In practice, it's turned out to be an "expert's tool" that's more expressive-convenient than C/++ to do all manner of things in the realm of systems programming, from embedded to HPC. And it's got great C/++ interop, so you can continue to leverage those software ecosystems.
In that respect, I'd say the biggest boon, rather than a killer project, would be for a famous programmer or shop to publicly and loudly adopt it as their everyday language, shouting its name for several years and saying things like "Wish we had started using this years ago, would have been so much better, look at all of these great apps and libraries we've been able to create and maintain more easily than if we'd stuck with C/++/Go/Rust/Swift!"
Every year millions of university students design new languages across the globe on their compiler classes.
To productise any of them, there is a whole ecosystem that needs to be added on top, editors, graphical debugging, packages, AI tooling....
Nim had a real chance at gaining a foothold, it just needed a company to back it. I think sadly that ship has sailed by now.