Forum Discussion

peternguyen2's avatar
peternguyen2
Frequent Visitor
2 months ago
Solved

One dataset exceeds 1GB

Hi,   All of our company's Power BI users are on a Power BI Pro license. Recently, we have encountered a dataset that exceeds 1 GB in size, which surpasses the storage limit included with the Pro l...
Need Help
  • Parchitect's avatar
    Parchitect
    2 months ago

     

    If only one semantic model exceeds 1 GB, the most cost-effective first option is usually Power BI Premium Per User (PPU), not a Fabric/Premium capacity.

     

    With Pro, the semantic model limit is 1 GB. PPU supports larger semantic models, up to 100 GB, and also gives you Premium features such as XMLA endpoint, deployment pipelines, 48 refreshes per day, etc.
     
    However, there is one important licensing point:
    If you put the report/model in a PPU workspace, then users who view that content also need PPU. Pro users alone cannot consume PPU workspace content.

     

    So the decision is usually:
    1. First try to optimize the model
       - Remove unused columns
       - Reduce high-cardinality columns
       - Disable Auto date/time
       - Use aggregations
       - Use incremental refresh
       - Review model size with VertiPaq Analyzer / DAX Studio

     

    2. If only a small group needs this large model
       → Use PPU for those users.

     

    3. If many users need to consume it
       → Compare PPU for all consumers vs Fabric/Premium capacity.

     

    4. If you need free-user consumption at scale 
       → Fabric F64 or higher / Premium capacity is usually the capacity route.

     

    I would also add one practical warning:

    The 1 GB number is model size in the Service, not necessarily the PBIX file size. Always check the actual semantic model size and optimize before changing licensing.
     

    🔍Parchitect
    Solutions Architect · Microsoft Fabric Specialist

    💡Helpful? Kudos are appreciated.
    ✔️Solved? Mark as Solution so others can find it faster.

Parchitect's avatar
Parchitect
Icon for Solution Sage rankSolution Sage
2 months ago

 

If only one semantic model exceeds 1 GB, the most cost-effective first option is usually Power BI Premium Per User (PPU), not a Fabric/Premium capacity.

 

With Pro, the semantic model limit is 1 GB. PPU supports larger semantic models, up to 100 GB, and also gives you Premium features such as XMLA endpoint, deployment pipelines, 48 refreshes per day, etc.
 
However, there is one important licensing point:
If you put the report/model in a PPU workspace, then users who view that content also need PPU. Pro users alone cannot consume PPU workspace content.

 

So the decision is usually:
1. First try to optimize the model
   - Remove unused columns
   - Reduce high-cardinality columns
   - Disable Auto date/time
   - Use aggregations
   - Use incremental refresh
   - Review model size with VertiPaq Analyzer / DAX Studio

 

2. If only a small group needs this large model
   → Use PPU for those users.

 

3. If many users need to consume it
   → Compare PPU for all consumers vs Fabric/Premium capacity.

 

4. If you need free-user consumption at scale 
   → Fabric F64 or higher / Premium capacity is usually the capacity route.

 

I would also add one practical warning:

The 1 GB number is model size in the Service, not necessarily the PBIX file size. Always check the actual semantic model size and optimize before changing licensing.
 

🔍Parchitect
Solutions Architect · Microsoft Fabric Specialist

💡Helpful? Kudos are appreciated.
✔️Solved? Mark as Solution so others can find it faster.