For really old software, it tends not to make good use of multiple cores anyway and you can simply emulate just a single core to achieve total store ordering.

Anything modern and popular and you can probably get it recompiled to ARM64

Unfortunately games are the most common demanding multithread applications. Studios throw a binary over the fence and then get dissolved. Seems to be the way the entire industry operates.

Maybe more ISA diversity will incentivize publishers to improve long-term software support but I have little hope.