Rewriting all the C code of Emacs in some other language would be a
tremendous job.
But one that can be "embarassingly parallel" and that doesn't require
necessarily much design work, so it's not terribly hard, many people are
able to participate.
It might not make Emacs better, so it might be harder to motivate people,
but the REmacs experience seems to suggest that it could work.
If execution of the other language is slower, the slowdown could make
Emacs unusable on old machines. Old machines such as the T200 and
T400 are very important for our freedom since we can run them with
Libreboot. We must not downgrade support for them.
My main machines run on Core 2 Duos, so I definitely care about those
"old" machines,