First one who gets a Rails-like showcase application. If that doesn't happen, both will get some action, but more on an Erlang/D level than Ruby/Python.
Which is why I don't doubt that there'll be quite a few people using Rust and/or Go. But I'm really not sure whether either of them will move beyond this tier into the PHP/Python/Ruby territory. A framework like Ruby would certanly help, and that can come out of nowhere. But I wonder whether the day of killer-feature frameworks aren't numbered, i.e. whether something like this can happen again.
So I think in the end it comes down to which of the two will do better in the Haskell/Ocaml/Erlang tier. Go has some rabid fanboys and quite a decent designer pedigree, good libraries and a bit of a headstart, whereas Mozilla is ripe with dumped applications and failed projects. So unless they're rewriting major parts of their infrastructure with Rust and actually release stuff, I'd bet on Go.
Mozzila giving out broken products? - well, dont you use firefox?
There actually might be more renewed interest in C++ (in its current C++11 reincarnation) than actually new interest in Go/Rust. (Chances of someone reviving Modula-3 or creating a killer framework for Ada will remain low)
Sure, you don't have the legacy libraries and architectural styles of C++ in Rust/Go, but that's a pretty moot point if we're talking about new projects. Go wild with generics, lambdas, concurrency, type inference etc..