I had the same idea but I don't think it will work very well just by using Optane as memory. While Optane is much faster than NAND flash it still a lot slower than main system ram nevermind vram and I suspect to run big AI models well you are going to need all those megatransfers.Im gonna bet Huawei, Microsoft, AWS, Apple are all allocating their semi design houses to fix that problem.
I think the big problem right now is memory size on devices for the bigger and more useful models. I think the tech that intel and micron made "3d xpoint" non volatile memory might be a nice solution. 3d xpoint can act like a storage but with memory like bandwidth. For normal programs there was never really a nice usecase outside of really high performance usages.
What may be possible is to have a two tier system. I have a cache system setup using Optane to accelerate my OS drive (m.2 NVME) and data volume (software raided SATA SSD) and it works well, makes the PC feel much more snappy:

(Romex Software is a Chinese company based out of Shanghai BTW, they ended up making a better cache software than both Intel and AMD)
A similar 2 tier system might be useful to run on consumer software for memory where the main memory is the faster cache level with a larger and slower Optane level underneath.
But the problem is Intel and Micron have stopped producing Optane because they couldn't do it profitably. I wish they would sell that technology to YMTC, maybe they could do it profitably.
Last edited: