As always in these cases, if the fork/rewrite is not compatible with the ecosystem, people won't migrate and the fork/rewrite fails.

neovim only got away with it because vimscript was so horrible and limited, the ecosystem already had an immense amount of churn and vim development was so centered around its developer. It was basically the perfect storm for a successful fork.

But imagine what it would take for the emacs community to switch from GNU emacs. You'd have to rewrite your config, learn new workflows when your old ones were working perfectly fine, learn the new technical concepts, find replacements for all your packages that don't even exist yet until a significant number of devs have switched over. Therefore the only forks that will ever succeed are the ones that start with the promise of full backwards-compatibility from the outset. And even then its more likely that the work of the fork gets incorporated back into GNU emacs at some point, as history shows.