💁 About this article
This article is part of Fugo’s Power BI knowledge base: a collection of resources answering common Power BI questions. We include notes throughout where Fugo’s integration may be helpful for displaying dashboards on digital signage.
Table of contents
If you’re coming to Microsoft Fabric fresh, the licensing can feel circular.
You see Fabric capacities (F2 through F2048), then Power BI Pro still shows up as “required,” and then Power BI Premium is mentioned even though it’s supposedly been replaced. On paper, it looks like overlapping products.
The confusion comes from assuming these licenses compete with each other. They don’t. They stack.
Power BI Pro: who is allowed to create and manage content
Power BI Pro is neither infrastructure nor capacity. It’s a user entitlement.
Any time a real person is doing one of the following, Power BI Pro is involved:
Creating reports
Publishing to a shared workspace
Managing datasets or semantic models
Collaborating with other users
Even in a fully Fabric-backed environment, Pro does not disappear. It simply becomes scoped to authors and power users, rather than everyone who consumes content.
This is why you’ll still see Pro required even when:
You’re on Fabric
You’re on Premium-style capacity
You’re well past the “small team” phase
Pro answers the question: Who is allowed to actively work inside Power BI?
Cost: Power BI Pro
$14 USD per user / month (standard list price)
That pricing model scales cleanly when:
the number of creators is small
most users actively collaborate
reports are consumed interactively, not broadcast
It scales poorly when:
hundreds of users only need to view
dashboards are read-only
reports are meant to be widely visible
This mismatch is what drives organizations toward capacity-based licensing later on.
Power BI Premium shifts from user licenses to shared compute
Power BI Premium introduced a different model entirely. Instead of licensing people, you license compute.
Premium assumed:
a small number of authors
a large number of viewers
predictable, centralized workloads
Premium didn’t eliminate Pro, it just changed who needed it. Viewers could be free users. Creators still needed Pro.
Cost: Power BI Premium (historical context)
Power BI Premium was priced as a fixed monthly capacity, regardless of how many people viewed reports.
P1: ~$4,995 USD per month (list price, now retired for new customers)
That made it cost-effective only once viewer counts crossed a certain threshold. Below that point, per-user Pro licenses were cheaper. Above it, Premium quickly won.
This mental model still applies today even though Premium SKUs themselves are no longer the primary path forward.
Microsoft Fabric: Premium capacity expanded beyond Power BI
Fabric did not introduce a new licensing layer on top of Power BI. It replaced Premium capacity and extended it across analytics workloads.
What Fabric actually changed:
Power BI became one workload among many
capacity sizing became more granular
pricing aligned to shared compute, not BI alone
From a Power BI perspective, Fabric behaves almost exactly like Premium with one critical boundary - the behavior of Power BI changes materially at F64.
Why F64 is the licensing breakpoint that matters
Below F64:
Fabric provides compute
Power BI still enforces per-user viewing licenses
report viewers need Pro
At F64 and above:
Power BI becomes fully capacity-backed
report viewers can use free licenses
Pro is limited to creators and admins
Cost: Fabric capacity and the F64 trade-off
Fabric capacity is priced by SKU, billed monthly (or hourly if pay-as-you-go).
Representative list prices:
F32: ~$4,194 USD per month
F64: ~$5,834 USD per month
Below F64, Fabric reduces infrastructure friction but does not reduce licensing costs for viewers. Report viewers still need $14/month Pro licenses.
At F64, viewer licensing effectively disappears.
This is why organizations often find that F32 + hundreds of Pro licenses costs more than F64 + a small number of Pro licenses even though F64 looks expensive in isolation.
Premium Per User: extending Pro without changing the model
Premium Per User exists to fill the gap between:
Pro-only teams
full capacity deployments
It gives individual users access to Premium features without committing to capacity.
What it does not change:
everyone accessing that content still needs PPU
licensing still scales linearly with people
Cost: Premium Per User
Premium Per User is priced per user, per month.
PPU costs more per user than Pro, but avoids upfront capacity spend.
specialist teams
transitional phases
feature testing
It rarely makes sense as a long-term, org-wide strategy.
Licensing explains access, not visibility
Fabric, Pro, and capacity licensing determine:
who can create
who can view
how workloads are executed
They do not determine:
whether dashboards run unattended
whether content appears on TVs
whether users are logged in
whether displays recover after a reboot
This is why licensing decisions often feel “finished” until teams ask how dashboards actually reach people.
At that point, delivery mechanisms matter:
Power BI in PowerPoint for meetings
Embedded analytics for applications
Digital signage platforms for persistent displays
None of this tells you how dashboards reach people
One subtle but important point: licensing only governs access, not delivery.
Fabric, Pro, and Premium decide:
Who can create
Who can view
Under what computational constraints
They do not decide:
Whether a dashboard runs unattended
Whether it appears on a TV
Whether it survives a laptop going to sleep
Whether anyone is logged in at all
That’s why teams often feel “done” with licensing and then immediately hit a second wall: Okay, but how do we actually put this in front of people?
That’s where things like:
Embedded analytics (application environments)
start to matter - but those are delivery layers, not licensing layers.
The practical takeaway
If you strip away the product names, the system looks like this:
Power BI Pro → human permission to work in Power BI
Fabric capacity → shared compute for analytics workloads
F64+ capacity → license-free viewing at scale
Once you understand that separation, the documentation starts to make sense, even when it’s badly written.