Many organizations have built mature analytics platforms on on-premises SQL Server, traditional data warehouse architectures, and dbt Core. Moving these workloads to Microsoft Fabric and dbt Cloud can look straightforward on paper — until existing assumptions about SQL behavior, development workflows, concurrency, testing and operations meet a very different platform.

In this session, I will share lessons from a real-world migration project in the financial services industry, moving an existing on-premises SQL Server-based data warehouse with dbt Core towards Microsoft Fabric and dbt Cloud. We will look at the architectural decisions involved, the migration path, and what happened when the original migration approach met the realities of the target platform.

Using concrete examples from the project, we will explore SQL compatibility, Warehouse vs. Lakehouse considerations, CI/CD and incremental model testing, development environments, locking and concurrency behavior, reconciliation, performance, governance and operational readiness. We will also look at how platform limitations and unexpected behavior forced us to adapt parts of the architecture and delivery process.

Attendees will leave with a realistic view of what such a migration involves, which assumptions should be validated early, where Fabric and dbt Cloud can add value, and which technical and operational pitfalls to consider before starting their own modernization journey.