nitty gritty
HBF / HBM (Sandisk and others)
is incurring substantial cost add ins that wbt /rram will not encounter
HBF and HBM
by expanding data channels in NAND - vertical stacking +tsv + optimised memory controller + logic based die + 2.5D implementation to integrate directly with a GPU is costly
WBT rram undercuts this equation by multiples then add in the volume cost advantage then add in scale iterations and ultra low power options and sram buffering which is in the pipe through the array optimisations Ilan Sever and team are developing + customers will never leave
cheaper than flash? oh yeah by orders of magnitude with more adoptin applications
the strategic pathway here is solid -again why did dbh and onsemi sign up?
the reasons are above
no brainer if you analyse the reason it will win against the competition imo
(20min delay)