Running discussion on SP, page-7054

Yep, there is no argument that ReRam has great potential, and will probably a main memory type in the next 10 years.

The risk I see is competition, both legacy & emerging technology .............& timelines..

Take the Tesla memory problem for example, the part number in question is H26M42001FMR (or similar), retails about $10, less for wholesale, capacity 8GB.. The replacement part number is MTFC64GJVDN-4M (according to some) 64GB @ ~$45, probably going to be under $10 by the time Weebit is ready. This is the discrete market, so flash limitation of ~28nm doesnt really come into play, what Weebit needs to do is scale to 8GB and be well under $10 retail.

The problem is Weebit's first embedded chip was 32MB, at 130nm, now when they make the same product at 22nm I assume the capacity would be about 180MB. Of course embedded doesnt use the whole chip, maybe 20%. So if ReRam took up the whole chip, you're looking at around 1GB.

So now we have the problem of scaling, as 1GB wont cut it for many application, Weebit will need to use multi level or multi layer in order to meet customer requirements, so thats probably a few generations off, maybe 5 years?????

Does anyone know the capacity of the 28nm demo chip? could prove if I'm completely wrong

What I'm waiting to see is customer signings on the embedded space!