Microsoft is retiring Power BI Premium per-capacity SKUs (P1 through P5) and moving every customer to Microsoft Fabric capacity SKUs (F SKUs). The migration is not automatic and it is not a rename: you manually purchase an F SKU in the Azure portal, reassign each workspace to it, then cancel the P SKU once you have validated the move. Do it before your current agreement term ends, because after that a grace-and-throttle clock starts that ends in your content becoming inaccessible.

This guide is written for the IT directors, data platform owners, and finance leaders who have to make that call in 2026. It covers the retirement timeline and the deadline that actually matters, the official P-to-F SKU mapping, the F64 threshold that decides whether you keep paying for viewer licenses, the real cost math, and the migration sequence that avoids an outage.

Key Takeaways

• P SKUs are being retired at the end of each customer's agreement term. Microsoft no longer sells new Power BI Premium per-capacity SKUs, and customers with expiring Enterprise or Microsoft Cloud Agreements can no longer add or renew them.
• The migration is a workspace reassignment, not a data move. You buy an F SKU in Azure and reassign workspaces to it; standard Power BI content carries over without moving data.
• F64 is the line that controls licensing cost. On F64 or larger, free (Viewer-role) users can consume Power BI content with no per-user license; on F2 through F32, every viewer needs Power BI Pro or PPU.
• The capacity price is roughly flat P1-to-F64; the savings come from consolidation. The business case is folding separate Synapse and Data Factory workloads onto the same Fabric capacity, not a cheaper base SKU.
• Missing the deadline has teeth. After your P SKU ends you get 30 days of grace, then throttling, then a full block on day 91 where data is retained but inaccessible.
• Pro and PPU are not affected. Only the per-capacity P1 through P5 SKUs are retiring; per-user licenses stay exactly as they are.

What is the Power BI Premium to Fabric migration, in one answer?

The short version: Microsoft Fabric is a superset of Power BI Premium, and F SKUs are the replacement for the retiring P SKUs. Fabric keeps every Power BI Premium capability and adds workloads like Data Engineering, Data Warehouse, Real-Time Intelligence, and Data Factory on one capacity. To keep running Premium features, you move to an F SKU.

What changes for you falls into four buckets:

• Where you buy: P SKUs came from the Microsoft 365 admin center; F SKUs are purchased in the Azure portal against your Azure subscription.
• How you pay: P SKUs used a monthly or yearly commitment; F SKUs default to pay-as-you-go (billed per second, one-minute minimum) with optional one- or three-year reservations for significant discounts.
• How you scale: F SKUs support on-demand resizing plus pause and resume, so you can stop billing during idle periods; the old Autoscale model does not exist on F SKUs.
• What you unlock: Azure-native features (trusted workspace access, managed private endpoints, Azure Monitor, Microsoft Cost Management) and Microsoft Azure Consumption Commitment (MACC) eligibility that P SKUs never had.

According to Microsoft's migration FAQ, the retirement applies only to per-capacity SKUs; Power BI Pro and Premium Per User remain active and unchanged.

When is the deadline to migrate off Power BI Premium P SKUs?

There is no single global cutoff date. Microsoft retires each P SKU subscription at the end of that customer's current agreement term. New P SKU purchases were removed from the flow back in 2024, and customers with expiring Enterprise Agreements or Microsoft Cloud Agreements can no longer add or renew P capacity. Enterprise Agreement customers can keep renewing existing P SKU capacity annually until their EA term ends, then must transition.

The date that matters to you is your agreement renewal. Per Microsoft's P-to-F decision guide, once your P SKU subscription ends the clock is unforgiving:

• Days 1 to 30 (grace): Your capacity keeps running at its current size at no charge, and you can still reassign workspaces to a new F SKU.
• Days 31 to 90 (throttled): Interactive operations are delayed, but you can still export and recover data.
• Day 91 onward (full block): All operations are rejected. Your data is retained but inaccessible until you migrate the workspaces or delete the capacity.

Treat the grace and throttle windows as a safety net, not a planned migration window. If your P SKU is still active, the right move is to plan the transition now and complete it before renewal.

How do Power BI Premium P SKUs map to Fabric F SKUs?

Microsoft's mapping is based on capacity units (CUs), where each P SKU v-core corresponds to 8 CUs. The equivalent F SKU is a safe starting point, but it is a reference for compute capacity, not a promise of identical licensing or features.

Official P-to-F SKU mapping

P1

• F SKU equivalent: F64 (64 CUs, 8 v-cores).
• Typical use: Small to mid-size organizations and departmental deployments.

P2

• F SKU equivalent: F128 (128 CUs, 16 v-cores).
• Typical use: Mid-size organizations with distributed team reporting.

P3

• F SKU equivalent: F256 (256 CUs, 32 v-cores).
• Typical use: Large organizations with high-concurrency environments.

P4

• F SKU equivalent: F512 (512 CUs, 64 v-cores).
• Typical use: Enterprise-scale deployments with high-volume refreshes.

P5

• F SKU equivalent: F1024 (1,024 CUs, 128 v-cores).
• Typical use: Very large, maximum-scale deployments.

Do not treat the mapping as your final size. Open the Microsoft Fabric Capacity Metrics app, review the last 30 to 45 days of CU consumption, and right-size from there. If you were running your P1 near 60 percent utilization on Power BI alone, size up, because new Fabric workloads like Copilot and Data Factory draw from the same CU pool.

Does migrating to Fabric change my Power BI licensing costs?

The single most important number in this decision is 64. On an F64 capacity or larger, users with only a free Fabric license and the Viewer role can view Power BI content with no per-user license, exactly as they could on a P SKU. Drop below F64 to save on capacity and you cross a licensing cliff: on F2 through F32, every viewer needs a Power BI Pro or PPU seat, and authors always need Pro or PPU regardless of capacity size.

On base capacity price, P1 and F64 are close to a wash. Independent 2026 analysis from EPC Group puts F64 at roughly 5,003 dollars per month on a one-year reservation, essentially level with a P1 under an EA. The same analysis pegs the crossover between Premium Per User (about 24 dollars per user per month) and an F64 capacity at roughly 219 viewers; below that, PPU is often cheaper. Because F SKU pricing is regional and changes over time, confirm your own numbers in the Azure pricing calculator rather than trusting a headline figure.

The migration rarely lowers your base capacity bill; it earns its keep by collapsing three or four separate data platforms into the one capacity you were already paying for.

The real savings show up when you retire adjacent infrastructure. A migration runbook from Power BI Consulting models a typical P1-to-F64 move saving 1,500 to 8,000 dollars per month by consolidating separate Synapse and Data Factory workloads onto the same capacity that already serves Power BI, with Copilot and Real-Time Intelligence included as net-new value. The exceptions where you pay more: sub-P1 tenants that downsize below F64 and lose free viewer inclusion, and tenants with heavy Autoscale usage, which Fabric bills separately per burst.

Getting this right is a modeling exercise, not a guess. BabyBots runs fixed-fee Fabric migration readiness assessments that baseline your actual CU consumption, map each P SKU to the right F SKU, and build the reservation-versus-pay-as-you-go case against your renewal date, so you commit to a capacity you can defend rather than the one the mapping table suggests.

How do you actually migrate from a P SKU to an F SKU?

The mechanics are simpler than the planning. Migration is a workspace reassignment: you purchase an F SKU capacity in Azure and reassign each workspace from the P SKU to the new capacity. Standard Power BI content (reports, dashboards, and small or default-format semantic models) carries over without moving data, so a same-region reassignment is a metadata change, not a data migration.

A disciplined sequence looks like this:

1. Assess. Inventory every P SKU capacity, export usage from the Capacity Metrics app, identify peak CU and any throttling events, and map each capacity to its target F SKU.
2. Provision. Enable the Fabric tenant setting and create the target F SKU in the same Azure region as the legacy capacity to avoid a cross-region move.
3. Pilot. Reassign one non-critical workspace, validate query performance, row-level security, and scheduled refresh over about a week.
4. Cut over in waves. Move non-critical workspaces first, then the bulk, then business-critical and executive-visible content last, monitoring CU against your baseline at each step.
5. Retire. Cancel the P SKU only after the migrated workspaces run clean. Buy the F SKU first, validate, then cancel, never the reverse.

One serious caveat: avoid cross-region reassignment unless a business driver forces it. Standard Power BI items survive a cross-region move, but large storage format semantic models and all Fabric items (Lakehouses, Warehouses, Notebooks, and Data Factory pipelines) do not, so you must back them up and recreate them. This is exactly the kind of decision that belongs in a planned Microsoft Fabric migration strategy sequenced by business triggers rather than run as a rushed capacity swap.

Frequently asked questions

Is the P-to-F migration automatic when my agreement renews?

No. Nothing migrates on its own. You manually purchase an F SKU capacity in the Azure portal and reassign each workspace from the P SKU to the new capacity. If your P SKU simply ends without an F SKU in place, your content enters the grace-then-throttle-then-block timeline rather than moving over.

Do report viewers still need Power BI Pro licenses after migrating?

It depends on the F SKU size. On F64 and larger, users with a free Fabric license and the Viewer role can view content with no per-user license, the same behavior as on P SKUs. On F2 through F32, every viewer needs a Pro or PPU license, and authors and editors always need Pro or PPU regardless of capacity size.

Are Power BI Pro and Premium Per User being retired too?

No. The retirement applies only to the per-capacity SKUs, P1 through P5. Power BI Pro and Premium Per User are separate per-user licenses that remain active, and your users do not need a license change as part of this migration.

Will Fabric F64 cost less than my Power BI Premium P1?

On base capacity, the two are close to level; independent analysis puts a one-year reserved F64 at roughly the same monthly cost as a P1 under an Enterprise Agreement. The savings come from consolidation, folding separate Synapse and Data Factory spend onto the same capacity, plus included Copilot and Real-Time Intelligence. Confirm exact figures in the Azure pricing calculator, since F SKU pricing is regional.

What happens if I cancel my P SKU before buying an F SKU?

Avoid that sequence. If the P SKU ends first, your content gets a 30-day grace period, then throttled interactive access from day 31 to 90, then a full block from day 91 where data is retained but inaccessible until you migrate or delete the capacity. Purchase and validate the F SKU first, then cancel the P SKU.

Does the retirement affect sovereign or government clouds?

No. Microsoft Fabric is not yet available in sovereign clouds, so P SKUs remain supported there and this retirement does not apply. Microsoft has said it will publish separate guidance when Fabric reaches those environments.

Where this is heading

Fabric is the convergence point for Microsoft's data, AI, and automation stack, and the P SKU retirement is the first of several consolidations on a published schedule stretching toward the end of the decade. The organizations that treat this as a forced capacity swap will do it twice; the ones that treat it as the on-ramp to a governed Fabric estate, with OneLake, Direct Lake, and Copilot in view, will size once and build on it. Your renewal date is the trigger, but the destination is a single analytics platform, not a like-for-like replacement.

Plan your migration before renewal

If your Power BI Premium renewal is inside the next two or three quarters, the time to model the move is now, not in the grace period. Book a BabyBots Fabric migration assessment and we will map your capacity, baseline your CU consumption, and hand you a right-sized F SKU plan and a wave-by-wave cutover sequence tied to your agreement date, in a single working engagement.

Sources

• Microsoft Learn — Power BI Premium P SKU to Fabric F SKU migration decision guide
• Microsoft Learn — Power BI Premium to Microsoft Fabric migration FAQ
• Microsoft Licensing — Important update to Power BI Premium licensing (March 14, 2024)
• Microsoft Fabric Community — Important update coming to Power BI Premium licensing
• EPC Group — Power BI Premium F64 vs PPU 2026
• Power BI Consulting — Power BI Premium P1/P2/P3 to Fabric F-SKU Migration Runbook 2026