Power BI Embedded Pricing 2026: What Each A SKU Costs, and What It Leaves Out

Power BI Embedded is Microsoft's way of putting Power BI reports inside an application you build, and it is priced unlike the rest of Power BI: you rent Azure capacity by the hour instead of buying a licence per person.

Microsoft's prices on this page were read from Azure's Power BI Embedded pricing page (Central US, US dollars), the Power BI pricing page and Microsoft Learn's capacity and SKU guide on 16 October 2026.

What does Power BI Embedded cost per node?

Each A SKU (Microsoft's name for a capacity size) is a node, a server of fixed size, billed for every hour it exists. Azure computes usage to the second and bills it hourly, so a node that runs for 12 hours and 15 minutes shows 12.25 hours on the invoice.

A7 and A8 exist in Microsoft's SKU table but show no price on the Azure page. Microsoft notes that prices are estimates, vary with your agreement and the exchange rate, and are shown here for Central US. Check your own region in the Azure pricing calculator before you budget.

Memory is the column that usually decides the size. Microsoft sets a node's memory as the limit for a single item, such as one semantic model, so a model (the dataset behind your reports) larger than that will not load on the node however quiet the traffic is.

Which licences does Power BI Embedded still need?

Two groups of people touch an embedded deployment, and Microsoft licenses them differently.

• Publishers need Power BI Pro or Premium Per User. That is $14 or $24 a user a month, paid yearly. Microsoft does allow publishing without a licence through a service principal (an app account rather than a person) calling the REST API, which suits an automated pipeline better than an analyst.
• Your customers need nothing, provided you embed for them. Microsoft calls this "app owns data": your application authenticates to Power BI, and the people using your app never sign in to Microsoft.
• Your own staff are a different case. Embedding for your organisation ("user owns data") on an A SKU, or on any F SKU below F64, requires a Pro or Premium Per User licence for every person who views the content.

That third point catches teams who buy Embedded for an internal portal. The capacity is cheap next to Premium, and then every viewer still needs a $14 seat.

A, EM, P or F SKU: which one should you buy for embedding?

Microsoft sells embedding capacity four ways, and they are not interchangeable.

The sizes line up: A1 has the cores of an F8 or EM1, and A4 matches F64 and the retired P1. P SKUs are no longer sold to new customers, which leaves A for embedding into your own product and F for anything that also needs the wider Fabric platform. The Power BI pricing guide covers the P-to-Fabric change in detail.

Can you pause Power BI Embedded to cut the bill?

Yes, and it is the largest saving available, with one condition attached. Microsoft's own FAQ is plain about it: when the service is paused the embedded content does not load, and you are not charged.

That makes pausing a good fit for an internal tool used in office hours and a poor fit for anything a customer opens. A customer-facing dashboard has no closing time, so most product teams run their node continuously and pay the monthly figure.

What does a Power BI Embedded deployment cost a month?

Two worked examples, using the Central US prices above. Neither includes the developer time to build the integration, which is often a large part of the first year's cost.

An internal portal on an A1, running ten hours a weekday. Pausing the node outside working hours gives about 220 hours a month.

• A1 capacity: 220 hours at $1.0081 = $221.78
• One Pro licence for the analyst who publishes: $14
• Pro licences for every viewer, because this is "embed for your organisation": $14 each
• With 30 staff viewing: $221.78 + $14 + (30 × $14) = $655.78 a month

A dashboard inside a SaaS product, always on. Customers can open it at any hour, so the node never pauses.

• A1 capacity, continuous: $735.91
• Two Pro licences for the people who publish: $28
• Customer viewers: no licence needed
• Total: $763.91 a month, or $5,909.25 if your models need the memory of an A4

The jump from A1 to A4 is the number to plan for. A1's 3 GB holds a modest model, and a product that grows its customer base usually grows its data with it.

What Power BI Embedded does not include

The capacity price buys somewhere for reports to run. Three costs sit outside it.

• The integration itself. Embedding for customers is built with Microsoft's JavaScript or .NET SDKs or its REST APIs, and someone has to write and maintain that code: authentication, embed tokens, and how reports appear inside your app.
• The reports. They are still built in Power BI Desktop by someone who knows it, including the DAX measures and any row-level security roles.
• Sizing. Microsoft gives peak renders per hour as guidance rather than a limit. As traffic approaches a node's ceiling, reports load more slowly, and choosing when to scale up is your job.

How Mammoth prices embedded dashboards

Mammoth Analytics takes the opposite approach to capacity. There is no node to size or pause: embedding is included on every plan, Free included, and viewers are free and unlimited on all of them. You pay for the people who build, from $39/month billed annually. Showing each customer only their own rows, the SaaS case above, needs the Mammoth Pro plan at $499/month billed annually, still with no node to run and no viewer licences.

A published Mammoth dashboard embeds three ways, chosen per dashboard:

• A public link, framed on any page you allow.
• An embed key for a private dashboard, where everyone who can see the page sees the same data.
• A signed token from your own server, for a dashboard inside your product. Your app signs a short-lived token for each viewer, so Mammoth never needs a login of its own, and on the Mammoth Pro plan the token carries which rows that viewer may see.

You can limit each dashboard to a list of sites, and a small script handles sizing, filters and page changes from the host page. Embedded analytics in Mammoth walks through all three, and Mammoth's pricing compares the plans.

When Power BI Embedded is the better choice

Mammoth embeds dashboards built in Mammoth, not Power BI reports running as they are, so the comparison holds when you are choosing where to build.

Power BI Embedded is the stronger choice when:

• You already have a library of Power BI reports and semantic models, and the people who maintain them.
• You need reports inside Teams, SharePoint or PowerPoint, which Microsoft supports natively.
• Your developers want full control of the embedded experience through Microsoft's SDKs.
• You are already committed to Azure or Fabric, and capacity sits on an existing agreement.

If none of those apply, and the goal is to put a dashboard in front of customers or clients without sizing capacity or licensing viewers, the per-node model is a cost you do not have to take on.

Frequently Asked Questions

Is there a free version of Power BI Embedded?

Only for development. With a Pro licence you can use free embed trial tokens to test, but production needs a purchased capacity, and until you buy one a "Free trial version" banner appears on every embedded report.

Do customers need a Power BI licence to view embedded reports?

No, when you embed for your customers. Microsoft states there are no licensing requirements for end users in the "app owns data" scenario. Embedding for your own organisation is different: below F64, every viewer needs Pro or Premium Per User.

What is the difference between an A4 and an F64?

They have the same eight cores. F64 is bought through Fabric, supports secure URL embedding from the Power BI service, and lets free Power BI users view shared content. A4 does neither, but it bills hourly with no commitment and can be paused.

How often can you change the node size?

As often as you need. Microsoft lets you change the node type at any time, and Azure bills each hour at the size that was running.