The day the control was in the hands of the experts from Wadiya it was over.
Their process advances got aladeen-ed.
I think it is the opposite. Management was to US-centric. When the market for GlobalFoundries was outside of the US.
1. They should have synced nodes internally and not synced nodes with competitors.
2. Fab7 should have never been allocated to mainstream only. Instead, should have been leading edge.
3. They shouldn't have split the fabs to different generations. Especially, with the 2011 5-year estimation that they could have increased leading edge revenue by 5x.
The only fab sync
Fab 1 & Fab 8 for 32SOI/28BLK.
It should have been
Fab 1, Fab 7, Fab 8 for 32nm SOI/BLK, 28nm BLK.
I don't think GlobalFoundries will ever be successful in leading edge if it is only one fab doing leading edge.
Currently, no one can get 22FDX from a United States source, nor can anyone get 22FDX from an Asian source.
I just don't see how anyone could see this. Then, be like well Singapore shouldn't be doing 22FDX or Malta shouldn't be doing 22FDX. The Chengdu option was an okay consideration, but it fails when Singapore is considered a cheaper/lower risk option.
Missouri GlobalWafers SOI to Malta = 22FDX US-source feasibility increased
Pasir Ris SOITEC to Singapore = 22FDX Asia-source feasibility increased
This also spreads to RFSOI/Photonics-on-SOI/etc.
I see whatever customer is on GlobalFoundries 14LPP/12LP/12LP+ ditching for Intel's IFS/IDM 2.0.
"IBM and Intel's plans to collaborate on critical, leading-edge research to advance logic and packaging technology development is an important milestone. By bringing together two of the best semiconductor research organizations, we will greatly accelerate innovations in the semiconductor industry and push the frontiers of technology for the benefit of clients and the world."
Friendship ended with GlobalFoundries. Now Intel is my best friend. (US supply of leading-edge semi for IBM)
~~~~
The general plan for GlobalFoundries is to split manufacturing capacity across the three fabs.
14LPP @ Fab8 => ~66,000 wpm
22FDX @ Fab1 => ~75,-90,000 wpm
--
If region usage is equal. 1/3 @ Americas, 1/3 @ EMEAs, 1/3 @ Asias
Fab 1/Fab 7/Fab 8 split means:
14LPP @ Fab1/Fab7/Fab8 = ~22,000 wpm per fab complex.
22FDX @ Fab1/Fab7/Fab8 = ~25,-30,000 wpm per fab complex
--
Fab 1 ~110K wpm, Fab 7 ~120K wpm, Fab 8 ~90k wpm
14LPP can be translated to 12FDX => 3x~22k wpm
22FDX => 3x~27,5k wpm per fab complex
28/32 => 3x~15k
40/45 => 3x~15k => Malta's prior reported capacity reached
55/65 => 3x~15k => ~94.5kwpm
80/90 => 3x~15k => ~109.5kwpm (Dresden's reported capacity reached)
110/130 => 3x~10k => 119.5kwpm. (Singapore's reported capacity reached)
--
Fab 1 can get one more module.
Fab 8 can get two more modules.
Fab 7 has rumored plans to build two modules/one complex.
So, even if the demand isn't the above per region they could fit more or less capacity for high demand or low demand nodes. Or, if they are crazy they can subsidize "Optimized layout migration" so 22nm FDSOI gets 28/32+40/45 designs and 45nm FDSOI gets 55/65+80/90+110/130 designs. Which is basically an advanced retapeout option to get a prior node chip to a FDSOI node chip. Which minimize the mature logic nodes they need to support.