When they first started ramping up ads and demonetising more videos for being insufficiently advertiser-friendly, they probably still had enough goodwill from users that if they’d immediately launched YouTube Premium and presented it as a way to both remove ads, and support video creators that couldn’t rely on ad revenue, it would have been decently successful. A good number of YouTubers who had to switch to sponsorships and Patreon could have been pushing for people to subscribe to Premium instead of play Raid: Shadow Legends, which presumably would have boosted subscriber counts, and might have been enough to make YouTube profitable and much more pleasant for both free and premium users than it is today. Instead, they burned through a large amount of goodwill before implementing Premium, so people were already more reluctant, and for a long while it only shared revenue with a select few channels who were already raking in ad money, and was unaffected by view counts, so early Premium subscribers were paying Logan Paul even if they never watched that kind of video, but weren’t paying the channels they actually watched.
It’s not Kessler Syndrome until it’s so bad that we can’t feasibly launch anything new. A single cascading collision chain might calm down again after breaking a lot of stuff without any catastrophic long term impact if all the debris ends up either in a stable orbit that can be predicted and avoided by other objects, or unstable orbits that decay until the debris falls out of orbit.