Yeah, I'm aware of what the chart shows - reading the source material is better than the chart. Snippet below.
Where I was confused was because we were discussing retention and how you came to the conclusion that it has 1 year of retention when 4DS have specifically stated to the ASX and to shareholders that retention is "seconds to days", followed by you quoting a chart highlighting 1 year, which as the paper points out is discussing energy usage because "retention is energy".
Its just a different benchmark or KPI to use when comparing different memories that aren't all in the same memory class.
But if we want to discuss energy usage, yeah, I agree with you, 4DS has claimed and I have no reason to doubt that 4DS RRAM IP uses less switching energy than DRAM over a year.
(20min delay)