@mjshack
yup, I'm glad they've finally released specifics about retention, I can finally put away the stump that used to be an axe.
I do have to disagree with you on your interpretation that it could be universal memory - the retention range explicitly excludes that, as you sacrifice endurance for retention, the longer the retention, the less writes you get, and even then the outside range retention is orders of magnitude less than other non-volatile technologies, as i said, its still a very useful, potentially very valuable technology, but its most certainly NOT universal memory.
I see applications as
1. Datacenter replacement for ram - thousands of times less power used for memory. potentially immune from "bit flipping" meaning ECC is dead too.
2. nvram "cache" applications where high availablilty systems could have "fast boot / last n seconds ago" restarts in the event of a reboot
3. iot devices, portable devices, significant power savings and potential for variable clock device applications, natural synergy with e-core type applications.
4. analogue application is clearly speculative, but confirms our own extrapolations that its something the company could be considering.
I agree that this signifies a pivot away from the "race to sell our patents" goal, and could well mean a much longer term end goal.
as to if thats a positive or negative, i'm on the fence, I'm invested cos i want money, so for me personally the longer term the goal the less interesting the stock is (at least until i get a house lol)
I'm on the whole pleased by this announcement, as the product now has enough performance specs to genuinely consider real world valuation.
Expand
(20min delay)