As far as I know WBT has best in class specs, cost per bit - least exotic materials.. as like no non standard material, best temp performance, with a hard RAD rating.
What actually sucked about this whole story is that Skywater wasn't the fab that was going to get us there. They just don't have the IP blocks - which have to be transferred like reram - to suit the needs of the customers that could benefit from reram. They do have a big defence customer base, but defence production chips do not change if they don't have to. They stay the same for as long as they possibly can.
But what the whole Skywater excercise proved was that WBT reram is manufacturability in a fab at an industrial scale. Showed the world its real. A lomg timeline to do this, but that's what it was. Household names don't manufacture at Skywater - as dar as I know.
We have two fab plays at the moment.
1. GF. The information to date makes it pretty clear that the fabbing at gf was not a partnership but instead WBT got it fabbed there, to show it could be done at 22nm. The chips work. They've demoed them. They should soon be qualified. The tech transfer to GF is complete. Any GF customer can request wbt reram. GF haven't made a big song and dance about it, because they spent a pretty Penn buying/licensing another reram - what was dialog, so they have to at least try and get some return on investment. But from the data that's been shared in presentations this reram is inferior. GF make money either way. I'm not aware of anybody using gf reram, please post if any of you guys know.
2. DB-HiTek. This is probably going to be the low hanging fruit. From memory lots of household names manufacturing in analog, mixed signal, power management, auto motive, at 130nm nodes. Drop in, bang bang, winner winner, chicken dinner. Should get wafers this year. Qualification to 85C before mid 2025. WBT should be holding the aces around this time.
Discrete. I am more than a bit dissapointed that there's been next to no news since 2022 on the progress of this, there was an announcement that eluded the selector could also be used in embedded, but this is actually the big money spinner. No customer needed. Automotive spi flash chips suck (typically nor flash). Ask tesla. You make a 2 gigabit wbt reram discrete spi chip to replace automotive flash, every automotive manufacturer will be at your door. They'll be the detectives, Coby will be Sharon Stone recrossing her legs. They'll be salivating. Why no news? I do recall they had a disclaimer in their quarterlies something to the effect of "we reserve the right not to disclose any information that will place the group at a competitive disadvantage".
Again, I'm holding, mid next year, if not before, it ain't easy getting rich.
gltah
Expand
(20min delay)