Power BI Premium P SKUs are being retired: plan your move to Fabric F SKUs

Microsoft has announced the end of life for Power BI Premium per capacity SKUs, the P1 to P5 subscriptions many organisations have built their reporting on. Going forward, capacity is bought as Microsoft Fabric F SKUs through Azure. Nothing breaks today, but if your company runs on a P SKU, your next renewal is now a migration project.

The good news is in the first lines of the Power BI blog: the Power BI Premium product capabilities will not change, and there is no immediate action required. What changes is how you buy capacity, and when you have to switch.

The dates that matter

The retirement hits you differently depending on how you buy. According to the announcement:

• New customers cannot buy Power BI Premium per capacity after July 1, 2024.
• Existing customers without an Enterprise Agreement (EA) can renew P SKU subscriptions until January 1, 2025. Renewals after that date have to move to Fabric capacity.
• Customers with an EA can keep renewing their P SKU annually until the end of their EA. If the agreement ends after January 1, 2025, they move to Fabric capacity at that point.
• Customers in sovereign clouds are not affected for now, as they do not have access to Microsoft Fabric yet.

Power BI Pro and Premium Per User (PPU) are not affected by this retirement. This is about capacity licensing only.

So the real deadline is not a single date. It is your own renewal date, or the end of your EA. Find that date first, because everything else follows from it.

Why Microsoft is doing this

Fabric went generally available last November, and it is built on the Power BI platform. Microsoft describes Fabric as a superset of Power BI Premium: everything Premium does, plus the other workloads such as lakehouses, warehouses and notebooks. Existing P capacities can already run Fabric items once Fabric is enabled.

Running two parallel ways to buy the same compute makes little sense once Fabric reservations are available. The blog points out that F SKUs are bought through Azure, can count towards a Microsoft Azure Consumption Commitment (MACC), and give access to Azure-only features such as trusted workspace access for OneLake shortcuts and managed private endpoints.

F64 is your P1, in compute terms

The capacity table on Microsoft Learn lines the two families up by compute. F64 has 64 capacity units, which matches the 8 v-cores of a P1. The rest follows the same pattern:

• P1 maps to F64
• P2 maps to F128
• P3 maps to F256
• P4 maps to F512
• P5 maps to F1024

F64 is also the line that matters for licensing. On F64 or larger, users with a free license and a viewer role can view Power BI content, the same as on a P SKU today. Below F64, every viewer needs a Pro or PPU license.

That is the trap I would watch for. Picking a smaller F SKU to save money can look smart on paper, until you count the Pro licenses you now need for every report consumer. If you rely on free viewers today, start your planning at F64.

What changes in the way you run capacity

The capabilities stay, but the operating model moves from a Microsoft 365 subscription to an Azure resource. That has practical consequences:

• F capacities are created and managed in the Azure portal, under an Azure subscription and resource group.
• Billing is per second on pay-as-you-go, with the option of a yearly reservation to bring the cost down.
• You can scale an F capacity up or down as your workload changes.
• You can pause an F capacity to stop billing when it is not needed, and resume it later. Pausing makes the content on it unavailable, so this is for dev and test, not your production reports.
• Pricing is regional, so check the price in the Azure region you plan to use.

For many Power BI teams, this is the first time their BI platform lives in Azure. Someone needs Azure permissions, someone owns the cost, and finance needs to know the invoice will look different.

What to do next

Moving the content itself is the easy part. You assign your workspaces to the new F capacity, the same way you assign workspaces to a P capacity today. The work is in the planning around it.

A sensible order:

1. Find your renewal date, and check whether you buy through an EA or another agreement. Talk to your Microsoft account representative or licensing partner early.
2. List every P capacity, its size, its region and the workspaces assigned to it.
3. Pull utilisation from the capacity metrics app. If a P1 is idling at 30 percent, or throttling every month end, this is the moment to resize rather than copy the old size.
4. Decide on billing. A capacity that runs all day, every day is a candidate for a reservation. Dev and test capacities may be cheaper on pay-as-you-go with pausing.
5. Sort out Azure ownership: subscription, resource group, who can create and scale capacities, and how costs are tagged and charged back.
6. Create the F capacity in the same region, move a few non-critical workspaces first, check refreshes and gateways, then move the rest.

Leave room in the plan. A capacity move is quick, but approvals, purchasing and Azure access can take weeks in a larger organisation.

Takeaway

After two decades in data, I have seen plenty of licensing changes that sounded dramatic and turned out to be paperwork. This one is mostly paperwork too, but it moves your BI platform into Azure, and the F64 line decides whether your free viewers keep working. Treat your next renewal as a small project, not an invoice.

When is your P SKU up for renewal, and have you checked whether you really need the same size on the other side?

Sources

• Important update coming to Power BI Premium licensing (Power BI blog)
• Important update to Power BI Premium licensing (Microsoft Licensing news)
• Understand Microsoft Fabric licenses and capacity (Microsoft Learn)
• Pause and resume your Fabric capacity (Microsoft Learn)