200mm was supposed to fade away as 300mm came online, and that worked from 2007 – 2014, but the trend reversed thereafter. Customers like building on 200mm fab lines because the manufacturing technologies are extremely mature and the costs are low. Many customers don’t get much benefit from moving to lower geometries, and they want to stick with the designs they’ve already paid
The highlighted is problematic for the transition to 300 mm and it is the same for a supposedly move from 300 to 450 mm, which have been in limbo for more than a decade.
300mm economic advantage becomes insurmountable at high volumes, but
the automotive is not near that point. Its a market of small lots, and really assorted IC.
200mm has the advantage of super-cheap masks, and small lots. This matters a lot in a
quite small automotive IC market. They would've stuck to 200mm, even without the certification being a showstopper for automotive ICs.
By the way G450C was Cuomo's pet project through SUNY which hosted the program, and went bust after NYC government dropped the ball on it after Cuomo bribery scandal.
Precisely because the automotive industry have different standards for autonomous applications, developers would rather stick with what they know to be proven. For the foreseeable future, as the EV matures in designs, engineering, and mass manufacturing, the EV will have more internal volume so the need for miniaturization, as expected with 300 mm geometries, will not be there, keeping the 200 mm wafers around for as long as the fabs can make money from them.
I see it going very much like you say, but here are some counter arguments.
Some companies may move just because of 300mm capacity guarantees being more solid than 200mm, but I see no industry-wide need to move. TI, I know, does a majority of its mixed signal on 300mm for every project they can't make on their own fabs exactly because of that.
Further increase in 200mm demand will lessen its economic appeal for sure. 200mm fabs can't keep increasing lead times to infinity.
And... even the certification is not a showstopper. You've better have seen horror story like run-cards on stuff sold to automotive. Having seen some myself, I can only wonder what's the point of all that certification now if buyers don't really care.
My clients will be satisfied that I am not losing any certification and time because these are established technologies. If I build/buy a 300 mm fab, it would be monumentally stupid, not to say losing money, to build common products with 200 mm geometries on a 300 mm wafer. So why should I risk at least 2 yrs of making no money by moving to 300 mm ?
Money wise, 200mm is surely a gold mine now.
SMIC's idea to use 200mm in a first tier fab in 2018 was laughed out, but now they are shovelling money.
And yes, and idea of using new 300mm fabs to produce at 200mm geometries does not make sense, even with the current demand crunch. Prices on 180nm-90nm need to go up like twice for that to happen, and if they will, then people will switch to lower geometries at least for logic.
This make 200 mm wafer technologies and associated equipment even more valuable, which means more money the foundries can charge.
Money, and time will tell.