Quants are often STEM PhDs, actually. There is a very famous Pearl student who is a quant now (Thomas Verma). Thomas is famous enough to have a constraint named after him.

It is true that what is considered worthwhile academic work is somehow socially constructed in the end, even in STEM. But in STEM there is a rigorous footing for these things that helps a lot with not running off to lala land (e.g. the process by which these things are socially constructed does not result in nonsense or arbitrary things being rewarded just because credentialed people did them). If a quant outsider constructs a very influential model, I could see that ending up in a Nobel, especially if it goes through a conventional publication process. I think though quants are generally kept very busy with non-academic things. You need space and time to do good work, and people outside academia or places like Google labs just don't have either.