What's new

TSMC set to build advanced 5nm chip factory in US

By 2023, Intel may still be stuck in 14nm + 10nm.lite.

Then USA will be 2 generation behind.

Intel can no longer innovate.
 
.
Wall Streets are major shareholder of TSMC-- not Chinese.

Who is hostage to who?

If you look into it, the major shareholders of TSMC are Arabs, and a lot of sovereign/pension funds from around the globe.

But do you think this will be the only one TSMC 5nm fab?
So I think you think too much.

TSMC has a nearly finished 5nm plant in Taiwan. By 2024, 5nm will be 2 nodes behind the bleeding edge, and nobody to pay to keep it running.
 
.
.
China are not that far behind. Not to worry.

https://wccftech.com/chinese-semiconductor-manufacturer-smic-to-introduce-7nm-node/

The potential of 7nm havent fully utilized given its so new.. As many say 7nm processor can still produced miracle.

SMIC - Chinese 7nm to Begin Production in Q4 2020
SMIC's N+1 7nm node is set to bring a variety of improvements over its current 14nm production node such as a 20% increase in performance, an improvement in logic density of 63%, and a total die area reduction of 55% according to Dr. Liang Mengsong, co-CEO of SMIC.
 
.
.
TSMC has a nearly finished 5nm plant in Taiwan. By 2024, 5nm will be 2 nodes behind the bleeding edge, and nobody to pay to keep it running.

only if you assume Moore's law will continue and that costs don't continue to escalate for leading edge nodes. I think there won't be too many nodes beyond 5 nm. Right now even 14 nm (now a trailing edge tech) fabrication isn't widespread, much less 7 nm.

Although, I have to say, currently the nominal "node" doesn't really mean much. I think gate width stopped being the definition of a process node already.

My prediction: anything after 5 nm will be hugely delayed or cancelled, just like how the 300 mm to 450 mm wafer transition was delayed, due to fast rising costs. 450 mm wafers were put on hold indefinitely because the size required for process chambers and handling was just too high, even with improved productivity. sub-5 nm could be the same way.

Glofo and Intel doesn't control the most advanced process tech, only Samsung n TSMC. That's why Trump wants TSMC. But I think people are waiting to lose the elections.

Samsung doesn't have good yield on their process. Right now I believe the majority of 7 nm revenue goes to TSMC. GloFo had 7 nm R&D but dropped it because it is too risky.
 
.
only if you assume Moore's law will continue and that costs don't continue to escalate for leading edge nodes. I think there won't be too many nodes beyond 5 nm. Right now even 14 nm (now a trailing edge tech) fabrication isn't widespread, much less 7 nm.

Although, I have to say, currently the nominal "node" doesn't really mean much. I think gate width stopped being the definition of a process node already.

My prediction: anything after 5 nm will be hugely delayed or cancelled, just like how the 300 mm to 450 mm wafer transition was delayed, due to fast rising costs. 450 mm wafers were put on hold indefinitely because the size required for process chambers and handling was just too high, even with improved productivity. sub-5 nm could be the same way.
The question is: Do mobile phones need <7mn chips at all? Do the APPs running on phones need faster and smaller chips at all?
 
. .

Pakistan Defence Latest Posts

Pakistan Affairs Latest Posts

Back
Top Bottom