The memory is organized differently, where VGA has to use 4 memory chips and spread a byte of memory among them, while MCGA just keeps it simple and uses one byte of memory per pixel.
Technically in MCGA mode VGA doesnt spread individual pixel bits among banks, pixels are kept "whole" as bytes in banks/planes. 13h is the only mode that doesnt benefit from all of that complicated EGA inherited machinery, yet it still became the most successful and practically synonymous with VGA :)
Btw OG 1987 $595 IBM PERSONAL SYSTEM/2 DISPLAY ADAPTER https://www.minuszerodegrees.net/video/IBM%20Announcement%20... VGA card used eight 64Kx4 DRAMs https://upload.wikimedia.org/wikipedia/commons/3/3f/IBM_VGA_... to build its 4 banks. Ram Address Bus is wired in parallel (can see pin10 A7 common for 2 rows of ram chips, also pin1 OE).