The data movement problem is the recognition that in modern computing systems, the cost of moving data — between memory and processor, between cache levels, between chips, between nodes — exceeds the cost of computing upon it. This inversion is the central consequence of the memory wall and the driving force behind architectural shifts toward unified memory, near-memory computing, and data-oriented design. The field has no agreed-upon metric for data movement cost, which means most optimization research still optimizes for operations rather than for the distances data must travel.