An old friend of mine did this work for NetBSD as his master's thesis at MIT. Rebuilding an emacs would, at a very rough estimate, be no more work, likely less. He had the benefit of being aware of other attempts to add SMP support to other unixen, though. 

One major problem with building a truly multi-core-optimized Emacs is that so much of the core concepts revolve around shared state, especially buffers. I *think* there's a workable design similar to a modern web browser with moderate-to-high sandboxing and worker threads, but someone would have to care enough about each of Emacs, modern browser implementation (including javascript and sandboxing), and language implementation to puzzle out the details. I don't know of such a person (but it would be a bit surprising if I did!), and I don't have the same sort of contact with grad students anymore. My hope is that there's still a window for such a person, and they haven't all been swallowed up by VSCode and co-pilot.