The killer problem with e-ink is the pricing betrays that the yield absolutely tanks as the area increases, which is true of semiconductors and other display technologies, but in the case of e-ink seems to be that much worse.
Part of the success of things like GPUs is their ability to degrade gracefully in the presence of one or two errors which would otherwise render the whole thing unusable.
If this is the case, then it would seem to me that the ability to produce modular e-ink displays, where segments are individual, independent displays, would be an option.
This is most tenable for large-scale displays (e.g., wall displays), and we're already used to segmented "television walls" in which borders between individual displays is strongly evident. The trick would be for such compound devices to both match one another's display characteristics (brightness / shading / hue), and for borders to be made as undetectable as possible (several possible mechanisms suggest themselves to me as I write this).
Given the sweet spot of 8--10" displays, compound devices made of multiples of these would seem to be at least a conceptual possibility. 16x8 displays would give 32" diagonal measure, 300 dpi, and assuming 50% BOM cost might run less than the 25" Onxy BOOX Mira Pro ($1,750).
LCDs are incredibly simple to manufacture compared to e-ink displays. Also, the best cleanroom technology in the world goes to LCD manufacturing facilities thanks to the incredible economies of scale available.
One dust particle per 12" square loses you 1 of 100 CPUs. It loses you 100% of your 50" TVs.
Also, CPU manufacturing facilities keep the insides of the machines cleaner than the facility, containing the wafers in transit between machines, and LCD manufacturers do not have the same luxury.
That lost CPU would still have potentially been worth several times as much as the 50" TV though. Intel server CPUs especially. Entry level laptop CPUs not so much. ;)
I assure you that losing 1% of the CPUs on a production line is a lot less of a disaster than losing 100% of the TVs. In one case, you have something you can do to improve yield, and in the other case you have a complete waste.
While that kind of makes some sense, it doesn't necessarily mean that LCD clean rooms are therefore the best clean room tech in the world.
I mean, I'm not saying they aren't, I just don't know. And the explanation you're presenting as the likely reason doesn't really bear that out. So far anyway. :)
I am saying it because I know it to be true. Someone later asked why LCD factories would invest in cleaner rooms than CPU fabs, and I explained the logic.
I think you're expecting some sort of first-principles argument as to why that ought to be true (or worse, asking someone to waste a bunch of time digging up a public citation on the details of factory design), which you're not going to get handed to you on a silver platter. To put it bluntly, if you want to think about it from first principles, go for it, it's not that hard. If you want a source for something covered under NDAs, you can go do the work yourself.
you can't just go by retail price though, it becomes a question of whether it's cheaper to implement a better clean room or to just slightly scale up production to account for breakage
Part of the success of things like GPUs is their ability to degrade gracefully in the presence of one or two errors which would otherwise render the whole thing unusable.