The landscape of cloud data warehouses is evolving rapidly. New requirements and technological innovations are driving the idea of combining different compute engines (such as Snowflake, DuckDB, and Polars) within a single data warehouse environment. But why would organizations even consider using multiple engines? This talk addresses the motivations—such as optimizing for specific workloads, cost efficiency, flexibility, and performance gains—that are pushing teams to rethink the traditional “one engine fits all” approach. The session begins with a historical perspective: How did we reach a point where multi-engine architectures became attractive? We then take a practical look at the benefits and challenges of leveraging multiple compute engines in cloud DWHs today, and assess whether it is already feasible and meaningful to use them in production scenarios. Finally, the talk offers a forward-looking outlook on how multi-engine cloud DWH architectures might develop in the years to come. The session is aimed at data engineers, architects, and decision-makers who want to understand the current state, motivations, and future potential of cloud data architectures.