Array ( [content] => [params] => Array ( [0] => /forum/threads/can-semiconductor-manufacturing-return-to-the-us-interview-with-dr-morris-chang-by-brookings-institution.15907/ ) [addOns] => Array ( [DL6/MLTP] => 13 [Hampel/TimeZoneDebug] => 1000070 [SV/ChangePostDate] => 2010200 [SemiWiki/Newsletter] => 1000010 [SemiWiki/WPMenu] => 1000010 [SemiWiki/XPressExtend] => 1000010 [ThemeHouse/XLink] => 1000970 [ThemeHouse/XPress] => 1010570 [XF] => 2021770 [XFI] => 1050270 ) [wordpress] => /var/www/html )
Morris explained that majority of top talents in US go to MBA, management consulting, finance, Wall Street, IT, computer programming, and Internet/startup companies.Did he say Morris Chung? Is that the proper pronunciation of Chang? Hard to believe he is 90 years old!
US Semiconductor Issues:
Lack of manufacturing talent
Cost difference between Taiwan and OR is 50% more
Arizona fabs were at the urging of the US Government
US semiconductor onshoring is a wasteful exercise in futility
US best design capabilities in the world
Taiwan is not nearly as advanced as US companies in design
Morris assumes there will not be a war between China and Taiwan
Using Oregon TSMC fab as an example is old news?
Morris explained that majority of top talents in US go to MBA, management consulting, finance, Wall Street, IT, computer programming, and Internet/startup companies.
I believe to work in a cold, hot, noisy, or restricted manufacturing environment is not an attractive option for many new college graduates.
In the past several years I happened to know several college graduates with EE degrees and just started their careers. All of them graduated from those "top" universities that Intel and TSMC always arranged to have onsite recruiting events in the past.
Among them only one went to work for semiconductor industry in the design field. Everyone else got a job at various high profile companies with starting annual salary of $160K or more. Some of them even got a guaranteed remote working option if they choose to do so.
The gap between software engineer and hardware engineer is getting larger. Some software engineers work from home or work in tropical islands.
I mentally tick the boxes the Morris ticks, all the negatives, yet the US continues to be the best place for semiconductor manufacturing reshoring. It’s not as good as it was in 2008, when all the 200mm fabs closed, and many people left the industry, but it’s a peer of Korea and Taiwan and will within 5-10 years be what it once was.
The pay gap and cost gap is sort of destroying Taiwan and Korea with a low birth rate. Taiwan and Korea engineers should earn like US, except for collusion, concentration and monopolies. If they don’t solve the problem (and it may be too late), Korea and Taiwan will age and disappear.
Part of the toxic employment mix in Asia is how limited the number of credentials are, how early in life they occur, and how limiting that is. Its just an overwhelming advantage that the US seeks both academic and real world credentials.
I think there is a simple solution to the US semiconductor problem.
Namely, move CS departments back into physics the way they
were in 1960s. Some examples: Stanford, UCB and MIT EECS
departments will not let scientist types have academic jobs. Best
example is Spice creator Larry Nagel not being allowing to have
an academic appointment so no progress in analog modeling
(in my view). I am an old guy. Back when I wrote physical
design programs for LSI Logic, fab employee attitudes were
that letting scientist types near fabs would reduce yields.
Add into this INTC with the help of the government trying
to be a monopsony (only one buyer).
Bob Widlar wrote lots of application notes for National Semiconductor from Puerto Vallarta, Mexico (for example, AN-446, TP14, TP15... I know there were others but TI seems to have watered down some of the content from National) -- but then again, he was Bob Widlar.The gap between software engineer and hardware engineer is getting larger. Some software engineers work from home or work in tropical islands.
I think there is a simple solution to the US semiconductor problem.
Namely, move CS departments back into physics the way they
were in 1960s. Some examples: Stanford, UCB and MIT EECS
departments will not let scientist types have academic jobs. Best
example is Spice creator Larry Nagel not being allowing to have
an academic appointment so no progress in analog modeling
(in my view). I am an old guy. Back when I wrote physical
design programs for LSI Logic, fab employee attitudes were
that letting scientist types near fabs would reduce yields.
Add into this INTC with the help of the government trying
to be a monopsony (only one buyer).
About the same age as Warren BuffettDid he say Morris Chung? Is that the proper pronunciation of Chang? Hard to believe he is 90 years old!
US Semiconductor Issues:
Lack of manufacturing talent
Cost difference between Taiwan and OR is 50% more
Arizona fabs were at the urging of the US Government
US semiconductor onshoring is a wasteful exercise in futility
US best design capabilities in the world
Taiwan is not nearly as advanced as US companies in design
Morris assumes there will not be a war between China and Taiwan
Using Oregon TSMC fab as an example is old news?
There is a historical regularity where the top semicondcutor company of one decade declines sharply the next. It happened first with NEC, then Motorola, and I think Intel will decline fast based on those examples. But Intel backed by government money is really a different business, more of a Chinese-style SOE (state owned enterprise).The Japan semiconductor industry is a good case study. Japan went from a leader to a follower during my career. I spent a lot of time in Japan early in my career, very little time at all in the second half.
Lots of EE graduates goes to software companies
In this era Bob Widlar would've been shown the door by some HR dork long before getting the chance to build anything.