Ann: Megabit Success, page-1588

Correct with that article.
That was testing of the cells only.
Now with the the cells packaged complete with selector to Mbit and I suspect that PL3 was packaged as 40nm and failed.
They have now increased the geometry to 60nm and now have a successful Mbit chip. What does that tell you?
You can reduce the cell size on it's own to whatever size you want, but when it has to be packaged into a an array, that is where the issues are and when you are using PCMO you have problems in lowering the geometry as many other companies have found out and gave up developing.
So the question remains that other NVM in many forms are at lower geometries and are in demand. WBT ReRAM as a discrete NVM is now at 22nm and they are working at lower geometry as we speak.
Global Foundries are qualifying the 22nm for mass production.
For a buyer to be interested in 4DS, they want to see that it is able to be scaled as a Mbit chip, not just the cells.
4DS is non-filamentary and if you look at WBT's cell architecture you will see it is an oxygen forming filament...it is not a HARD Filament!