Author Topic: Secretive Giant TSMC’s $100 Billion Plan To Fix The Chip Shortage  (Read 597 times)

0 Members and 1 Guest are viewing this topic.

Offline Trader

  • Frequent Contributor
  • **
  • Posts: 319
  • Country: us
 
The following users thanked this post: mawyatt

Online Psi

  • Super Contributor
  • ***
  • Posts: 8398
  • Country: nz
Re: Secretive Giant TSMC’s $100 Billion Plan To Fix The Chip Shortage
« Reply #1 on: October 20, 2021, 01:41:37 am »
So they are investing 1/2 an Elon Musk over 3 years to expand their factory.
That's no small amount of money.
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 23545
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Secretive Giant TSMC’s $100 Billion Plan To Fix The Chip Shortage
« Reply #2 on: October 20, 2021, 03:46:32 am »
Nice, strategic moves like this make perfect sense.

mawyatt, any comments with knowing you spent years in this industry ?
Come visit us on stand 1002 at EMEX: https://www.emex.co.nz/visitor-information/about-emex/
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6229
  • Country: nl
Re: Secretive Giant TSMC’s $100 Billion Plan To Fix The Chip Shortage
« Reply #3 on: October 20, 2021, 06:38:47 am »
Nothing secretive here. All fab owners are scaling up, not only for the high end chips but also the jelly bean parts.
Demand is high.

I know here many people know the following but non tech people have less knowledge with the logistics of chipmanufacturing. If a wafer has for instance 60 layers this means each wafer has to be cleaned, doped, photolayered, exposed, etched, for 60 times before it is finished. Then tested, cut and packaged.
Nowadays state of the art nand chips have over 128 layers, do the math.

So these productruns take a long time from start to finish and therefore you run it for a very large quantity of the same chip. Now if the shortage is a large amount of different chips you feel the pain. Everyone is waiting in line till their chip is made.
Only the extreme high end chips need the smallest new (<10nm euv) litho machines and only for the bottom few layers. Above laying layers the older duv litho machines are used since they are cheaper in $/hr but those same machines are needed for the jellybean parts, so yes you have an overlap.

It is not a cookie factory where you can bake ten thousand cookies on monday, then switch the plates and bake ten thousand different cookies the next day.
Then a new fab takes three years to build and costs in excess of $18.000.000.000 using more power than a large town.
« Last Edit: October 20, 2021, 06:42:08 am by Kjelt »
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 1653
  • Country: us
Re: Secretive Giant TSMC’s $100 Billion Plan To Fix The Chip Shortage
« Reply #4 on: October 20, 2021, 04:58:40 pm »
Nice, strategic moves like this make perfect sense.

mawyatt, any comments with knowing you spent years in this industry ?

This makes good sense for TSMC and the fact they are building a fab in Phoenix, Az in US is also good. TSMC got to where they are today, not by luck, or following short term thinking (Wall St mentality), but by taking the long term path and heavily investing in it. They attracted the right folks that can keep them going, I know because we lost a really good engineer/scientist to them ~15 years ago. Did you note the credentials of the TSMC founder? Harvard, MIT and Stanford graduate, and his technical vision shows!!

We became impressed with them some time ago when they created a program that provided free or almost free silicon to universities, allowing students and professors to actually fabricate custom chips themselves. Tried to do this with IBM back in~2000, but the university management balked and dropped the ball because they wanted someone to cover the university cost for setting up the program. Short term mentality again, it's everywhere and a killer of innovation and true progress, by only focusing on the near term bottom line  |O

TSMC and Samsung have really taken over the semiconductor reins the past decade, and I'm not sure intel will be able to catch up, even with the $50B US subsidy, and they are slated to get a large portion of that pie!! Being US based and a founding father of the semiconductor world, I'm hoping they'll succeed in regaining the leadership role they and IBM had in the past, but not "betting" on it ::)

Anyway, exciting times for the semiconductor world and hat's off to those that chose to take a long-term investment in such :-+

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: tautech, 2N3055, sandalcandal, duckduck


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf