When choosing the dimensions for an MDC (Multidimensional Clustering) table, poor space utilization can occur due to low cell density attributable to the low cardinality of the dimensions selected. Low cardinality means that there are fewer unique values compared to high cardinality, leading to many cells being sparsely populated, resulting in significant wasted space. High cardinality dimensions would typically lead to a better distribution of data across cells, thus utilizing space more efficiently.