My assessment is that 4DS has a chip mixture right now that, if...

My assessment is that 4DS has a chip mixture right now that, if this was the end of the 18 month collaboration period, they would be happy with. However, they are still 3 months off the end of their agreed 18 month collaboration window with IMEC and honestly, you'd be a mug not to take advantage of the booked fab and lab time that they have remaining (and have paid for).

There is no such thing as too much data (remember, all this IP and knowledge will be invaluable for a potential purchaser) Once 4DS walks out of IMEC, if they wanted to get back in to do more work, they would have to take a number and join the back of the queue behind everyone else. How stupid would these guys look if they gave up 3 months of access to the lab and in 6 months time turned around and said, "oh wait, we need to get back into IMEC for another iteration."

My highlights of the announcement -

Endurance cycling into the 10s of millions -

From a page I found on the web, check the first page

https://www.cypress.com/file/202316/download

If you cycle every 5 minutes with endurance of 100,000 cycles, the product life is 1 year

10 million /100,000 = 100

100 year product life cycling every 5 minutes

20 year product life cycling every minute

5 year product life cycling every 15 seconds.

Compatibility with backend of line temperatures

This means they can thermally bond the chip to the megabit array - great!

I'd like more detail on retention but will be happy with the fact they are improving it every iteration. If the end user is cycling every minute and retention is 2 minutes that isn't an issue. Also the data can be cycled over to another memory block before that time is up and if that has to be every 2 minutes that's a 10 year product life. Remember the key thing is, while the data is locked into the memory chip, it can be powered off and that saves a huge amount of energy, even if you have to move it around every 2 minutes to beat the retention time.

I'd like information on yield, but that's something you can probably only know once you've started fabricating enough wafers to produce a statistically significant result.

If people are worried about the lack of reaction, I wouldn't be - stocks go up by the stairs, but down by the elevator. If we got 10% increase every day from now, after 10 days of trading we would be up 150%. Last year, once the new financial year rolled over, this stock started seeing good gains.

As to the open ended nature of the next iteration I'm sure we'll get another announcement letting us know the next fab phase has started and an indicative completion date.