Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/cramer-calls-on-u-s-to-build-factories-to-address-chip-shortage-unemployment.13727/page-2
        )

    [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] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Cramer calls on U.S. to build factories to address chip shortage, unemployment!

ASMC 2021 is organizing a panel discussion (May) to look at these issues (where will fabs be built and why). There are many ways to think about the past and many good points were made in this thread.
I would argue that the 200 to 300 mm transition was not kind to US manufacturers. Most did not make the cut. Additionally, in the era of the megafab (running on 300 mm), anything "game changing" has to be large and requires a lot of money and a lot of people.
The megafabs need tons of labors and Taiwan and Korea have proven very good a synchronizing thousands of engineers and PhDs and make them go in the same direction. We can match these countries in terms of innovation but not in execution, especially at that scale.
Finally, I would agree that the GE culture which has infected nearly all manufacturing activities in the USA has been very detrimental in the long run. It boils down to a Faustian bargain: command high prices through monopoly positions or get out. Most got out and here we are today. Short of WW3, IC manufacturing is unlikely to ever come back and if it does in a small way, it will be because Samsung and TSMC have run out of native talent.
 
thanks, I am a recovering wise ^@! and they keep taking my chip away at the meetings. You are right we not only gave away the leadership in the core thing the little we gave up the many little things needed to make the core thing - willingly. "Not strategic, not our core competency." Look at GE, the leadership there was a what to do playbook in the day, well retrospectively maybe not.... I guess it turns out making hard things to make might be important to somebody someday. The Government could help for goodness sake we are facing down gov sponsored competition from others. My concern is that other than where they are needed they will try to "help" - sometimes you have to say - for goodness sake please stop helping... sometimes. - Semis are a strategic initiative gov sponsored now in Taiwan, Korea, China, Europe so... hey maybe we might be interested too.. Where i worked once we were in love with Boston Consulting group - that worked out swell. As Daniel said it was not for lack of money - it was the money was used to appease "the investors" vs spending on getting ahead in technology. What I saw was our customers were no longer the customers, wall street were the customers. Big difference... you can only sell your seed corn so many times and then... uh ohhh... turns out all that stuff and technical talent are important after all - "lets just get it back, it can't be that hard" says the finance guy who knows the price of everything but the value of nothing. Remember one of Dilbert's pointy haired bosses key management principles -"anything I don't understand must be easy."
and I don't think it was a conspiracy either it was all out in the open it was a strategy. The management were doing it on purpose, the government was incentivizing it (or dis-incentivizing building or expanding in the USA. But its like the automotive companies and I do have a question about that. Why can TSMC and Samsung and others build fabs here and make them work but other than Intel (arguably if its working or not) and TI - no NA based semiconductor companies have had much interest in NA for a long time. So... is it something in the management culture? is it the finance -wall street orientation of NA based companies? What gives - why can they do it here but we cant do it here?
 
@kg5q I ran into hundreds of people like you in my 30 years in NA chip making, all are shaking their heads. Like others, I am now pursuing opportunities in Asia as there is little to no native growth here, TI, SAS and GF exceptions. Intel is kryptonite for all external suppliers so one is unlikely to go there for opportunities. The gravity of Taiwan and Korea is such that we are now a satellite. Geostationary is the best we can hope for.

There are dozens of factors that played into the situation, some important, others less:

1- Excessive growth of finance and financial metrics - See "Trade Wars are Class Wars" for a great overview. Long story: the primacy of the US$ to settle payments outside of the USA and the important of treasuries is acting as a permanent break on manufacturing.
2- No coordination between states led to extreme decentralization of fab locations = not cost competitive. It is cheaper to travel business class to Taiwan, stay in a nice hotel for 3 weeks than to make the equivalent # of calls in the US. Moreover, things like shipping costs to cleaning facilities, repair etc are all much cheaper since the support factories are right next door. In comparison, a factory like GF Malta has to ship parts to Texas and Arizona for reconditioning, which means more expensive carrying and shipping costs.
3- I do not want to go there but state of science education in the USA is lamentable. We are still not sure if we should study evolution in HS curriculum! As a result, 75% of all graduate students are foreigners. There is nothing wrong with foreigners per se (most stay here) but when it comes time to voting for a new science park, our polity is not well informed. See anti-VAX as latest example.
4- Political paralysis is helping vested interests who like the status quo. We are unable to do the simple things so forget about the hard ones. My favorite example: we cannot eliminate the penny and replace the paper $.
5- Short term vs long term mentality. Another favorite example: then-Governor Christie vetoing new tunnel under the Hudson because of short term electoral considerations. Result: flooding of the extant tunnel during Sandy that led to widespread corrosion issues resulting in traffic problems that are still going on. And the new tunnel is not even started!
6- Regulatory costs, permitting, EPA, OSHA and most importantly healthcare. Puts us at a permanent disadvantage when each worker costs $13-$15k per year before walking in the door when that burden is shared collectively in Taiwan and Korea.
7- and many others - marginal cost of software is zero so investments go there - Amazon, Tesla, FB, Apple do not care one bit where chips are made - workforce trending old - semi in USA has a bad reputation for layoffs...

We do better with entrepreneurial companies (non-existent in this space) and higher pricing (we love monopolies). Short of Musk investing in a large SiC factory or Bezos going after carbon nanotubes, I am keeping my passport current - there is much Korea and Taiwan is my future.
 
you are spot on. A good article about the financialization of American business is above. The also reference a great book I read called "makers and takers" which talks about this. A person who used to work for me called me (before Covid) from the American Airlines Lounge at Ohare and said "guess what i am looking at?" I donno - a copy of Harvard business review - that says "don't let metrics undermine your business!" - well.. really?

yeah I saw it and sprayed my coffee out... here are the high priests of metrics and data everywhere never enough - data data everywhere and not a drop to drink... saying maybe we went too far... really? its like the firemen who start fires so they can be there to put them out. Pay us to train you to put metrics on everything - now woops now its too much maybe those pesky customers were important after all. Einstein said " not everything that can be counted counts, and not everything that counts can be counted." Motorola Semiconductor no longer exists because of that sentence. Focus on everything except what we needed to win. We minored in the majors and majored in the minors until it was too late.
 
I'm not convinced Intel fabs can make chips like TSMC does. TSMC has perfected their process recipes over the last 30 years working with a huge ecosystem of customers, partners, and suppliers. You can't download that and replicate it. UMC, GF, and SMIC have all stalled trying to FinFET it with TSMC. What makes you think Intel can do it and better?

And for those who think Intel can spin off their fabs like AMD did think again. That was a one time deal that involved a lot of back room dealing. And remember, IBM literally paid GF to take their fabs. If Intel wants to sell their fabs they should convert them into condos.
I hear this a lot and I am afraid it's a bit misleading. IBM didn't really pay GF to take their fabs. Essentially IBM gave them their fabs for nothing and in return GF would keep substantially all of the employees, continue their operations and most importantly they had to do advanced node development and then make adv node chips for them. The 1.5bil was actually IBM paying GF for 2-3 years of next node development, which they did and then once that $$ was spent they threw in the towel.
 
Are we really sure Intel is the only one who can save US semi industry? I hope we can have some out of box thinking and new strategy instead of being forced to choose Intel. On the other hand, unless Intel comes back to the business to make chips for other companies and becomes very competitive, it won't satisfy the need from AMD, Nvidia, Qualcomm, IBM, and scores of DoD suppliers.
What exactly does saving the US semi industry mean? To me it has to include advanced node development and manufacturing. Who else in the US can do next node development. No one but Intel has the people or the money to do it. IBM did it until it became too expensive for them. GF tried to do it with the IBM dev team remnants and guess what...too expensive for them (a little more complicated than that but still generally true). So who else could possibly do it in the US?

If saving the industry just means manufacturing newer node chips on US soil I'm not convinced there's really a problem other than the very low volumes that the DoD needs which could be done by SEC, Intel or TSMC if they build their tiny fab in AZ. Again the US govt missed their chance to secure their supply when they let IBM and then GF get out of advanced nodes.
 
"No one but Intel has the people or the money to do it."

That's why some people, including me, is thinking why throwing taxpayers' money onto Intel will solve the problems that Intel hasn't been able to solve for the past several years.

Intel does have a lot of money, $23 billion cash alone at the end of 2020.
 
Every government project I have worked on or abserved was a mess or wasteful and inefficient, sometimes to the extreme. I did have the pleasure of having a look at the communications department at UC Chico years ago which was very well run while being very innovative, efficient and creative in getting things done in a totally superior manner. Every place else in government, including a national lab, was and is a money, talent, time and resource wasting joke. Don't look to the government for answers, you must work despite the government and hope for the best.
 
Every government project I have worked on or abserved was a mess or wasteful and inefficient, sometimes to the extreme. I did have the pleasure of having a look at the communications department at UC Chico years ago which was very well run while being very innovative, efficient and creative in getting things done in a totally superior manner. Every place else in government, including a national lab, was and is a money, talent, time and resource wasting joke. Don't look to the government for answers, you must work despite the government and hope for the best.
I don't think government "project" would work here. On the other hand, government "program" or "policy" might help. Don't you think Taiwan government played a role in the success of TSMC and Taiwan semiconductor/computer industries in general?
 
"No one but Intel has the people or the money to do it."

That's why some people, including me, is thinking why throwing taxpayers' money onto Intel will solve the problems that Intel hasn't been able to solve for the past several years.

Intel does have a lot of money, $23 billion cash alone at the end of 2020.
That is fair, they pour a lot of money into R&D and have struggled the last few nodes for sure. But the flip side of that is $23B in cash will only build one new fab these days. If they or anyone else in the US were to try and become a leading edge foundry (not just an IDM) they would almost certainly need help from the gov't because it's not going to be profitable for a while when you need to build new fabs and catch up on the technology. To be clear, I wasn't saying that the gov't should pour money into Intel, I was just saying that they are the only US semi manufacturer that could possibly do it. US missed the boat long ago, they never should have let IBM close shop, they were THE leading edge foundry for the US govt.
 
Back
Top