Topic: TSMC says efforts to rebuild US semiconductor industry are doomed to fail (Read 246 times)

Real nonsense, I agree!
In order to be sure, it is quite enough to look at the product life cycle at such an enterprise! There are materials on YouTube, I strongly recommend that you get acquainted with them. So, in order not to waste a lot of time, the critical influence of a person in this process is minimal. Everything else is 101% automated, with 100,500 sensors and quality control circuits. What people ? There is a real participation of people, in production, well, let it be 5%, and then I am sure of a clean process, or rather the life cycle of chip production.
But I agree that in the scientific research part, the "percentage of people" is higher. But these are unique specialists with great motivation, of which there are only a few!

I have a hard time believing TSMC is letting a guy paid with 2$ an hour drive a manitou through their warehouse with 1 ton of chips hanging 5 meters in the air. 


This is where you are mistaken, by the same logic Ferrari would have lower profit margins per car than Renault.
But let's stop speculation since TSMC is a publicly traded company so:


https://investor.tsmc.com/chinese/encrypt/files/encrypt_file/reports/2022-07/95e05d031e382c1beaf2b7f088e2dee32e13a0bc/2Q22EarningsRelease.pdf


Apple is at 25%, just an example, Amazon is at 4%.

I believe when he meant cheaper labour, he meant people who do not do the high end parts of it, but more like "carry this box here to there" type of people. But I agree, it was never something that will cost cheap, it is a high cost job and the results are not as profitable as any other business.

The margins are low because it's chips, and you can't just produce a billion of them in a day, meaning you are selling it in limited orders and even with high demand, there is a logical max price for it, you can't just charge 10k per chip right? So that means you can increase the price, but eventually it will be limited. However, I do agree with you that it can be done, it is not as difficult as they make it out to be, it is just another business.

For a start - we are talking about Taiwan - NOT China. Despite what China's Leaders say/think about the matter for the foreseeable future Taiwan is NOT part of China and will remain so. Someday that may change but won't be anytime soon.

China has ZERO advanced node chip making capabilities, their 'bleeding edge' is the 7nm node and even there it is problematic for them because China has never had access to the still by any standard fairly advanced lithography equipment needed to improve their yields. Smaller than 7nm is out of the question for them and with the chip technology embargo in-place even the 7nm node may no longer be possible for the (yes, 'the' as in their only one that was capable of 7nm) Chinese Foundry to use

~~ Silicon Valley does not have to 'catch up'. Why? Simple: Foundries like TSMC et al are not the ones who are designing the chips and the basic processes used to make them. Companies like Micron, Cisco, Broadcom, AMD, nVidia, etc design them. Recently even Intel has shifted production of some chips to TSMC.

Foundries are given the chip designs by their customers who work with the foundries to implement their designs into chips using whatever process node they want to use that the foundry can provide. Guess where most of the actual design work for leading/bleeding-edge chips is done - the US, Europe and Japan. Not Taiwan (and certainly not China).

~~ Manufacturing cost is a different story. All aspects of chip making is heavily automated as there is no other way to do it so that bit is out of the picture. Oh, and just *where* is the equipment made? All advanced process equipment is designed & made in, you guessed it - the US, Europe, and Japan. Now, yes Taiwan does have lower labor costs but that is their only advantage.

~~ Hardware Patents - again, most are held by US, European and Japanese owned companies with #1 on that list being the EUV lithography systems. Now Process Patents - that's a different story. Process Patents are a Foundry's highly optimized Secret Sauce but if need be there are always ways around that with the most obvious being Licensing. Barring that, just do things in a significantly different way. Ya know, like Intel, GF & Samsung have done with their Foundries. The whole point of TSMC opening a Foundry in the US is simply to have a long-term safe place to operate that is (mostly) free of major political unrest with zero chance of China causing problems.

its not just that wages in taiwan are half that of american (50% discount to import rather then create domestically)

but also if they time their contracts right they can get another 10%-20% discount via playing forex movements

along with this
if making domestically american companies then have to pay lots of different taxes and costs. where as importing something. they have more loopholes to tax dodge to get further discount

1. If you think someone in this industry is working for 10$ a day you're deeply mistaken, they are highly paid and everyone gets offers from different companies, even Bitmain tried to bribe TSMC workers to come work for them and the whole poaching scandal nearly resulted in them being cut
2. If chips are in high demand if prices are skyrocketing how are the margins low?
3. 10-15 years? Let's be serious. Besides none of the materials in chops are that impossible to get as obviously, the US is still manufacturing chips, Europe is manufacturing chips, and there won't be any huge next-day change where Intel needs 1000 tons of hafnium per day starting tomorrow.


Or maybe not:

For Chip Makers, the Flip From Shortage to Glut Intensifies
https://www.wsj.com/articles/chip-makers-cut-costs-as-demand-slump-supplants-pandemics-chip-shortage-11667560050

The United States has an officially voiced problem - the possible difficulties of providing industry, and the military-industrial complex, in the first place, with the necessary high-tech components. Difficulties - from insufficient volume, to interruptions in supply, up to a complete stop.
Given that the United States is still one of the most high-tech and richest countries, while already having experience in its own production of chips, building, or rather, restoring chip production, is only a matter of time. In this case, "two birds with one stone" will be killed. The second hare is pulling the "blanket" of chip production (simpler chips, for the automotive industry, household appliances, etc.) from China, in order to weaken China's position in the high-tech market. By the way, from China, American high-tech industries will also migrate to the United States or nearby countries.

Not China, but Taiwan is the biggest producer of the  Semiconductor chips in the entire world. US simple can't complete with them due to some reasons,

1. High labor cost in US for skilled workers
2. The margin is not attractive enough considering the competition
3. Establishment of raw material procurement is extremely difficult as the companies supplying raw materials are already under contractual agreement for next 10 - 15 years

It can only succeed if any US company can start the production in a low cost country and get huge tax concession from the government for employment generation. It is not easy!

This statement of efforts to rebuild US semiconductor industry being doomed to fail sounds a bit self-serving because what would this mean for TSMC if in fact it is successful?  It's not in their best interest if efforts to rebuild US semiconductor industry are successful.

If it's something for 'defense' then nothing is impossible for the US. They put a lot of effort when it comes to this aspect of their country more than anyone else in the world. They just have to get the right people to do the job for them and they're basically all set. It'll be a hard feat, yes, but if these guys are decades ahead of their competition when it comes to military technology, then it just means that they understand enough about electronics and the chips that they are using to be able to reverse-engineer it - and even manufacture it with the right tools.

For now TSMC can take the crown of being the major producer of lots of advanced chips in the world, but a country such as the US with a bigger economy, a lot of dependable engineers and technicians, and basically an infinite amount of resources to spend can catch up if it wanted to.

The US has the potential to become a leader in microelectronics. 

They know how to attract first-class engineers and other technical specialists to their country.  They can always offer them a highly competitive salary and an excellent benefits package.  They also have the opportunity to build the necessary plants and simulate the entire necessary technological process.  Microelectronics products are used for a variety of purposes (including military ones). 

Therefore, ensuring autonomy in such industries is a strategic task for a country like the United States.

It is of course expected TMSC would say that it’s doomed to fail because if USA achieves this then it would be their end. Why would USA need them if they can build their own chips? Which is what they are afraid and this is not just being afraid of business results, it also means that if USA and Europe could provide their own chips anytime they want, then in a case where China attacks Taiwan, why would anyone care?

If Taiwan is only one producing the chips then everyone would care about such a war because they want their chips, but if they all build their own chip factories then nobody would stop China from attacking Taiwan and that is why TMSC doesn't want that to happen.

I can't believe that the Americans were dumb enough to lose their advantage in chip manufacturing 30 years ago.
Taiwan now has the industrial capacity and the qualified workforce. If a small island like Taiwan was able achieve such economic miracle decades ago then the USA should be capable of rebuilding it's semiconductor industry. Saying that this is "doomed to fail" sounds way too pessimistic.
Perhaps the US chip manufacturers will have to outsource their production facilities in Mexico or Costa Rica, because of the cheap labor.
The main problem here is the US government. You can't just throw 52 billion dollars at the US companies and expect fast results.
There will be semiconductor shortages for sure and expects all electronics to become way more expensive in 2023.

The reason for him saying that is right in the article:


The foundry price was 18 billions, he spent 44 billions on twitter, do you actually think that money was the problem, besides, if there was a shortage of chips, how was that possible in the context of these small margins you claim?

Interesting, to this day I had considered the backing of USA to Taiwan to be almost completely unconditional, I thought it was about keeping the Chinese CCP away from the island inhabited by a free people, I had no considered the fact of the Taiwanese chip production playing a part in this backing, but it makes sense. 

I assume that USA has realized how strategic and important the production of chips is for their national security, they would not want to risk its production by leaving it on hands of a single foreign country, Taiwan or others.

The way I see it, the chip production (and the hardware production) in Asia has been long preferred by giants within the industry like Apple, I thought it was only because the cheap labor force. I wonder how competitive would the American chips will be in retail price and I hope USA does not abandon the people of Taiwan if they succeed to rebuild the national chip industry.

I think the US might still defend Taiwan even if they will succeed in creating their own semiconductor industry - after all, US is a superpower and it's not  going to cede this status without a fight, especially since US military is clearly more capable than any military in the world by far.

And TSMC sounds like they are not happy about the prospects of losing their monopoly rather than being worried about getting invaded by China because the US will lose a reason to get involved.

The chinese chips are terrible compared to TSMC chips. If there was patent infringement, it's not like it matters. There is no global patent police that has jurisdiction.

It takes years and billions of dollars to build these factories. The U.S. tech sector could eventually catch up if they're inclined to do so, with impending Chinese invasion of Taiwan awaiting. TSMC obviously doesn't want the U.S. to become independent on their chip production because it would lessen the reliance U.S. companies have on them. If the U.S. does not have a reliance on Taiwan, there is no incentive for the U.S. to protect them from invasion.

No because entrepreneurs are too busy doing "hot" stuff like blitzscaling a Something-as-a-service business at 200mph only to crash into a tree.


I don't think so, because of minimum wage concerns.


That will be the biggest problem for the US. No I don't think they will infringe on Chinese patents and then be allowed to continue operating in China, so its more of an aerial battle for silicon innovation.

My question is this....

~ Will Silicon valley be able to catch up with the technological advances that was made by China over all these years?
~ Can they manufacture these chips at the same cost that China are doing, because China are almost 100% mechanized?
~ What hardware patents are there on the existing technology, that are owned by China ...and will they allow the US to infringe on that.

This is going to be a long process, with a lot of political struggles to implement this. China will not take this sitting down.. they will sabotage all efforts to get this off the ground, because they want to retain that leverage and monopoly that they have over chip manufacturing. 

If the US produces a lot more of their semiconductors than Taiwan then I think that'll put them more of a difficult position with China than they have now (the US might not need to come to their aid as much, they might do like is being done with Ukraine now rather than it being considered an attack on them directly).

I'd envisage chipmaking becoming either like steel production or many other industries financed by governments (transportation, energy, water). If it's playing an integral role in keeping a government/country secure and functioning it should be easy to source production internally if necessary. Italy is in the top ten of countries that can produce planes and that's directly to support their military efforts (this may be outdated, I was looking at the stats in March).

When wars happen, countries generally strive towards socialism - militaries are "socialist" (heavily funded and controlled by the government), ownership of infrastructure becomes more and more important too there are examples of this from the past like kindertransport).