What would be the point? Kernels aren't swappable, and the value of an OS is in the non-kernel parts - drivers and software.
So a new kernel, even if it is enterprise-grade and production ready, won't run all the things that Linux does. Only for very niche cases, it makes sense (bare-metal hypervisors, or example).
> Projects in C/C++ seem to be making much faster progress eg. SerenityOS than the Rust community. What is holding Rust back in this area?
It's slower to write in, maybe? Unlike traditional projects where the language doesn't matter as much as the libraries available, with OS projects having a huge library doesn't help much.
For example, it might be just as fast to write a data processor in Rust as in C++, because the majority of the code is in already written libraries - you just glue together input libraries, database libraries, etc.