It make sense to move T-SQL to an engine supporting it such as Fabric Warehouse, but I think this kind of migration requires dipper evaluation in the context of future costs and possibilities. MS Fabric is still new. Databricks is a more mature solution, but in your case, you would need to translate T-SQL to Spark/SparkSQL (it's possible to simplify it). Databricks has integration with Power BI and it works well with SQL Warehouse.

I think it's not just mater of moving, but analyzing it in wider perspective: cost, features, team capabilities, etc.