It’s much smaller than the original table. If you compute lots of these, then voila, you have an OLAP cube.
I wouldn’t call this an “OLAP Cube”. It’s just an aggregated fact table. A collection of those with their corresponding dimensions is a “data mart”.