Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/semiconductor-dependency-imperils-american-security.16254/
        )

    [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
)

Semiconductor Dependency Imperils American Security

prime007

Active member
Eric Schmidt (ex-CEO of Google) and Graham Allison (Harvard professor of government) wrote a commentary in the WSJ arguing that American national security is at risk if it is completely dependent on Taiwan for advanced semiconductors. You can read their arguments in the link below:
 
Isn't the noble gas shortage due to the Ukraine war going to hobble the industry in a few months? Russia must be a very serious threat to security if we're sending Ukraine a CHIPS Act worth of small arms and overpriced MANPADS while simultaneously giving up access to critical resources and letting our infrastructure continue to rot.

Either that, or Russia is doing to America in Ukraine what America did to Russia in Afghanistan in the 80s. That would mean we're in an era of great stagnation and corruption, for which Trump and Biden are the perfect avatars. Joe was elected on "nothing will change" and he has totally kept that promise, which means things continue to decline as they have been for decades. It also probably means that a few big players will get massive tax breaks to build Potemkin semiconductor villages in political hotspots like OH and AZ, and that's that.
 
Isn't the noble gas shortage due to the Ukraine war going to hobble the industry in a few months? Russia must be a very serious threat to security if we're sending Ukraine a CHIPS Act worth of small arms and overpriced MANPADS while simultaneously giving up access to critical resources and letting our infrastructure continue to rot.

Either that, or Russia is doing to America in Ukraine what America did to Russia in Afghanistan in the 80s. That would mean we're in an era of great stagnation and corruption, for which Trump and Biden are the perfect avatars. Joe was elected on "nothing will change" and he has totally kept that promise, which means things continue to decline as they have been for decades. It also probably means that a few big players will get massive tax breaks to build Potemkin semiconductor villages in political hotspots like OH and AZ, and that's that.

"Isn't the noble gas shortage due to the Ukraine war going to hobble the industry in a few months?"

It seems major foundries are not worry about that. TSMC stated that they are getting the gas they need.
 
I read the Allison-Schmidt piece this morning. I was disappointed in it. I don't understand starting out the article with the infamous Morris Chang quote. A quote from an expert with far greater semiconductor expertise than either author says the US (and by implication the Euro) efforts to make headway in chip leadership are for naught? That was confusing at best. I was also not impressed by the statement that Intel and GF "excel at producing slower chips". Not by design, but by lack of success in developing the state of the art. And then they close by saying the US needs "a national effort". I'm not sure what that means. I suppose it could be something like the National Institute of Health, which operates national labs which investigate directed research, and funds research programs in academia and private companies. This looks like the central planning model, which I think is the wrong answer. Far too much dependence on central planners.

I was hoping A-S would focus more on making it easier for research, development, and production to be done in the US. Lower corporate tax rates; right now as a country we are fixated on raising them. Reduce useless cultural and environmental paperwork and approvals. Especially approvals. Make it the objective to get things built most efficiently, not to figure out how to slow them down.

Dramatically improve the US primary and secondary educations, which are mostly uncompetitive, and regulate the hugely expensive mess that our renowned universities have become into streamlining. Schmidt knows a lot about academia. Talk about something that needs a national effort. Reward STEM education with a very substantial national scholarship program. And, as it stands, over 40% of STEM students in the US are immigrants, yet we do not have a strategy for ensuring they can easily stay here once they graduate, even if they want to stay. That's dumb. Attempting large-scale semiconductor development without enough qualified scientists and engineers is a fools errand.

Formulate environmental regulations that are designed for success while protecting the environment, not procedural slowdowns, approvals, and massive paperwork. I was recently reading what we've done to Space-X:


Disgusting. I wonder what Intel has to do for its planned Ohio fabs? I'm guessing TSMC in Taiwan has a much straighter line to walk.
 
I strongly believe that one one of the reasons, the biggest reason, TSMC is diversifying into different countries is political, expanding the silicon shield to protect Taiwan. Japan is a key foothold and of course the US. I hope India and Europe are next. There is strength in numbers, absolutely.
 
I strongly believe that one one of the reasons, the biggest reason, TSMC is diversifying into different countries is political, expanding the silicon shield to protect Taiwan. Japan is a key foothold and of course the US. I hope India and Europe are next. There is strength in numbers, absolutely.
I thought TSMC just said no to Europe, at least for the time being?
 
I thought TSMC just said no to Europe, at least for the time being?

TSMC said not at this time, which is what they said about the AZ location before the deal was signed. TSMC will definitely build in EU the question is where and the answer is where the money is.
 
I read the Allison-Schmidt piece this morning. I was disappointed in it. I don't understand starting out the article with the infamous Morris Chang quote. A quote from an expert with far greater semiconductor expertise than either author says the US (and by implication the Euro) efforts to make headway in chip leadership are for naught? That was confusing at best. I was also not impressed by the statement that Intel and GF "excel at producing slower chips". Not by design, but by lack of success in developing the state of the art. And then they close by saying the US needs "a national effort". I'm not sure what that means. I suppose it could be something like the National Institute of Health, which operates national labs which investigate directed research, and funds research programs in academia and private companies. This looks like the central planning model, which I think is the wrong answer. Far too much dependence on central planners.

I was hoping A-S would focus more on making it easier for research, development, and production to be done in the US. Lower corporate tax rates; right now as a country we are fixated on raising them. Reduce useless cultural and environmental paperwork and approvals. Especially approvals. Make it the objective to get things built most efficiently, not to figure out how to slow them down.

Dramatically improve the US primary and secondary educations, which are mostly uncompetitive, and regulate the hugely expensive mess that our renowned universities have become into streamlining. Schmidt knows a lot about academia. Talk about something that needs a national effort. Reward STEM education with a very substantial national scholarship program. And, as it stands, over 40% of STEM students in the US are immigrants, yet we do not have a strategy for ensuring they can easily stay here once they graduate, even if they want to stay. That's dumb. Attempting large-scale semiconductor development without enough qualified scientists and engineers is a fools errand.

Formulate environmental regulations that are designed for success while protecting the environment, not procedural slowdowns, approvals, and massive paperwork. I was recently reading what we've done to Space-X:


Disgusting. I wonder what Intel has to do for its planned Ohio fabs? I'm guessing TSMC in Taiwan has a much straighter line to walk.

"I was also not impressed by the statement that Intel and GF "excel at producing slower chips".

The original text was: "U.S. companies such as Intel and GlobalFoundries excel at producing slower chips that are used in everything from televisions to tanks."

It's funny. I don't know GlobalFoundries but I'm sure such opinion is not what Intel can accept. Many Intel processors for servers and PCs are very fast. Otherwise why there are so many datacenters are using them? The chips for televisions and tanks have different performance goals, different reliability requirements, and different national security consequences. Mixing them in one bag and try to suggest a policy is a flowed approach.

For example, 45 million TV were sold in US in 2021 alone while less than 11,000 units of various models of M1 tanks have been made since 1979. Government can and should subside those chips needed for M1 tanks especially the government is THE only customer. On the other hand, how can government keep subsiding the chips manufacturing for consumer products such as TVs, cars, washing machines, microwave ovens, or coffeemakers? It's not sustainable!

Morris Chang once said he thinks to use government subsidies for domestic chips manufacturing on national security related applications does make sense. But he believes it won't work for semiconductors destined for consumer applications.

I'm looking for opinion leaders who really understand the economy and semiconductor industry to make some useful and practical suggestions.
 
Last edited:
Back
Top