?

What Affects Microsoft Fabric Pricing in the UK?

Several factors influence the final cost.

Workload demand

Data pipelines, warehouse queries, notebooks, semantic model refreshes and Power BI activity all consume capacity.

Concurrency

Two organisations with the same data volume can have very different capacity requirements if one runs workloads sequentially while the other runs them simultaneously.

Active hours

A capacity that runs 730 hours per month costs considerably more than one that is active only during working hours.

Storage

OneLake storage has its own cost model, so data retention and storage tier should be included in the budget.

Region

Microsoft pricing varies by region, making the UK South and UK West selection relevant when preparing an estimate.

Purchasing model

PAYG provides flexibility, while reservations can reduce the cost of predictable capacity usage.

Capacity management

Scaling, pausing, overage controls and workload optimisation can all affect the final bill.

How to Choose the Right Fabric Capacity

Capacity selection should start with workload evidence.

Before selecting a SKU, assess:

1. How much data will be processed?

2. How much data will be stored?

3. Which Fabric workloads will run?

4. How frequently will pipelines execute?

5. How often will semantic models refresh?

6. How much concurrent activity is expected?

7. Are there predictable workload spikes?

8. How many hours per month will the capacity actually run?

9. How many users will consume Power BI reports?

10. Does the organisation benefit from the F64 licensing threshold?

For a new implementation, starting with PAYG provides an opportunity to measure actual demand.

The goal is not to find the smallest possible SKU.

It is to find the smallest capacity that reliably handles the workload without creating recurring performance problems or unnecessary overage.

If an environment consistently approaches its capacity limits, scaling up is justified.

If a capacity remains substantially underused, the business should investigate whether the SKU can be reduced or whether workloads can be scheduled more efficiently.

Is Microsoft Fabric Worth the Cost for UK Businesses?

Whether Fabric is worth the investment depends on the workloads it replaces or consolidates.

For organisations already invested in Microsoft technologies such as Power BI, Azure and Dynamics 365, Fabric can bring data integration, engineering, warehousing, analytics and real-time workloads into a shared environment.

That can reduce the complexity of maintaining separate platforms for different parts of the data estate.

But Fabric is not automatically the most economical option for every reporting requirement.

A business with straightforward reporting needs may not need F64 or a larger capacity.

A business running multiple pipelines, a warehouse, engineering workloads, semantic models and a large Power BI audience has a different cost equation.

The useful comparison is therefore not simply the Fabric capacity price.

It is the total cost of operating the required data platform and delivering the required reporting and analytics capabilities.

Final Word

Microsoft Fabric pricing in the UK is ultimately a capacity-sizing decision.

Using the September 2026 UK South reference rate in this guide, an F2 works out at approximately £225.72 per month when active for 730 hours, while an F64 works out at approximately £7,222.91 per month.

Those numbers provide a useful starting point, but they are not a complete Fabric budget.

The final calculation needs to account for active hours, region, OneLake storage, Power BI licensing, Spark workloads, overage and the purchasing model.

For predictable workloads, Microsoft currently advertises up to 40.5% savings on eligible Fabric reservations compared with standard PAYG pricing.

For organisations evaluating F64, the Power BI licensing implications deserve particular attention. Free users can consume Power BI content on qualifying F64+ capacity when they have the appropriate Viewer role, but publishing and other collaborative activities can still require paid Power BI licensing.

The practical buying lesson is simple: do not choose a Fabric SKU from user count alone. Measure the workloads, understand concurrency, estimate active hours and then select the capacity that gives you enough headroom without paying for compute that remains idle.

Contact Dynamics Square UK

Choosing the right Fabric capacity is only one part of the implementation decision.

Dynamics Square UK can help you assess your data workloads, plan a Microsoft Fabric environment, evaluate capacity requirements and understand how Fabric fits into your existing Microsoft ecosystem.

Whether you are expanding Power BI, consolidating data platforms, building a Fabric Data Warehouse or planning a broader Microsoft Fabric deployment, the right capacity model can make a significant difference to both performance and cost.

Contact Dynamics Square UK to discuss your Microsoft Fabric requirements and pricing options.