Ann: Weebit Nano packages its first memory chips, page-69

I think things seem to be going very well so far. Its early though, with the pre-work around ability to shrink being very important in ensuring the first runs will more than likely work than not.

We dont want to move too fast though. Whilst there is optimisation going on at 40nm if they move too fast to 28nm then changes that increase yield and speed and efficiency at 40nm and are usable at 28nm may not get into the first runs at 28nm. Quick to first silicon wafer is important but quick to best marketable silicon wafer will save time in the longer run through not needing multiple runs to get to viability and also make the product far more marketable.

The key point with WBT is they have worked on industry expected timeframes, calling out a 12 month optimisation phase on 40mn. Now, we dont know a few things. What yield was like off first wafer and what the chip performance specs are like off same. Given they were not afraid to send off the chips to others and went to the effort of full commercial pinout packaging. That probably wasnt too bad. Now if the chip performs to all those parameters off first wafer, then that optimisation phase coould conceivably finish sooner. It would also portend well for 28nm.

I think there are a couple of things we can be quietly assured on here.
1. is the confidence in the partner CEA-Leti. These guys are leading edge and so far havent missed a beat.
2. The incorporation of WBT reram specifications into the Silvaco tCAD tools can only assist in speeding up the move to 28nm by making changes, simulation and testing much easier. If you look at both Silvaco and Leti websites you'll realise just how close the relationship is. This is a very powerful advantage and will also hasten takeup later on.
3. they know what they have to do at 28nm and 300mm wafers are not a problem. Leti have a 300mm fab line in Grenoble and they have partner relationships with STM and Intel who also have 300mm fab lines. I dont think either Leti or STM are 28nm fab lines yet though, but fatman32 pointed out STM have committed to a 300mm line sometime this year (in Italy I think?) and that would have to be below 28nm.

I guess the key point here is know what is realistic and WBT's progress to date and likelihood of success comes into some perspective.