The size of the cache probably isn't a concern: 20GB of hard disk space would be enough for 200,000 items even if each page was even if the individual pages were 100kB each (and they're not; they're more like 70kB once you count the images). How much would a 20GB hard drive cost assuming you can still find the things? The HTML is only 16kB, but can be reduced a fair bit by more considered styling (CSS selectors can be a bit more sophisticated than single classes, you know; and an HTML element can have more than one class; and font tags were deprecated last century - oh, and you'd probably want to add "sans-serif" as a generic backup body font). The CSS file and image files can have client-side cache control headers added so that they aren't repeatedly requested by the same client. But no, size isn't really the concern.
Maintaining the cache so that it doesn't continue to list deleted items, or fail to recognise new items is more tedious. It would be more productive to first investigate and find out just what is being slow, and look at speeding that up.
