Back to Blog

Power BI Features by License Type - What You Actually Get for the Money

May 31, 2026•8 min read•Michael Ridland

Power BI licensing confuses almost every Australian client I talk to. Not because the SKUs are complicated in isolation, but because Microsoft keeps shifting the meaning of "Premium" while leaving the old terminology in the documentation. People come to us having paid for Pro for two years, used about a tenth of what it can do, and now they're being told they need Fabric capacity for something they thought they already had. The conversation usually starts with "we already pay for Power BI, why is this not working?"

Microsoft's features by license type page is the canonical reference and it does a reasonable job of laying out what each tier gets you. What it doesn't do is tell you which features actually matter for an Australian mid-market business, what the gotchas are, or how to think about the upgrade path. That's the gap I want to fill here.

The four tiers that matter

There are essentially four meaningful licensing positions for Power BI today:

  1. Free - personal use, no sharing.
  2. Pro - per-user license, the standard for most knowledge workers.
  3. Premium Per User (PPU) - per-user license with capacity-tier features.
  4. Fabric capacity (F SKUs) - capacity-based, formerly known as Premium capacity (P SKUs).

There's also Power BI Embedded for ISVs and the Microsoft Fabric capacities that include Power BI workloads, but those follow the same logic as F SKUs for end-user feature parity, so I'll fold them together.

Forget about everything else you might have heard. The above four are what you're actually buying.

Free is for analysts working alone

The Free tier gives you Power BI Desktop and a personal workspace in the service. You can build reports, you can save them to My Workspace, and you can publish to the web in very limited ways. What you cannot do is share with another person inside your organisation. The moment you want to collaborate, even on a single report with one colleague, somebody needs a paid license.

We see two valid use cases for Free. The first is the solo analyst doing exploratory work that nobody else needs to see. The second is the data engineer who builds models that they hand off via .pbix files to other teams who do their own publishing. Outside those two scenarios, Free is a dead end and you'll churn through it within a week.

Worth knowing: Power BI Desktop is also free. You don't need a license to install Desktop. So an organisation can absolutely have data analysts building things in Desktop without buying anything, as long as the deliverable is the file itself rather than a shared report in the service. Some Australian regional councils we've worked with do exactly this for ad-hoc analysis.

Pro is the default and probably what you want

Pro at around AUD 14 per user per month is the workhorse. Every consumer and every author of shared content needs Pro, with the caveat that if your reports live in a Premium-backed workspace, the consumers might not need Pro. We'll come back to that.

What Pro gets you that Free doesn't:

  • Shared workspaces with role-based access.
  • App publishing to broader audiences within your org.
  • Subscriptions, sharing, embedding in Teams and SharePoint.
  • Row-level security on shared content.
  • Datasets up to 1 GB per dataset, 10 GB of storage per user.
  • Eight refreshes per day on imported datasets.

For most Australian businesses I work with under about 500 staff, Pro across the board is the right answer. It's predictable, it's per-seat, it's easy to budget for, and it covers 90 percent of the use cases. We rarely recommend anything more sophisticated until specific pain points emerge.

The pain points that push you off Pro are usually one of these: dataset sizes pushing past the 1 GB limit, refresh frequency requirements above 8 per day, paginated reports needed for operational use, deployment pipelines, or you've got a large audience of view-only consumers and the per-user pricing stops being efficient.

Premium Per User splits the difference

PPU is the awkward middle child. It costs roughly AUD 30 per user per month and gives an individual user access to most of the features that used to require Premium capacity, but only within PPU-licensed workspaces. So you can have a small team of authors on PPU and they get larger datasets, deployment pipelines, paginated reports, AI features, and more frequent refreshes, but anyone consuming their content also needs PPU.

PPU is a great fit when you have a small expert team that needs the bigger toys but you don't have an audience big enough to justify capacity. We've used it for a couple of Sunshine Coast and Brisbane clients where the analytics team is six or seven people and the consumers are also analysts who need to interact deeply with the data. Once the consumer audience gets above thirty or forty people, the maths usually shifts back toward Pro plus capacity.

The gotcha with PPU is that you can't mix it cleanly with Pro on the same workspace. A workspace is either Pro or PPU, and everyone touching it needs the matching license. Plan your workspace topology around this from day one.

Fabric capacity is where it gets interesting in 2026

Fabric F SKUs replaced the old P SKUs for new customers in 2024. The change matters because Fabric capacity isn't just Power BI capacity anymore - it's a shared compute pool that runs Power BI alongside Data Factory, Synapse engineering, real-time intelligence, and the rest of the Microsoft Fabric stack. You're buying compute, not seats.

What this unlocks for Power BI specifically:

  • Free consumption for unlicensed users (read-only) at F64 and above. This is the killer feature. You can publish a workspace from an F64 capacity and anyone in your org can view those reports without a Pro license.
  • Massive datasets - 10 GB per dataset on F2, scaling up.
  • Pageable XMLA endpoints for tools like Tabular Editor and DAX Studio against shared models.
  • Deployment pipelines, AI insights, autoscale, and most of the premium-only features.
  • Direct integration with OneLake, lakehouses, warehouses and the Fabric data engineering surface.

The F64 threshold matters because of that free-consumer feature. For an organisation with a thousand users where only fifty are authors, F64 is dramatically cheaper than Pro for every consumer. The breakeven point is roughly between three and five hundred consumers, depending on how you scale the capacity and whether you can pause it overnight.

A few honest opinions on Fabric capacity:

  • Autoscale is finicky and we usually recommend disabling it until you've got a stable usage baseline.
  • The Capacity Metrics App is essential and somehow still feels like a beta product. Make sure your platform team is comfortable reading it.
  • You can pause F SKUs to save costs out of hours, and most clients should be doing this. A capacity paused from 7pm to 6am AEST cuts your bill roughly in half if your business is purely Australian.

How to decide

The decision tree we walk clients through is pretty simple.

If you've got under 30 people who need to consume content, buy Pro for everyone and move on. Don't over-engineer this.

If you've got between 30 and 300 consumers, run the maths on whether Pro is still cheaper than F64 plus Pro for authors. Usually Pro still wins until you're closer to 300 or you have specific Premium feature requirements.

If you've got over 300 consumers or you need any of the capacity-only features (paginated reports as a key deliverable, datasets over 10 GB, XMLA write, deployment pipelines), go capacity. The capacity sizing question is its own analysis and depends heavily on refresh patterns and concurrent user count. We typically start clients on F64 or F128 and tune from there based on the metrics app data.

If you're already on Microsoft Fabric for data engineering, you've got capacity anyway, so use it. The Power BI workload uses the same compute. Just make sure you've got the workspace assignment set correctly and you understand how Fabric capacity throttling interacts with Power BI rendering.

Don't forget the non-license things

A lot of the friction I see with clients isn't actually licensing, it's the bits that licensing doesn't fix. On-premises data gateways still need to be set up regardless of license. Sensitivity labels and DLP still need configuration. Workspace governance, naming conventions, certification, and dataset lifecycle management aren't included in any license tier. You have to actually do that work.

We help a lot of Australian organisations get this foundation right when they're moving from scattered Pro usage to something more deliberate. Our Power BI consultants page covers the kinds of engagements we run, from licensing reviews through to full Fabric implementations. And if you're thinking about the broader business intelligence picture rather than just Power BI in isolation, our AI for business intelligence page lays out how we think about modern BI in an AI-augmented world.

The honest bottom line

Power BI licensing is fine once you understand the four tiers and what actually changes between them. The hard part is resisting the temptation to over-buy or under-buy based on what a sales conversation made it sound like you needed. Pro is the default. Capacity is for scale and specific features. PPU is for small expert teams. Free is for solo work. Anything else is noise.

The most expensive mistake I see is paying for Premium capacity that nobody's actually using the premium features of. The second most expensive is sticking with Pro for a 2,000-person organisation where everybody consumes reports daily, and then complaining about the bill. Both are easily diagnosed with a half-day review.

If you'd like a hand sorting out your Power BI licensing position, or you're trying to build a business case for the move to Fabric, that's exactly the conversation we have most weeks. Get in touch and we can take you through it.

Reference: Power BI features by license type - Microsoft Learn