Not able to recreate that.
But if the file you add already exist and you are overwriting (removing and adding), then it must rebuild the entire zip.
Also if the zip is missing its content catalogue at the end of the file (depending on what is created it from the beginning). the entire file need to be parsed instead of just jump to end and add it
However MC is not optimized to update zip is fastest way, normally it is not a problem but if you got 200.000+ items in the root of the zip i can cause it to be slow. Zip is handled like a virtual filesytem, you can navigate into the zip as a volume.
Because of this the zip archive must load the entire zip before doing anything to it. and for each file you add it must check so the file do not already exists. and so on.
However there are some optimization added in the coming version. (14.x) It will load a large zip faster.. but it will still be slow.. The way zip is build, having 200k+ items in a zip is not optimal. I think 7Zip handles large number of files better.