Power BI Embedded Pricing, A SKU vs F SKU Cost and Licensing
Every A and F capacity priced from Azure's own price list, side by side, with the two numbers most guides skip: when pausing beats a reservation, and how many internal viewers it takes before F64 is cheaper than Pro seats.
Sources: Azure retail price list, Microsoft Learn · read 30 September 2026
Ask the demo shop a question, or type your own:
Reading the schema, writing SQL… Writing SQL… Running (read‑only)… SQL AgentSQL wrote
▋
A real MySQL shop: customers, products, orders and signups over the last year. AgentSQL writes the SQL, runs it read-only, and answers.
Direct answer
Power BI Embedded is priced per hour of capacity, never per viewer. On Azure's price list the smallest A SKU, A1, costs $1.0081 an hour, $735.91 for a 730-hour month, and A4, the compute equivalent of F64 and P1, costs $8.0565 an hour, $5,881.25 a month. Fabric F SKUs can embed the same content at $0.18 per capacity unit an hour, or at $5,002.67 a month for F64 on a one-year reservation. Customers of your app need no Power BI license; your own staff each need a $14 Pro seat unless the capacity is F64 or larger. Authors always need Pro.
Cheapest production capacity
$735.91
A1, a full month, pausable
A SKU vs F pay as you go
30% less
same compute, every tier
Free internal viewers from
F64
$5,002.67 a month reserved
›_ power bi embedded pricing, the A SKU list
Six A SKUs, priced by the hour since 2017.
Azure retail price list, service "Power BI Embedded", US East. West US 2 and Central US read identical.
Microsoft's pricing page for Power BI Embedded sends you to the Azure calculator, which renders its numbers in JavaScript. The same figures sit in Azure's public retail price feed, and that feed shows something the calculator does not: the effective date on every A SKU meter is 1 October 2017. These rates have not changed in nine years, while Microsoft has repriced and renamed almost everything around them.
| SKU | Power BI v-cores | Fabric equivalent | Per hour | Per month, always on |
|---|---|---|---|---|
| A1 | 1 | F8 | $1.0081 | $735.91 |
| A2 | 2 | F16 | $2.0081 | $1,465.91 |
| A3 | 4 | F32 | $4.0242 | $2,937.67 |
| A4 | 8 | F64 | $8.0565 | $5,881.25 |
| A5 | 16 | F128 | $16.1210 | $11,768.33 |
| A6 | 32 | F256 | $32.2506 | $23,542.94 |
Monthly figures are the hourly rate times 730, the hour count Azure bills as a month. Microsoft Learn also lists A7 and A8 as equivalents of F512 and F1024, available only in some regions; they have no US East retail meter. EM SKUs are sold through Microsoft 365 licensing on a yearly commitment and do not appear in Azure's retail feed.
›_ power bi embedded cost, three ways to pay
The same compute costs three different amounts.
Monthly cost, always on, for each A SKU and its Fabric twin. Fabric at $0.18 per CU-hour pay as you go, or $938 per CU a year reserved.
At every tier the order is the same. Fabric pay as you go is the most expensive way to buy the compute, the A SKU is about 30 percent cheaper than that, and a one-year Fabric reservation is about 15 percent cheaper again.
The reservation only wins if the capacity runs all month. An A1 costs less than an F8 reservation whenever it runs under 620 hours, which is about 85 percent of the month. Pause it overnight and on weekends and the A SKU is the cheap option by a wide margin.
A reservation also locks you in for a year. The A SKU has no commitment and can be scaled up or down in minutes, which matters while you are still learning how much load your embedded reports generate.
›_ power bi embedded license rules
Who still needs a Pro license on embedded capacity.
From Microsoft Learn, "Capacity and SKUs in Power BI embedded analytics". The capacity pays for compute; it does not always pay for viewers.
| Scenario | Who views | Capacity it runs on | Viewer license |
|---|---|---|---|
| Embed for your customers (app owns data) | External users of your own app | Any A, F, EM or P SKU | No Power BI license for end users |
| Embed for your organization (user owns data) | Your own staff, signed in with Entra ID | Any A, F, EM or P SKU | Pro or PPU per viewer, unless the capacity is F64 or larger, or a P SKU |
| Teams, SharePoint and PowerPoint | Your own staff inside Microsoft 365 | Any A, F, EM or P SKU | Pro or PPU per viewer, unless F64 or larger, or a P SKU |
| Secure URL embedding from the Power BI service | Your own staff | F, EM or P only, not A | Pro or PPU per viewer below F64 |
Software vendors
Your customers are free to view
In the app owns data pattern your application authenticates with a service principal, and Microsoft states there are no licensing requirements for end users. An A1 serves every customer you have, however many that is, until the load outgrows one v-core.
Internal teams
Below F64, every viewer pays
Embedding into SharePoint, Teams or an intranet page for your own people does not dodge per-user licensing. On an A SKU or anything under F64, each viewer needs Pro at $14 or Premium Per User at $24 a month on top of the capacity.
Everyone
Authors always need Pro
Publishing to a workspace takes a Pro or PPU license. The one exception Microsoft documents is a service principal publishing through the REST API, which suits automated pipelines rather than analysts building reports by hand.
›_ power bi embedded for internal users
About 305 viewers before F64 pays for itself.
A1 always on plus a $14 Pro seat per viewer, against one F64 on a one-year reservation with free viewers. Authors are needed on both sides, so they cancel out.
| Internal viewers | A1 plus Pro seats | F64 reserved | Cheaper |
|---|---|---|---|
| 25 | $1,085.91 | $5,002.67 | A1 plus Pro seats |
| 100 | $2,135.91 | $5,002.67 | A1 plus Pro seats |
| 250 | $4,235.91 | $5,002.67 | A1 plus Pro seats |
| 305 | $5,005.91 | $5,002.67 | Break-even |
| 500 | $7,735.91 | $5,002.67 | F64 reserved |
The arithmetic is short. F64 reserved costs $5,002.67 a month, A1 costs $735.91, and the gap of $4,266.76 buys 304.8 Pro seats at $14. So up to about 305 internal viewers, the small capacity plus seats is cheaper; past that, F64's free viewing wins.
Swap A1 for an F8 reservation at $625.33 and the break-even moves to about 313 viewers. Most viewers already holding Pro through Microsoft 365 E5 changes the whole calculation, so count who actually lacks a seat before buying anything.
The same threshold shows up on our Power BI pricing page, which runs the break-even on Pro against F64 alone and lands at 357 read-only users.
›_ honest fit
When embedding Power BI is right, and when your staff just need answers.
Embed Power BI if
You sell software and want polished, branded dashboards inside your product for your customers. That is what Power BI Embedded was built for, and on the app owns data pattern an A1 at $735.91 a month, paused outside your customers' working hours, is one of the cheapest ways to ship interactive reports to thousands of external users.
Look elsewhere if
The plan is to embed reports in an intranet so managers stop emailing the data team. Below F64 every one of those managers needs a Pro seat, someone has to build and maintain each report, and the question they ask next week will not be on it. You would be paying capacity plus seats to answer yesterday's questions.
That second case is the one AgentSQL is for. It connects read-only to your Postgres, MySQL, Snowflake or BigQuery database, turns a question typed in plain English into SQL, runs it, and returns a chart, a table and a one-line answer with the SQL shown beside it. Nobody builds a report first. The Team plan is $59 a month billed yearly for up to 10 seats, which is less than an A1 runs in three days and less than five Pro licenses.
We do not embed dashboards in your product, so if customer-facing analytics is the goal, keep Power BI Embedded on the list. If the goal is staff answering their own questions, try the console at the top of this page on a real question first. Our self-service BI software guide lays out the other options.
›_ how to size power bi embedded
Four steps to a capacity bill you can defend.
01
Name the viewers
External customers or your own staff. That one answer decides whether viewers are free or $14 each below F64.
02
Start on A1, measured
Run real reports under real load and watch the capacity metrics. Scaling an A SKU up takes minutes and has no commitment.
03
Count the idle hours
If the capacity can be paused more than about 15 percent of the month, stay on A. If it must run around the clock, reserve F.
04
Add the authors
Budget a Pro seat for every person who publishes, on any capacity. Five report builders add $70 a month.
Government buyers have one more constraint: Microsoft Learn notes that Azure F SKU capacities are not supported in the GCC environment, where only EM and P SKUs are available, while GCC High and DoD do support them.
›_ frequently asked
Power BI Embedded pricing questions, answered.
How much does Power BI Embedded cost?
Power BI Embedded is billed by the hour for capacity, not per user. Azure's price list shows the smallest A SKU, A1, at $1.0081 an hour, which is $735.91 for a full 730-hour month, and A4 at $8.0565 an hour, or $5,881.25 a month. The Fabric equivalents run $0.18 per capacity unit an hour, or less on a one-year reservation.
Is Power BI Embedded free?
No. You can build and test with free embed trial tokens under a Pro license, but Microsoft limits those to development, and embedded reports show a "Free trial version" banner until a capacity is bought. Production embedding always needs an A, F, EM or P capacity, and the people who publish content still need Pro or Premium Per User.
Do users need a Power BI license to view embedded reports?
It depends on who the viewers are. When you embed for your customers, the app owns data pattern, end users need no Power BI license at all. When you embed for your own staff, every viewer needs Pro or Premium Per User unless the capacity is F64 or larger, or a legacy P SKU. A SKUs never remove that requirement.
What is the difference between Power BI Embedded and Power BI Premium?
Power BI Embedded is an Azure service sold as A SKUs, billed hourly with no commitment, built for software vendors embedding reports in their own apps. Power BI Premium is a Microsoft 365 subscription sold as P and EM SKUs with monthly or yearly commitments. Microsoft is retiring P SKUs and steers new buyers to Fabric F SKUs, which cover both uses.
Can you pause Power BI Embedded?
Yes. A SKUs can be paused and resumed, and billing stops while the capacity is paused, which is the main reason to pick A over a reservation. Paused, embedded reports stop rendering, so this suits development, testing and business-hours apps. An A1 run 10 hours a day on 22 weekdays costs about $221.78 a month.
Which is cheaper, an A SKU or an F SKU?
Pay as you go, the A SKU is about 30 percent cheaper than the F SKU with the same compute: A1 at $735.91 a month against F8 at $1,051.20. A one-year F reservation flips it, F8 drops to $625.33, about 15 percent under A1. Keep A if you pause it more than roughly 15 percent of the month.
Is Power BI Premium capacity being retired?
Yes, the per-capacity P SKUs are. Microsoft's Fabric licensing documentation says it is retiring the Power BI Premium per-capacity SKUs and that new and existing customers should consider Fabric F SKUs instead; P capacities remain only for customers on an active Enterprise Agreement. F64 is the compute equivalent of P1 and A4.
›_ power bi costs
›_ other bi pricing