At this time, no concrete information on the price of Microsoft Fabric has been officially communicated by Microsoft. However, we were able to get an idea of the numbers that should be released shortly and finally be able to answer the question everyone is asking: How much does Microsoft Fabric cost?
First, if you want to know more about Microsoft Fabrics and its functionalities, read this article : Microsoft Fabric, 5 features to discover that will leave you speechless!
Pricing itself follows the basic principles of Azure: we pay for Compute and for Storage.
In Compute terms, this is scaled for all Microsoft Fabric features in a Pay-as-you-go format, which means it will be possible to scale-up and scale-down resources and opt for cost optimization strategies. Concretely, Microsoft Fabric introduces Capacity Units (CU) which represent a pool of computing power needed to execute queries and tasks. These CUs can be purchased according to the same principle as in the previous price model of Microsoft Power BI Premium per Capacity, i.e. according to SKUs predefined by Microsoft varying from F2 to F2048 (compared to EM1 to EM3 and P1 to P5). For performance comparison, an F64 will be the equivalent of a Microsoft Premium per Capacity P1 license (it is also important to note that Power BI Premium P SKUs have access to all Fabric features by default). In terms of Power BI licenses, if you go for a F64 or a higher capacity, read users will not need Power BI Pro license only users who create content. Users will be able to create data warehouses, use notebooks, and manage their capacity without a Power BI Pro license just as it was possible to do with a Power BI Premium per Capacity license. For those who opt for a lower capacity version than F64, one Power BI Pro license per user will be required to consume the content.
Price for 1 CU pay-as-you-go is $0.18/hour (USD) for WEST US region 2. F2 offers 2 CUs, F4 offers 4 CUs, F8 offers 8 CUs, F16. .. and F2048 offers 2048 CUs. So, for a scenario with an F64 SKU that is open for 730 hours (average number of hours per month), will cost $8,409.60/month ($0.18 per hour * 730h * 64 CUs). It is important to note, as mentioned previously, that cost optimization strategies can be put in place to optimize and reduce performance according to busy periods.
In addition to Compute’s costs, there are OneLake storage costs which will be detailed around 0.023$/GB/month (again depending on the WEST US 2 region). To these charges, inter-region data transfer network charges may apply depending on the source/destination of each storage access.
