Required! Man, this sounds like the bank telling me to pay the 10$ card maintenance or fuck off!
That is the law of demand and supply of block capacity directly speaking.
Still haven't told me the major drawback of rising the block size to 4MB or 8 MB.
I already have. It's the same drawback as to why enhancing the privacy on the protocol might not be worth it: maintaining conservative approach. Mess around with backwards compatibility, do reckless experiments for the sake of progression, and you just have a new toy in the grand computer software collection that ultimately failed.
Just to clarify, so you don't ask me again: The extra storage that the node operators will need to possess isn't the issue per se. The issue lies in the notion of breaking backwards compatibility to enjoy some short-term relaxation, despite the immeasurable mountain of evidence that suggests otherwise in a provably successful conservative approach.
You're talking about old nodes rejecting new blocks!
Backwards compatible software means to run two instances of it, and ensure that both can seamlessly communicate and interact with each other. When prior-version and post-version nodes have a mismatch in their chain, it indicates that something has broken along the way.
By the way, I see you keep mentioning global adoption. You do know it can't work unless the chain weights a VISA data center, right?