Microsoft’s Power BI licensing page lists capacities, SKUs, and per-user tiers in a table only a licensing specialist could love. Here’s the version for someone who just needs to buy the right thing — and the short answer up front:
Most small and mid-sized businesses only need Power BI Pro. Premium exists for specific, identifiable reasons. If none of them apply to you, you’re done — buy Pro at ~$14/user/month and spend the savings on building better reports.
Now the longer answer, so you can check whether you’re the exception.
The three tiers in plain English
- Power BI Pro (~$14/user/month). The standard licence. Everyone who publishes or views shared reports needs one. Full modeling, DAX, scheduled refresh (8×/day), sharing, apps, row-level security — the complete core product.
- Power BI Premium Per User, PPU (~$24/user/month). Pro plus bigger datasets (100 GB vs 1 GB), 48 refreshes/day, paginated reports, and some advanced features. The catch: every consumer of PPU content also needs PPU, so it prices like Pro-plus-70% across the audience.
- Premium / Fabric capacity (from ~$300/month for F2, ~$5K+/month at P1-equivalent). You rent dedicated compute instead of licensing every user. The historically famous benefit: at F64 capacity and above, report viewers no longer need Pro licences — viewing becomes free at any scale.
The actual decision factors
1. How many people view reports?
This is the big one, because it’s pure arithmetic. Viewers × $14 vs capacity cost:
| Report viewers | Pro cost/month | Cheapest sensible option |
|---|---|---|
| 10 | ~$140 | Pro |
| 50 | ~$700 | Pro |
| 250 | ~$3,500 | Pro (probably) |
| 500+ | ~$7,000+ | F64 capacity starts winning |
The crossover where free viewing beats per-user licensing sits in the several- hundred-viewer range. Below that, viewer count alone never justifies Premium.
2. How big is your data model?
Pro caps a published dataset at 1 GB compressed. That sounds small; it isn’t. Power BI’s compression is excellent — a well-modeled 1 GB dataset can hold tens of millions of rows. In our experience, models that blow the 1 GB cap usually have a modeling problem (unneeded columns, wrong grain, imported data that should be aggregated) rather than a licensing problem. Fixing the model is cheaper than buying Premium — it’s a routine part of our Power BI consulting work.
Genuinely large models — billions of rows, years of transaction-level history — do need PPU or capacity. You’ll know if that’s you.
3. Do you need more than 8 refreshes a day?
Pro refreshes 8 times daily; PPU/Premium allow 48 and near-real-time options. Most businesses running daily or hourly reporting are fine with 8. If you need data fresher than hourly, look at PPU first — it’s the cheap way to get 48.
4. Do you need to share outside your organization… at scale?
External guest users need licensing too. A handful of guests: Pro handles it. Publishing dashboards to hundreds of external partners or customers: that’s capacity territory (or “embedded” licensing — a topic for another post).
5. Paginated reports?
Pixel-perfect, printable, invoice-style reports (the old SSRS format) now work on Pro as well, with some limits. Heavy paginated workloads still sit more comfortably on PPU/capacity, but this is no longer the Premium-only feature it once was.
The upgrade path that actually makes sense
- Start with Pro for everyone. It’s the default for a reason.
- Hit a real wall — dataset size, refresh frequency — and move just the affected workload or users to PPU. It’s an incremental $10/user, not a $5K/month platform decision.
- Consider capacity (F64+) only when viewer count climbs into the hundreds or you’re embedding reports in customer-facing products.
Microsoft’s Fabric push has made small capacities (F2, F4…) cheap enough to experiment with, but note the fine print: below F64, viewers still need Pro licences — small Fabric capacities buy features and compute, not free viewing. That detail catches a lot of buyers.
The bottom line
Don’t let licensing anxiety stall a reporting project: the worst case is you start on Pro, outgrow something specific, and upgrade the specific thing. The money mistake isn’t choosing the wrong tier — it’s buying Premium “to be safe” when a $14 licence and a well-built data model would have done the job.
If you’re weighing this for a real project, book a free 30-minute call — we’ll look at your data size, audience, and refresh needs and tell you exactly which licences to buy (we don’t resell them, so we have no reason to upsell you).