Understanding the Power BI Embedded Capacity Workspace

Power BI Embedded Capacity and Workspace Guide

Power BI Embedded lets ISVs and enterprises embed Power BI reports in custom applications and portals without requiring end users to have Power BI licenses. It runs on Microsoft Fabric F-SKU capacities. This guide covers capacity SKUs, workspace design, multi-tenant architecture, and performance optimization for Power BI Embedded deployments.

Key facts

• Power BI Embedded uses Microsoft Fabric F-SKU capacities: F2 ($263/mo) to F64+ ($5,257/mo+).
• End users of embedded reports do not need Power BI Pro or PPU licenses — capacity covers them.
• Power BI Embedded is used by ISVs embedding analytics in SaaS products and enterprises sharing reports with external stakeholders.
• Direct Lake mode delivers near-Import-mode query performance without refresh windows — available on F-SKU capacities.
• EPC Group has completed 1,500+ Power BI deployments including Power BI Embedded implementations for ISV clients.
• EPC Group Power BI Embedded implementation: $50,000–$250,000 depending on tenant count and complexity.

Power BI Embedded — ISV vs. Enterprise Use Cases

Two distinct audiences use Power BI Embedded for different purposes.

• ISV (Independent Software Vendors) — Embed Power BI reports in SaaS products. End customers see analytics without knowing Power BI is the engine underneath.
• Enterprise (external distribution) — Distribute analytics to customers, partners, or suppliers who don't have Power BI licenses in your tenant.

Both scenarios require F-SKU capacity. The workspace and tenant architecture decisions differ significantly between the two.

Fabric F-SKU Capacity Sizing

Capacity size determines how many concurrent users, datasets, and report renders you can support. Choose based on your embedded analytics workload.

• F2 ($263/month) — Small workloads. Up to 4 GB memory. Roughly 30 reports. Good for dev/test or early-stage ISV products.
• F4 ($526/month) — Mid-market. Semantic model refresh under 10 minutes. 50–100 concurrent users.
• F8 ($1,052/month) — Growing ISV. Multiple semantic models. 150–200 concurrent users.
• F16 ($2,103/month) — Large ISV or enterprise with high concurrent user counts.
• F64 ($5,257/month) — Enterprise scale. Power BI plus Fabric data engineering, lakehouse, and real-time intelligence.

Right-size every 90 days. Microsoft adjusts F-SKU memory allocations and Direct Lake availability with each major service update.

Multi-Tenant Architecture Options

ISVs must decide how to isolate tenant data. Two patterns dominate in 2026.

Single Workspace, Multi-Tenant (Row-Level Security)

• All tenants share one workspace and one set of datasets.
• Row-level security (RLS) isolates each tenant's data at query time.
• Lower operational overhead — one workspace to manage, one dataset to update.
• Risk: weaker data isolation for regulated industries. Compliance audits may require demonstrable physical separation.
• Best for: cost-sensitive ISVs with non-regulated data and simple schemas.

Workspace-Per-Tenant

• Each customer tenant gets their own dedicated Power BI workspace and dataset.
• Physical isolation of tenant data — each dataset contains only that tenant's records.
• Higher operational overhead — workspace provisioning automation required at scale.
• Required for HIPAA, SOC 2, and regulated-industry data where isolation must be demonstrable in audits.
• Best for: healthcare, financial services, and government ISVs with strict compliance requirements.

The Noisy Neighbor Problem

When multiple tenants share a capacity, one tenant's heavy usage can degrade performance for all others. This is the "noisy neighbor" problem.

• Cause: large dataset refreshes, complex queries, or many concurrent users from one tenant consuming most capacity memory.
• Solutions: separate capacities for high-usage tenants; autoscale to absorb peaks; stagger refresh schedules; monitor per-tenant resource consumption.
• Detection: Power BI Premium Metrics app shows per-dataset and per-workspace memory consumption in real time.

Direct Lake Mode for Embedded

Direct Lake mode changes the performance economics for Power BI Embedded deployments. It is available on F-SKU capacities.

• Traditional Import mode: data is copied into Vertipaq. Refresh windows required. Stale data risk between refreshes.
• Direct Lake: semantic models query OneLake-resident Parquet files at near-Import-mode speed. No refresh window.
• Benefit for ISVs: removes the need to schedule and orchestrate dataset refreshes at scale across hundreds of tenant workspaces.
• Performance benchmark: a Fortune 500 finance organization achieved sub-800 ms query time vs. 30-minute refresh cycles with Import mode.

Frequently asked questions

What is Power BI Embedded?

Power BI Embedded is a Microsoft service that lets ISVs and enterprises embed Power BI reports in custom applications and portals. End users do not need Power BI Pro licenses. The capacity SKU covers all embedded report users.

What SKU do I need for Power BI Embedded?

Power BI Embedded uses Microsoft Fabric F-SKUs. Start with F2 ($263/month) for small workloads. Most production ISV deployments start at F4–F8. F64 ($5,257/month) is the enterprise sweet spot for high concurrent user counts plus Fabric data engineering.

What is the difference between workspace-per-tenant and RLS?

Workspace-per-tenant physically isolates each customer's data in separate datasets. RLS isolates data logically through query filters in a shared dataset. Regulated industries (HIPAA, SOC 2) typically require workspace-per-tenant for demonstrable audit isolation.

What is the noisy neighbor problem in Power BI Embedded?

When multiple tenants share a Power BI Fabric capacity, one tenant's heavy workload can consume most capacity memory and slow queries for all other tenants. The fix is per-tenant capacity assignment for high-usage customers or autoscale configuration.

Does EPC Group implement Power BI Embedded for ISVs?

Yes. EPC Group designs and deploys Power BI Embedded for ISV SaaS products and enterprise external analytics portals. We handle capacity sizing, multi-tenant architecture, Direct Lake configuration, and compliance controls. Engagements start at $50,000.

Schedule a Power BI Embedded consultation

Talk to an EPC Group Power BI architect about your Embedded capacity design, multi-tenant architecture, or ISV analytics implementation. Call (888) 381-9725 or request a 30-minute discovery call.