Few people have a problem that Optane directly solves. Its utility is narrow in the best case and you would need to specifically engineer a software stack to take advantage of it for some small part of the memory address space.

In many cases it is desirable for memory to be volatile. If high-throughput paging mechanics are required, like databases, this technology doesn't serve much of a purpose. Storage densities per server have only been increasing. It has some obvious utility for transaction logs and such but falls into a notorious dead spot in the market -- people with low volumes don't need the performance and people with high volumes can achieve similar results by other means for a price that isn't that different from the Optane premium.

I've tried to find a use for it, and I haven't been able to. That doesn't mean a use case doesn't exist, but if it was "the biggest new idea" you'd think the use case would be more obvious for more applications.