How Much is Fabric | Strategies for Managing and Estimating Fabric Capacity Expense

Description

Learn how to accurately estimate and control Microsoft Fabric capacity costs in today’s cloud‑driven world. We’ll cover key drivers like workload patterns, utilization, and scaling, and walk through practical models, examples, and best practices to align Fabric spend with your organization’s performance and budget goals.

Key Takeaways

• •How many of you want to use Fabric, but
• Welcome and Introductions
• How is Fabric Licensed
• What Consume Capacity
• Offers Deep Integration
• Uses Open Data Formats
• Capacity is Associated

My Notes

Action Items

• [ ]

Resources & Links

Slides

Warm Up
Questions
•How many of you are using Fabric now?
•How many of you want to use Fabric, but
don’t know how much it will cost?
•How many of your bosses are worried about
how much you are spending on Fabric?
•How many of you want to save money?
Let’s Get Fabric
in Shape
Gym Analogy
✓New Years Resolution
✓Join Gym
✓Strength Training
✓Squats
✓Bench Press
Is this the way you are
approaching Fabric Licensing?
Nathan Giullian
Founder | Birch Run Consulting
Jim Hill
Principal Business Intelligence
Engineer at BambooHR
AGENDA
• Welcome and Introductions
• What is Fabric
• How is Fabric Licensed
• What Consume Capacity
• Show me the Numbers
• Tools to Monitor
• Tips and Tricks
• Q&A
What is Fabric?
How is Fabric
Licensed?
• A Unified Analytics
Platform
• Integrates Data
Engineering, Data
Governance, Data
Warehousing, Real-time
Analytics, & Business
Intelligence
• Leverages OneLake as
Single, Centralized Data
Lake
• Offers Deep Integration
with Power BI & Other MS
Services
• Uses Open Data Formats
• Licensed by Capacity
• Capacity is Associated
with a Workspace
• Capacity can be
Assigned to More Than 1
Workspace
• Capacities can be
Increased or Decreased
by an Admin
• Capacities can be
Turned On and Off by an
Admin
Microsoft Fabric Community Conference 2025
What are the Capacity Levels?
SKU
Capacity unit (CU)
Pay-as-you-go
Reservation
F2
$262.80/month
$156.334/month
~41% savings
F4
$525.60/month
$312.667/month
~41% savings
F8
$1,051.20/month
$625.334/month
~41% savings
F16
$2,102.40/month
$1,250.667/month
~41% savings
F32
$4,204.80/month
$2,501.334/month
~41% savings
F64
$8,409.60/month
$5,002.667/month
~41% savings
F128
$16,819.20/month
$10,005.334/month
~41% savings
$33,638.40/month
$20,010.667/month
~41% savings
$67,276.80/month
$40,021.334/month
~41% savings
• Developers
need individual
license for all
SKUs
• Free End User
Licenses
• Copilot
Capabilities
F256
F512
F1024
$134,553.60/month
$80,042.667/month
~41% savings
F2048
$269,107.20/month
$160,085.334/month
~41% savings
What Activities Consume Fabric Capacity?
Power BI
Dataset refreshes
Report queries and rendering
Dataflows and data transformations
Real-time Analytics
KQL database queries
Streaming ingestion and interactive analytics
Data Engineering
Notebooks and Spark jobs
Lakehouse operations (reading, writing)
Data prep and transformations
SQL-based Workloads (Data Warehouse)
T-SQL queries on data
warehouses/lakehouses
Data modeling, indexing, and query execution
Data Factory
Pipelines, data ingestion, and orchestrations
Copy activities and data movement
Machine Learning / AI
Model training and scoring
(when run within Fabric environment)
How Do You
Know Where to
Start?
With Fabric You Have Choices
ETL Tool
Storage Option
Dataflow
Database
Notebook
Dataflow
Pipeline
Lakehouse
Warehouse
Process for
Estimating
Fabric Needs
What Does
Microsoft
Recommend?
Use the Fabric
Trial License
We wanted to do
more!
FCA open
source tool
We wanted to do
more!
Fabric Cost Analysis
How the Data was Collected
• Three Tables
• 100M
• 10M
• 1M
• ETL Tools:
• Dataflow Gen2
• Notebook
• Pipeline
• Target Destinations:
• Fabric Database
• Dataflow Gen2
• Lakehouse
• Warehouse
It’s Science!
Show me the
Numbers
in Power BI
Take Aways
• Lakehouses perform the best of the storage options
(Dataflows Gen2 do well also but aren’t as open)
• Pipelines used the least CUs, but took longer
• Notebooks were the fastest, but used more CUs if the
spark instance wasn’t terminated immediately
• Dataflows Gen2 are a good balance between speed,
cost, and technical skill
DirectLake
Report CU Consumption Example
Findings:
• Reports Have High Consumption
• DirectLake vs Import
Import
Recommendations:
• Start with Fabric Trial
• Separate ETL and Reporting
Workspaces
Recommendations
NOTE: Your Individual Results will Vary
• Start Low - Work up
• Best Performance
• ETL Tool: Pipeline then Notebook
• Storage: Lakehouse
• Turn Things Off that Aren’t in Use
• Reserved Capacity
• Monitor Usage Regularly
• DirectLake models have low CU consumption
Monitoring Fabric
Cost Saving Strategies
Questions to Ask Yourself
• Are my ETL workloads continuous or batched?
• When are users consuming reports?
• Do I need deployment pipelines?
• Is the extra effort to save costs worth the time required?
• Do you need the F64 only features?
✓ Copilot – No Pro licenses for consumers
Tips & Tricks to Manage Fabric Costs
• Separate Report Workspaces from ETL Workspaces
• Use surge protection
• Consider a separate capacity for Prod and Dev
• Consider Moving to a Reserved instead of Pay-as-you-go
licensing
• Turn Capacities on and off (for Pay-as-you-go only)
What Have We
Learned?
Summary Take Aways
• ETL Tool: Pipelines are most cost performant
• Storage: Lakehouses are most cost performant
• Leverage the Fabric Trial
• Monitor Usage in the Fabric Capacity App or FCA
tool
• Adjust regularly
Microsoft Fabric Community Conference 2025
Sound off.
The mic is all yours.
Influence the product roadmap.
Join the Fabric User Panel
Join the SQL User Panel
Share your feedback directly with our
Fabric product group and researchers.
Influence our SQL roadmap and ensure
it meets your real-life needs
https://aka.ms/JoinFabricUserPanel
https://aka.ms/JoinSQLUserPanel
How was
the session?
Complete Session Surveys in
for your chance to WIN
PRIZES!