(If the question is why no generic fold over any monad exists in Rust, the short answer is that Rust isn't Haskell. Which is great, it means Rust can specialize in things Haskell can't and vice versa. Personally my favorite example is that Haskell avoids shared mutable state by avoiding mutable state, and Rust avoids shared mutable state by avoiding shared mutability. Both have their downsides but each works better for different applications.)
Having the same function exist with the same name helps communicate with other programmers, even if you can't express the fact that it's actually the same function at the language level. Just like in a language without generics you might still have add(int, int), add(decimal, decimal), add(string, string) and add<T>(list<T>, list<T>) functions. Rust already offers a function called "and_then" on most of its monadic types, and a programmer can see and understand that these are all in some sense "the same function" even though there's no way to abstract over that.
edit: Ah yes, I see that Scala chooses to use the name 'fold' for what is essentially the 'maybe' function in Haskell.
I don't think there is any very useful generic notion of 'fold' that corresponds to this usage. (I see how the operation is a kind of fold, but the type signature of the function isn't flexible enough to be used for any foldable type.) So that is why Rust and Haskell don't name this operation 'fold'.
maybe' def f value = foldr (const . f) def value
Although in pratcice almost noone uses Maybe's foldable instance (and it also doesn't scale to other datatypes, like Either, nearly as well, because of the multiple type parameters).
https://doc.rust-lang.org/std/option/enum.Option.html#method...