Este artículo actualmente está disponible solo en inglés. La traducción estará disponible próximamente.
Power BI Premium vs Pro vs Embedded: Licensing Decision Matrix
Microsoft's Power BI licensing has six SKUs, each with different features, audiences, and price models. The wrong choice means you either pay $5,000/month for capacity you do not need, or you blow up at 100 users on a Pro license that cannot deliver. This article is the decision matrix we use with ECOSIRE clients to right-size Power BI licensing in one meeting.
We will walk through Power BI Pro, Premium Per User (PPU), Premium capacity (now Fabric F-SKUs), Power BI Embedded (A-SKUs), and what changes when Fabric gets involved. Numbers are list prices as of the Power BI Desktop 2026-01 release and Microsoft Fabric Q1 2026 pricing. Real prices vary with EA discounts and CSP partner pricing — your account team can usually shave 10-25% off list.
Key Takeaways
- Power BI Pro ($14/user/month): the baseline. Required to share content. Buy this for every author and every viewer in a small/mid org without Premium.
- Premium Per User (PPU) ($24/user/month): unlocks AI features, paginated reports, deployment pipelines, larger model size — for power users only, not viewers.
- Premium capacity / Fabric F-SKUs (~$5,000/month entry, F64): per-org capacity that lets unlimited Free users view content. Crossover point: ~300+ viewers.
- Power BI Embedded (A-SKUs) ($735/month entry, A1): for ISVs embedding analytics in their own SaaS. Pay per capacity, not per user.
- Microsoft Fabric absorbs Premium capacity and adds OneLake, Data Factory, Synapse — same price tier, more value, but more services to learn.
- The decision boils down to: how many viewers do you have, do you need the AI/large-model features, and are you embedding in your own product?
- Fabric F2 ($263/month) is too small for production Power BI workloads — minimum F64 for unlimited viewers.
- Embedded A-SKUs do NOT support self-service authoring in the Service — they are for app embedding only.
The Six SKUs in One Sentence Each
- Power BI Free: Desktop authoring + your own My Workspace. Cannot share with others. Useful for personal exploration.
- Power BI Pro: All Free features + ability to share content in workspaces with other Pro users. The default per-user license.
- Premium Per User (PPU): Pro features + Premium-only capabilities (paginated reports, AI, larger datasets, deployment pipelines). Audience must also have PPU.
- Premium Capacity (F-SKUs): Org-level capacity on Microsoft Fabric. Unlimited Free users can view Premium content; authors still need Pro.
- Power BI Embedded (A-SKUs): Capacity for embedding analytics in your own application. No interactive Service consumption.
- Microsoft Fabric: Superset of Premium that adds OneLake, Data Factory, Synapse, Real-Time Analytics — same capacity model.
Detailed Comparison Matrix
| Feature | Free | Pro | PPU | F64 (Fabric) | A1 Embedded |
|---|---|---|---|---|---|
| Monthly cost | $0 | $14/user | $24/user | $5,002 | $735 |
| Power BI Desktop authoring | Yes | Yes | Yes | n/a | n/a |
| Share content with others | No | Yes | Yes | Yes (Free viewers OK) | API only |
| Maximum dataset size | 1 GB | 1 GB | 100 GB | 400 GB | 3 GB (A1), scaling |
| Refresh frequency | 8/day | 8/day | 48/day | 48/day | 48/day |
| Paginated reports | No | No | Yes | Yes | Yes |
| AI capabilities (auto-ML, smart narrative) | Limited | Limited | Yes | Yes | Yes |
| Deployment pipelines | No | No | Yes | Yes | No |
| Free viewers can consume | No | No | No | Yes | n/a |
| Embedded in apps | No | Yes (Pro pricing) | Yes | Yes | Yes (primary use) |
| XMLA endpoint (read/write) | No | No | Yes | Yes | Yes |
| OneLake / Fabric features | No | No | No | Yes | No |
| RLS support | Yes (Desktop) | Yes | Yes | Yes | Yes |
| Sensitivity labels | No | Yes | Yes | Yes | Yes |
| App workspace sharing | No | Yes | Yes | Yes | API |
Power BI Pro: The Baseline
Pro is the entry point for any organization that wants to share Power BI content. Every author needs Pro. Every viewer needs Pro (unless the dataset is in a Premium capacity).
Buy Pro when:
- You have 1-300 total users (authors + viewers).
- You do not need datasets above 1 GB compressed.
- You do not need AI features, paginated reports, or deployment pipelines.
- You can refresh up to 8 times per day.
Math example: 50 users × $14/month = $700/month. This is dramatically cheaper than the F64 entry point and covers most small/mid-market needs.
Pain point: above 300 users, Pro becomes more expensive than F64 capacity, and you also lose features (no paginated reports, no AI). Recalculate at every quarterly headcount.
Premium Per User (PPU): The Author's Upgrade
PPU is the answer to "I want Premium features but I have only 5 power users." It is per-user, like Pro, but unlocks Premium-only features for that user.
Critical gotcha: anyone consuming PPU content must also have PPU (or a Pro license against Premium capacity). PPU does not act like capacity — it does not enable Free viewers.
Buy PPU when:
- You have a small number (5-30) of advanced users who need paginated reports, AI, deployment pipelines, or 100 GB datasets.
- The audience for those reports is also a small number of advanced users.
- You are not yet justified in buying F64 capacity.
Math example: 10 finance analysts × $24/month = $240/month. That same audience getting Premium features via F64 would be $5,000/month — PPU is dramatically cheaper for small audiences.
Watch out for: scope creep. Once finance loves the paginated reports, sales wants them, then operations wants them. By the time 50 users need PPU, F64 ($5K/month for unlimited) becomes the better deal at 50 × $24 = $1,200 vs $5,002. The math flips at ~200 users on PPU = ~$5,000.
Premium Capacity / Fabric F-SKUs: The Org-Wide SKU
Premium capacity (now sold as Microsoft Fabric F-SKUs) is org-level capacity. You buy a SKU sized in capacity units (CUs), workspaces are assigned to it, and content in those workspaces can be consumed by Free users (no Pro license needed for viewers).
The key Power BI features at F64:
- Unlimited Free user viewing (the headline benefit).
- 400 GB dataset size.
- 48 refreshes per day.
- Paginated reports for all users.
- AI features.
- Deployment pipelines.
- XMLA read/write.
- Embedded API entitlement.
- All Fabric workloads (OneLake, Data Factory, Synapse, Real-Time Analytics) — though you pay capacity for any usage.
Capacity sizing:
| SKU | Monthly cost | CUs | Power BI viewers feasible |
|---|---|---|---|
| F2 | $263 | 2 | Tiny dev/test |
| F8 | $1,051 | 8 | <50 viewers |
| F16 | $2,103 | 16 | <100 viewers |
| F32 | $4,206 | 32 | <200 viewers |
| F64 | $5,002* | 64 | 200-500+ viewers (Free viewers enabled) |
| F128 | $10,003 | 128 | 500-1000+ viewers |
| F256 | $20,006 | 256 | 1000+ viewers, heavy workloads |
*F64 is special because below F64 (i.e., F2-F32) the unlimited-Free-viewer entitlement does NOT apply — viewers still need Pro. F64 is the practical entry for the "free viewer" benefit.
Buy F-SKU when:
- You have 300+ viewers in the org.
- You need paginated reports across the org.
- You want Fabric (OneLake, Data Factory) in addition to Power BI.
- You are running into the 1 GB / 8-refreshes-per-day limits.
Power BI Embedded (A-SKUs): The ISV SKU
Embedded is for the SaaS company that wants to put Power BI dashboards inside its own product, branded as theirs.
Differences from F-SKUs:
- A-SKUs are billed hourly, can be paused, and start at A1 ($735/month if running 24/7).
- A-SKUs do NOT entitle anyone to use the Power BI Service interactively — they are app-embedding only.
- Authors still need Pro for development; the A-SKU is consumed at runtime by app users.
- Authentication is via service principal + embed tokens (see our RLS guide).
- Pause when not in use to save money (e.g., business hours only).
A-SKU sizing:
| SKU | Hourly | Monthly (24/7) | Notes |
|---|---|---|---|
| A1 | $1.0083 | $735 | 3 GB dataset, dev/test, light prod |
| A2 | $2.0167 | $1,470 | 5 GB dataset, small prod |
| A3 | $4.0333 | $2,940 | 10 GB, medium prod |
| A4 | $8.0500 | $5,876 | 25 GB, larger prod (≈ F64 features) |
| A5 | $16.10 | $11,752 | 50 GB |
| A6 | $32.20 | $23,503 | 100 GB |
Buy A-SKU when:
- You are building embedded analytics inside your own SaaS for external customers.
- Customers should NOT need a Power BI license.
- You can pause capacity outside business hours.
Real-world example: ECOSIRE runs A2 capacity for an Odoo customer portal that serves ~150 of their clients. Capacity is paused weekends, costs ~$900/month, and supports unlimited per-tenant dashboards with embedded RLS.
Microsoft Fabric: The Superset
Fabric replaces Power BI Premium capacity SKUs (P-SKUs were retired) with F-SKUs that include the same Power BI Premium entitlement plus all Fabric workloads. If you buy F64, you get Power BI Premium AND OneLake AND Data Factory AND Synapse Data Warehouse AND Real-Time Analytics — all on the same capacity.
The implication: if you were going to buy Premium for Power BI anyway, Fabric is a free upgrade in features. You may not use OneLake or Data Factory immediately, but you have them when you need them.
The trap: more workloads on the same capacity = more risk of throttling. F64 has 64 capacity units total. Power BI alone does not max it; Power BI + heavy Data Factory pipelines + Synapse warehouse queries can. Monitor capacity metrics ruthlessly or step up to F128.
Decision Tree
Use this sequence:
- Are you embedding in your own SaaS product? Yes → A-SKU. Done.
- How many users (authors + viewers)?
- 1-50 → Pro for everyone ($14/user).
- 50-300 → Pro for everyone (still cheaper than F64 below 300).
- 300+ → Probably F64 ($5,002 unlimited Free viewers + Pro for authors).
- Do power users need Premium-only features (paginated, AI, 100 GB datasets)?
- Yes, but only 5-30 power users → PPU for them, Pro for viewers.
- Yes, broad audience → F64 (Premium capacity).
- Do you need Fabric workloads (OneLake, Data Factory)?
- Yes → F-SKU sized to your usage (start F64).
- No → Stick with Pro until 300+ viewers, then F64.
Crossover Math
A useful sanity check:
- F64 ($5,002/month) ÷ Pro ($14/month) = 357 users.
- Below 357 users, Pro for everyone is cheaper.
- Above 357 users, F64 with Free viewers is cheaper, AND you get all Premium features.
- This is the right answer ignoring features. If you need paginated reports across 50 users, F64 wins regardless of math.
Operational Considerations
- EA discounts: Microsoft Enterprise Agreement customers typically get 10-25% off list. Don't pay sticker.
- Annual commit on Fabric: F-SKUs sold via annual reservation can be ~40% cheaper than pay-as-you-go monthly.
- Capacity monitoring: deploy the Microsoft Fabric Capacity Metrics app to all F-SKU workspaces. Set alerts on overage.
- Right-sizing rhythm: review SKU once a year, or after 25%+ user growth.
- Don't mix Embedded and Premium: one capacity for embedded SaaS, another for internal analytics. Capacity bleed-over creates operational pain.
Common Mistakes
- Buying F64 for 50 viewers: 50 × $14 Pro = $700/month vs $5,002 F64. F64 is wasted money below 300 users (unless you need Fabric workloads).
- Buying PPU for the whole org: PPU does not enable Free viewers. 100 × $24 = $2,400/month and the audience still needs PPU. F64 unlimited at $5,002 is cleaner above 200 users.
- Using A-SKU for internal users: A-SKUs are for embedded apps. Internal users should be on F-SKU.
- Forgetting the Pro license for authors on F-SKU: F-SKU enables Free viewing, but authors who publish content still need Pro. Budget accordingly.
- Not pausing A-SKU off-hours: an A4 paused 12 hours/day saves $2,900/month.
Frequently Asked Questions
What happened to P-SKUs?
P-SKUs (Premium capacity, e.g., P1, P2, P3) were retired in 2026. Microsoft Fabric F-SKUs are the replacement, with the same Power BI features plus the broader Fabric stack. Existing P-SKU customers were migrated automatically.
Can I buy just Power BI without Fabric?
Pro and PPU are still available standalone. Premium capacity is now sold as F-SKUs (Fabric capacity), but you can use it for Power BI only and never touch the other Fabric workloads. You pay for capacity regardless of which workloads you use.
Is there a free tier for embedded?
No. Embedded is exclusively the A-SKU paid model. Pro licenses cannot be used to embed in third-party apps for unlicensed users — Microsoft's terms require A-SKU for that scenario.
How does Fabric capacity differ from compute units in cloud SQL?
Fabric capacity is a "smoothed" pool of compute units. Power BI rendering, dataset refreshes, Data Factory pipelines, and Synapse queries all draw from the same pool. Microsoft applies "smoothing" so a brief spike does not throttle, but sustained over-use will. Monitor with the Capacity Metrics app.
Where can I get help right-sizing our Power BI license?
ECOSIRE sizes Power BI licensing as part of every deployment. We model your user count, workload mix, and growth trajectory to pick the SKU that delivers without overspend. Talk to our Power BI consultancy team or browse Power BI products for license-aware templates. For embedded setups, see our embedded analytics service.
Power BI licensing rewards careful right-sizing. The penalties for getting it wrong — overspend on F64 below 300 users, or feature starvation with Pro above 1000 — are large but visible. Use the decision tree above before any Power BI rollout, recheck every 12 months, and you will keep the bill aligned to actual value delivered.
Escrito por
ECOSIRE TeamTechnical Writing
The ECOSIRE technical writing team covers Odoo ERP, Shopify eCommerce, AI agents, Power BI analytics, GoHighLevel automation, and enterprise software best practices. Our guides help businesses make informed technology decisions.
ECOSIRE
Desbloquear decisiones basadas en datos
Paneles de Power BI personalizados, modelado de datos y soluciones de análisis integradas.
Artículos relacionados
Power BI para Odoo: 12 patrones DAX listos para producción
12 patrones DAX probados en batalla para datos de Odoo en Power BI: inteligencia de tiempo, cohortes de clientes, antigüedad de inventario, pérdidas y ganancias de varias empresas y uniones de claves compuestas.
Actualización incremental de Power BI para tablas de más de 10 millones de filas
Guía de actualización incremental de Power BI para tablas de más de 10 millones de filas: diseño de particiones, RangeStart/RangeEnd, políticas de actualización, plegado de consultas e híbridos de DirectQuery.
Seguridad a nivel de fila de Power BI: patrones dinámicos frente a estáticos
Análisis profundo de Power BI RLS: roles estáticos frente a dinámicos, patrones USERPRINCIPALNAME, tablas de seguridad, jerarquías de administradores, pruebas de RLS y RLS integrado para SaaS.