I wouldn't be surprised if it did take off, classic Wasm semantics were horrible since you needed a lot of language support to even have simple cludges when referring to DOM objects via indices and extra lifeness checking.
WASM-GC will remove a lot of those and make quite a few languages possible as almost first-class DOM manipulating languages (there's still be cludges as the objects are opaque but they'll be far less bad since they can at least avoid external ID mappings and dual-GC systems that'll behave leakily like old IE ref-counts did).
You still need to usually install plenty of moving pieces to produce a wasm file out of the "place language here", write boilerplate initialisation code, debugging is miserable, only for a few folks to avoid writing JavaScript.
There will always be enthusiasts to take the initial steps, the question is if they have the taste to make it a coherent system that isn't horrible to use.
Counted out over N languages, we should see something decent land before long.
WASM-GC will remove a lot of those and make quite a few languages possible as almost first-class DOM manipulating languages (there's still be cludges as the objects are opaque but they'll be far less bad since they can at least avoid external ID mappings and dual-GC systems that'll behave leakily like old IE ref-counts did).