Oh, maybe I used misleading words. It is unexpectedly long for an intro. If you tell people that it will take a while to reach sandbox, you create different expectations. You could also create a third option between skipping and playing intro. A summary in a few pictures. With your style of breaking 4th wall pre-intro, that would be fitting, with that girl laughing at you or you complaining, that your nice intro was being skimmed over instead of being played.
png is a compressed but lossless format. I do not understand why there are png that are not fully compressed. Is this from the olden days of weak hardware? Even the cheapest smartphone has more than enough power to decode those. I loaded one of those and saved it as png with compression 9 and it shrunk by 10-20%. And for the bigger png, I had the tool count colors. While same color areas compress well in png, if it is too many colors, this is bad, because png is lossless and preserves each and every pixel. But your creation tools created those pixels specifically to have a color shading, just as in real life images. So you can use all the tech to keep files small, as is used for normal photos. Your absolute limit would be the point where the ink lines get fuzzy. I only tried with a 1.5MB background. And even in the original, there is dithering if you look closely.
Original size, 1.5MB, jpg75 110kB. While there might be a slight fuzzyness with the black lines, this is also only 1/15th of the size and the original also has this blocky dithering in the green area. Interestingly, when saving this in paint, the png was 52kB and when I saved it again with other tool and compression 9 it only was 35kB. Default png compress is outdated by 10+ years I suspect. They should make default the max. Oh, and that snipped has 5800 colors.
(Of course you need not go 1/15th, but you do have a lot of room for compression)