Navigating Power BI & Fabric Licensing

Power BI licensing has certainly evolved over the years, and with so many changes, it’s no wonder questions keep coming up. Below are a few that I have heard and seen:

  • What Power BI license do I need?
  • How can I find out what Power BI license I have?
  • Do we need both Pro and Premium Per User (PPU) licenses?
  • Does Power BI Embedded and Premium Capacity charges stack?
  • Is Premium Capacity the only way to offer free read-only access?
  • Does a the Fabric Admin need a Pro license?
  • Which Power BI license is needed to view reports?
  • What is MS 365 E5, and why should I care?
  • What are the differences between Microsoft Fabric and Power BI Licensing?
  • What license do I need to share with my colleagues?
  • What happens when Power BI Premium is retired?

I know, it’s a lot – but don’t worry. My goal with this blog is to tackle as much as I can. I have to admit, when I first decided to write about Power BI licensing, I thought, “Ha… easy…” But I quickly realised I was wrong. Sure, I knew the fundamentals, but it took some research and testing to validate and understand everything. So, consider this your in-depth guide.

Also, a quick side note, all blogs are released in advance on our website – be sure to check it out: Metis BI Blog | Power BI Tips, Best Practices & Insights

So, What Power BI Licensing Types Are Available?

To keep things simple (and to save myself from having to update the diagram too often), we’ll focus on the following five licensing models:

Licensing Diagram

That said, I can’t continue without mentioning a few others that come to mind. First, there’s Power BI Premium (P-SKUs), which is being retired. While I’ll reference it throughout this blog, it won’t be part of our main list. Then, there’s the Fabric Trialsince it’s a 60-day trial, I’m not including it officially, but I know many people are using it. Finally, there’s Power BI Report Server, which I’ll cover briefly towards the end of this blog.

Fabric Free / Power BI Desktop

That’s right, it’s no longer called “Power BI Free” – now it’s “Fabric Free”! Not the biggest fan of the name change, but let’s get into what it means. Power BI is a tool you can download for free from your browser or the Microsoft Store. Just a note on this, downloading from the MS Store means automatic monthly updates. Power BI Desktop is an application installed locally on your machine and from here, you can do a lot.

You can ingest, transform, and model your data. You can create DAX calculations(think business calculations), visualise your data to identify those previously unknown patterns and trends and most importantly export data to Excel, CSV and PowerPoint 🙃 All of this is possible with a Fabric Free license. Just keep in mind, if you want to publish to your personal “My Workspace” (basically your personal folder in the cloud), you’ll need the Fabric Free license assigned to your account.

Starting with Fabric Free is a great way to get comfortable with the tool. Like I explained above, you have access to all the fundamental features and capabilities you need to get started. In fact, when I’m helping people get going, this license works really well.

If you’re worried about the number of data connectors or the data volumes you can work with – don’t be! You aren’t limited to a certain number of data connectors(unlike some other BI tools out there). And when it comes to data volumes, you get 1GB semantic models (same as Pro Licensing). But what does that actually mean? Well, it’s not exactly just 1GB – that’s after compression 🤯 So, your model could be 10 to 15 times larger before compression. Also, remember that you can go beyond this limit if your machine can handle it- don’t recommend it though. The 1GB post-compression limit really matters when you want to publish to Power BI Service, also known as the PowerBI.com, Power BI Cloud, or now with Fabric – the Fabric Service? You get the idea though.

If you want to see whether you are on a Fabric Free license (or any other), whilst in the Power BI Service, click on the profile icon (circle with your pic) in the top right.

You’re probably now thinking, “Laz… I’m sticking to a Fabric Free license…Desktop does it all.” Well, I should mention that it limits you on everything sharing and collaboration. Let’s be honest, as great as a Power BI solution can be, it’s pretty useless if it’s only available to one person and can’t be shared, right? After all, we want our colleagues, teams, departments, and entire organisation to see the insights we generate. With Fabric Free, the main limitation is that you can only publish to your personal “My Workspace”, which isn’t ideal for collaboration. To further add to this, if you wanted to “publish” your solution to what is known as a “App Workspace” where others have access, you can’t. A quick note, please don’t think the solution is to share the PBIX (Power BI Desktop filetype) over email 🙅 With that said, I feel like it’s now a good time to move on to the next licensing model!

Power BI Pro

The easiest way to describe Power BI Pro is that it’s like Fabric Free, but with sharing and collaboration capabilities. Let’s get the cost out of the way first, it’s a per user license costing £8.20 per month – you can check it out here. For users in the US, it’s priced at $10 per user – see here. After checking in the MS 365 Admin Centre, you can purchase it monthly or annually, but it will always come to £8.20 per month.

A common question I often hear is, “Why am I hearing about Microsoft 365 E5when talking about Power BI?” Well, Microsoft 365 is a subscription suite that includes Office apps, Teams, cloud storage, and more. The E5 plan includes Power BI Pro, so if you’re subscribed to the E5 enterprise plan, costing £50.30per user/month, you won’t need to pay an additional £8.20 per user/month for Power BI Pro – it’s already included! 💸

Note: If your organisation has E5 licenses, all users get “free” read-only access. Many organisations start to consider a Premium capacity (Fabric) to enable free read-only access for everyone. So, if you reach a certain number of Pro licenses, it may be more cost-effective to simply switch to a capacity-based model. For instance, an F64 reserved capacity starts at £3,733.89, which is similar to the cost of 455 Pro users at £8.20 each. So, with E5, you can get read-only access without extra costs. Keep in mind though, just because you are on a capacity-based model, it don’t always mean better performance 😉

As mentioned, one of the core differences between Power BI Pro and Fabric Free is the ability to share and collaborate. What does this mean? With a Power BI Pro license, any user in the organisation can access and share solutions that others have shared or published to an app workspace. This goes beyond just “My Workspace,” which is limited to individual access only. With Pro, users can also create shareable links to distribute solutions to others, but sharing is strictly possible only if both the person sharing and the recipient have a Pro license.

Now, I don’t want to mislead anyone by saying the only difference between Pro and Free is sharing and collaboration. There are a few other limitations, though these features may not be as commonly used. For instance, Subscriptions aren’t available without a Power BI Pro license – just keep this in mind.

Premium per User (PPU)

Moving on to Premium per User or PPU! Think of it as the middle ground. Some organisations need more advanced analytics but don’t have the budget for a Premium or Fabric capacity. That’s where PPU fits – it sits right between Power BI Pro and Premium (Fabric).It’s a per user license priced at £16.40 per user/month, which you can see here. Also, if you already have Power BI Pro or E5 licensing, you get a discount on PPU! I’m not sure of the exact breakdown in GBP (I think its £8.20), but in USD, a Power BI Premium license is $20 per user/month, which is reduced to $10per user/month.

Okay, so what makes PPU so, premium? Well, if you take a look at the diagram above, you’ll see that your semantic model limit jumps to 100GB, the total storage per user increases to 100TB, and the number of refreshes shoots up to 48 per day – compared to just 8 with Power BI Pro. On top of that, PPU comes with extra features like Azure Cognitive Services, enhanced Dataflow capabilities, Datamarts, Deployment pipelines, XMLA endpoint connectivity, enhanced automatic page refresh, and real-time DirectQuery partitions in Incremental Refresh.

Here’s something really important that often catches organisations by surprise. Let’s say you need an advanced feature, like Deployment Pipelines. So, you assign a PPU license to one person, who then creates the necessary workspace with Deployment Pipelines, and invites other Pro users to simply consume the content. Here’s the catch – those Pro users won’t be able to just consume or access that content! To view or interact with any artefact from a PPU workspace, every user must also have a PPU license, not just Pro.

To finish up with Premium per User (PPU), this license doesn’t mean that the Fabric Metrics App is needed to monitor memory and CPU usage, as it’s not applicable to PPU licenses. We use the Metrics Apps to monitor our capacity-based licensing, which actually helps simplify resource management for PPU. However, admins still have the ability to manage PPU licenses, users, and settings through the Admin portal, making it easier to configure these options effectively.

Fabric

We have now arrived at our first capacity-based licensing model. As it’s Fabric, we will be spending more time going through this. So, what does capacity-based licensing really mean? Think of it as a container filled with dedicated resources – like Capacity Units (equivalent to vCores) – that are solely for your use and your organisation’s needs. Unlike the previous licensing models we discussed, which are classified as shared capacity, dedicated means you have your own set of resources.

When MS Fabric was first announced, it ran parallel to Power BI Premium. Essentially, several FSKUs, ranging from F2 to F2048, matched an equivalent P SKU in functionality and capacity. For instance, F64 corresponded to a P1, while F0124 aligned with a P5. But with the launch of Fabric, new SKUs have been added, offering more options and flexibility. Check out the following table for the details and to see how new SKUs have been introduced.

Fabric to Premium SKUs

Where is Premium Gone?

Before going further into Fabric, Power BI Premium deserves it’s last moment… P SKUs are going to the 🪦 That’s right, as of July 2024 P SKUs are no longer being sold to new customers. As of January 2025, existing customers won’t be able to renew their P SKU contracts. As of January 2026, Power BI Premium PSKUs will reach their end of life for customers without an Enterprise Agreement.

So, from what I understand there is no need to panic! You can simply reassign workspaces to the new F SKU equivalent capacity, and there is an option to bulk reassign, which saves time. By “bulk reassign”, I imagine this means reassigning workspaces from existing an P capacity to new F capacity.

Fabric… how much?

As mentioned earlier, we have F SKUs ranging from F2 to F2048, each with different resources, primarily Capacity Units (CUs), and relevant pricing. A key difference between F SKUs and P SKUs is that Fabric now offers both Pay-as-you-Go (PAYG) and Reserved Capacity options. This choice affects the overall cost depending on the type of capacity you choose. Below, you can seethe detailed breakdown for each SKU and associates cost for PAYG and Reserved.

Fabric Monthly Pricing

Before we finish up with the costs, there is one more thing… You need to pay for storage separately. Below is taken from the MS site which you can find here.

  • OneLake Storage: General data storage at£0.0172 per GB/month.
  • OneLake BCDR Storage: Backup storage at £0.0310per GB/month, for recovery and disaster scenarios.
  • OneLake Cache: Cache storage for faster performance, at £0.1837 per GB/month.

How to Provision and Access a Fabric Capacity?

If like me, you area Power BI expert and focused on front-end analytics, we would usually go through the 365 Admin Centre. However, provisioning an F SKU requires us to do so through the Azure Portal, similar to provisioning A SKUs. Also worth noting, you can enable MS Fabric at the tenant or the capacity level. In regards to how we can access Fabric workloads, there are three ways which I listed below:

  • Premium Capacity P-SKUs: Existing Premium capacities, which are being retired.
  • Fabric Capacity F-SKUs: New Fabric capacities, available as PAYG or Reserve.
  • Fabric Trial License: For those looking to dip their toes in all the new and exciting experiences.

Note: You would probably think, Copilot comes with any of the F SKUs. However, keep in mind, for the F SKUs you must have an F64 or above to access MS Copilot. It’s not available with any of the lower F SKUs.

Understanding Compute Power

An F64 SKU provides the same compute power of a Power BI Premium P1 SKU. We now use Capacity Units(CUs) with MS Fabric SKUs not vCores as we always had with Power BI Premium. To make more sense of this, a single Power BI Premium vCore is equivalent to 8CUs.

Access to MS Fabric and Power BI Artefacts Across FSKUs

Exploring MS Fabric’s capabilities reveals an important distinction in accessing different SKUs, ranging from F2 to F2048. While you can leverage MS Fabric features and workloads across any F SKU, it’s important to understand the access limitations for Power BI artefacts. F SKUs below F64 do not provide free read-only access to Power BI artefacts.

For instance, while users within an F4 SKU can access Fabric artefacts, accessing Power BI artefact such as reports will require those users to have a Power BI Pro or PPU license. This aligns with a P1 SKU where users consuming reports did not need their own Power BI Pro license, but only a Fabric Free license.

No matter which FSKU you’re using, if you want to develop a semantic model, create a report, or publish a solution, you’ll need a Power BI Pro or PPU license. This also applies to your Power BI Service Admins or now, Fabric Admins.

Can Fabric Handle Larger Volumes?

When it comes to Fabric, it’s worth noting that the total size of your semantic model is quite similar to Power BI Premium. For example, with a P1 capacity, the limit is 25GBper semantic model – after compression. The same goes for the F64 SKU. Importantly, this limit applies to each individual artifact rather than the accumulated total –  unlike how it worked with Premium Gen1 in the past. Below, you can see the breakdown of memory limits per F SKU, plus more:

Fabric Max Limits

Note: If you’re wondering when I’ll dive into the all-new DirectLake storage mode, I’ve got you covered! Check out my in-depth blog here, all about the different storage modes, including DirectLake. Here, I’m intentionally focusing on licensing, so head over there if you’re keen to learn more about storage.

Embed

We have now arrived to the last licensing model we will be exploring together, Power BI Embedded. Now, as the name suggests, this is used to embed our reports into custom applications like web apps or websites or into our internal applications such as SharePoint, Teams, Dynamics, etc.

When it comes to embedding Power BI, you have a few options depending on the specific scenario and more importantly, your authentication needs. For this reason we need to be aware of the terms “Users Own Data” and “App Owns Data”.

When we are embedding within internal applications such as Teams or SharePoint, this is classified as “User Owns Data”. To do this, users must have a per user license (Pro/PPU) or the report must be hosted in a F, P or EM SKU workspace. For the authentication, users would be authenticated through Entra (formerly Azure Active Directory).

When we are embedding within a custom application, this is classified as “App Owns Data”. This is ideal for custom applications needing Power BI reports embedded for external clients. The reports must be hosted in a F SKU or an ASKU workspace. A service principal is required, as it manages the connection and tokens. Authentication is handled by the custom application, so in other words the users do not need to sign in with a Power BI account. This is recommended for scenarios without Pro requirements and where end-users won’t know Power BI is being used.

Keep in mind, A SKUs are billed on an hourly basis, allowing for flexibility to pause, resume, and scale up or down as needed. They are also suitable for development and testing environments, offering an efficient way to manage resources without using Premium(Fabric) capacities. EM SKUs, do not provide direct access to the Power BI Service. While all F SKUs support embedding capabilities, free read-only access is only available with F64 or higher, enabling users without a paid license to consume content. See the diagram below for specific resources associated with each embed SKU.

Fabric to PBI Embedded SKUs

Power BI Report Server

I don’t have much to add here, except that given an F64 SKU is equivalent to a P1 SKU, it seems reasonable to assume that Power BI Report Server (PBIRS) should also be included. However, I haven’t found any definitive information from Microsoft confirming this yet. It’s something to keep an eye on for further updates.

Licensing Scenarios

This section will explore the different Power BI licenses available and guide you in choosing the best licensing model for various scenarios. Whether you need to embed reports, provide self-service analytics, or scale up to enterprise-level solutions, understanding these licensing options will help you select the right fit based on features, capacity, and cost.

Scenario 1

An organisation has 120 users, all needing access to Power BI to varying levels:

  • 70 users will be pure report consumers, accessing content through the Service.
  • 50 users will be developing semantic models, authoring reports, and to publishing these reports for others.

To cover everyone’s needs, all 120 users require a Power BI Pro license to create, share, and view shared reports.

Cost breakdown:

120 users x £8.20 per month = £984 per month.

Scenario 2

An organisation has 150 users, all needing access to Power BI to varying levels:

  • 40 users will be developing semantic models, authoring reports, and to publishing these reports for others. They will also need to take advantage of some Premium features suchas Deployment Pipelines.
  • 110 users will be pure report consumers, accessing content through the Service.

To cover everyone’s needs, all 150 users require a Power PPU license to create, share, and view shared reports.

Cost breakdown:

150 users x £16.40 per month = £2,460 per month.

Scenario 3

An organisation has200 users, all needing access to Power BI to varying levels:

  • All users in the organisation have MS 365 E5 licensing, meaning they have a Power BI Pro license available.
  • All 200 users are using their Power BI Pro license for authoring and sharing.
  • As their needs evolve, they now need Premium features such as 48 refreshes per day and larger model limits.

To cover everyone’s needs, all 200 users require a Power PPU license to create, share, and view shared reports. However, as they all have an E5 they can purchase the Power BIPPU Add-On.

Cost breakdown:

200 users x £50.30 per month = £10,060 per month (This is not a Power BI license)

200 users x £8.20 per month = £1,640 per month

Scenario 4

An organisation has600 users, all needing access to Power BI to varying levels:

  • 100 users will be developing semantic models, authoring reports, and to publishing these reports for others.
  • 500 users will be pure report consumers, accessing content through the Service.

To cover everyone’s needs more cost-effectively, you could purchase an F64 SKU along with 100additional Power BI Pro licenses. In this scenario, Power BI Pro for 600 userswould cost £4.9K monthly, which exceeds the £3.7K for an F64 reserved capacity. On top of this, you need 100 Pro licenses for authoring and sharing, adding£820 per month, making the total £4.5K. This offers a cheaper alternative to purely relying on Pro licenses, with access to more advanced features. However, note that dedicated capacity does not always guarantee better performance compared to shared capacity – you must monitor your capacity!

Cost breakdown:

Single F64 Reserved Capacity = £3,733 per month

100 users X £8,20 per month = £820 per month

Total Cost = £4,553

Scenario 5

An organisation has 1000 users, all needing access to Power BI and Fabric workloads to varying levels:

  • 25 will be data engineers working with various Fabric experiences with “large” volumes. A minimum of 25GB per semantic model will be required.
  • 50 will be using Power BI for developing and sharing reports.
  • 250 will be self-service users, developing and sharing their own reports.
  • 675 will be pure report consumers, accessing shared content through the Service.

For this scenario, the best approach would involve a combination of a Fabric Reserved Capacity and Power BI Pro licenses. To support the necessary Fabric workloads and the 675 report consumers, an F64 reserved capacity should be utilised at £3.7K per month. The 50 Power BI developers and 250 self-service users can all use Power BI Pro licenses.

Cost breakdown:

Single F64 Reserved Capacity = £3,733 per month

300 users X £8,20 per month = £2,460 per month

Total Cost = £6,193

Scenario 6

An organisation has6000 users across various regions. All 6000 are covered with an E5 licensewhich includes a Power BI Pro license. However, Marketing requires their teamto take advantage of some Premium only features.

  • 6000 users in the organisation require a Power BI Pro license which they receive through E5.
  • 100 Marketing users ‘needs Premium features that are not available with Pro licensing.

For this scenario, the best approach would involve the organisation continuing to use their Power BI Pro licenses through E5 and for Marketing to purchase 100 individual PPU licenses. As they have E5, this will be the add-on price and not the full price of £16.40 per user/month. Keep in mind, if anyone else from the organisation wants to access any marketing workspaces hosted in a PPU capacity, they will not be able to with a Pro license.

Cost breakdown:

6000 users x £50.30 per month = £301,800 per month(This is not a Power BI license)

100 users x £8.20 per month = £820 per month

Scenario 7

A law firm has E5licensing for all its internal reporting needs. However, they are building a client-facing application that requires some Power BI reports to be embedded within it/

  • The firm has E5licenses that include Power BI Pro, meaning internal collaboration is covered.
  • They need to embed Power BI reports into an external client-facing application.

In this scenario, the best approach would involve using Azure Embedded A SKUs to support the Power BI embedded functionality for external clients. The A SKU provides dedicated capacity for embedding, with the specific SKU size depending on the anticipated load and number of users accessing the embedded reports. Since the law firm does not require any internal Premium features, an A SKU would keep costs in check while meeting their needs. To determine the best A SKU size, they can start with an hourly capacity model to assess the CPU and memory requirements and adjust accordingly based on monitored usage.

Summary

To wrap up, choosing the right Power BI license is key to making the most of its capabilities for your team. Each option is designed to fit different needs, from simple sharing to large-scale embedded solutions. If you have questions or need further guidance, feel free to reach out – we’re here to help!

About the Author:

Lazaros Viastikopoulos

Power BI Expert 🚀 Helping Organisations Transform Data into Insights by Overcoming Challenges with Proven Approaches in Reporting, Governance, Training, Optimisation, Storytelling. Ask About Our Proven Power BI Methods.

Lazaros Viastikopoulos | LinkedIn

Reference:

Viastikopoulos, L (2024). Navigating Power BI & Fabric Licensing. Available at: Power BI Licensing Guide: Choosing the Best License for Your Needs [Accessed: 04th November 2024].

Share this on...

Rate this Post:

Share: