Ann: 4DS Technical Update, page-404

For what it is worth, this is what I am guessing occurred. A certain Dr that is now not employed screwed up the etch. It was probably something painfully simple and when they first looked at it they thought it was an easy fix. Then they looked at it again and realized they were going to have to redo the etches, which immediately imposed a significant delay and took them beyond the current IMEC agreement - hence the need to renegotiate. Simultaneously they had these great numbers from the fully crystaline PCMO and when WD looked at it they indicated they wanted that improved performance. So then the other Dr's thought, if we are going to have to redo the etches to fix the stuff up from the other Dr, we may as well go to the fully crystaline PCMO, but there is a bunch of stuff we are going to have to tweak to get this right and it is going to have to go through a whole new simulator road trip which will take a few weeks, and then more weeks to make the etches. Oh crap. Meanwhile the conditions under which they have to test the non-platform lots don't allow them to properly limit current fluctuations and they need the Imec controller to know whether the endurance reduction is a feature of the test environment or a consequence of the fully crystaline PCMO - which is why they didn't go that way in the first place.

Then God knows who wrote the release, but I am guessing it was one of the engineers - because that is how it read.

The comment about the "no established SCM market" is simply engineering speak rather than marketer speak for the fact that they don't know what the relative price to characteristics are because there isn't yet a wide range of solutions for storage class memory. Hence in each lot they have been tweaking qa bunch of stuff to get trade-offs between retention, endurance, speed and power consumption and cost and they don't yet know what the sweet spot, and they won't know until they start building it and do market surveys - which isn't 4DS's problem because they are not selling completed devices. 4DS is going to sell the tech to someone else (like WD) who will make a range of SCM devices for the customers and they are the ones who will want to be able to adjust parameters to produce devices of differing performance and differing price breaks and targeted controllers to optimise the behaviour under the conditions faced. For example a blindingly fast drive might have lower endurance, but be packed with twice the memory to allow cells to be masked out invisibly as they fail - all handled by the controller, while another slower device might be much cheaper because it doesn't need the cell masking and redundancy. Both devices have a market but the differ widely in cost. None of this is known until you are actually designing the end consumer device... Which is not what 4DS is doing. 4DS is providing a tech than can be set up and configured and made in multiple ways to get a variety of performance characteristics so the manufacturer can make all those decisions.

I think, it was a completely inept market statement prepared in shock, panic and finally relief when they had made all the hard decisions and got everyone to agree.