TSMC’s dominance poses risks to the global economy
wsj.com
wsj.com
Does this mean there are absolutely no alternatives? Likely not, but TSMC are essentially the only game in town. At scale. At quality. With a useful track record of known production characteristics.
The combination of it being a very specialized and expensive industry with high upfront costs, extremely high RnD, cut throat competition and patent landmines means that as time has gone on ther have been fewer and fewer players. The only ones really capable of filling orders for chips using process tech from the last 5 or so years would be TSMC, Samsung, Global Foundries and Intel. Intel just recently opened up to third party production so they wont have an effect on the global supply for awhile.
The combination of it being a very specialized and
expensive industry with high upfront costs, extremely
high RnD, cut throat competition and patent landmine
Obviously, labor costs are much lower elsewhere. But the massive capital expenditures involved mean that labor is a comparatively much lower % of overall costs compared to, say, spinning up a factory to make cheap shirts.
I wonder if Western nations/companies will be spurred by this chip shortage to start prioritizing their own foundries. Like energy independence, except... silicon independence.
The problem isn't that TSMC is the only fab. Its that semiconductor fabrication is inelastic, in other words it cant respond quickly to changing demand. The lead time to start producing a given chip design is fairly long and a fab like TSMC will have its available capacity contracted out in advance. They were already running at capacity before COVID because their 7nm process was very popular.
You cant just build a fab overnight and start producing chips. Just building the place and getting the appropriate equipment installed takes over a year and then there is a lead time for adapting the chip design to your process. You actually design your chips with the proprietary fab process in mind. So if you had contracted with TSMC but they cant fill your order and you have to move to one of the newly available Intel fabs your engineers must adapt your design before you can start producing again. Every step of this process takes time that means that no reaction due to COVID would be showing results yet. There will be a shortage for the foreseeable future.
It's very hard for a company to take the latter seriously (like the supply chain issues that arose due to covid) when it is a very rare event that will likely have other large effects on your market and your competitors also.
Governments stepping in to subsidise/force use of other companies will likely reduce the short term profitability or commercial appeal of your product.
So 2 of the 5 permanent members of the UN is already not going to do this.
And it is not only limited to non-western countries, the same argument holds for Spain with Catalonia for instance.
Real-politik is complicated.
It's best to think of the UN as a meeting place where consensus can form, rather than as a law making body with a well functioning appeal and dispute resolution process for its own deliberations.
I think if I were Taiwan, it would be US bases or we leave the NPT and build some warheads...
My previous comment on a previous post: https://news.ycombinator.com/item?id=26382543
Seems relevant.
Even the US has been very carefully not fully committing to a defense of Taiwan, giving China the impression (according to some reports) that they might get away with it if they just act quickly enough.
Of course, none of this affects non-military threats to TSMC's production, like earthquakes.
If you want to sustain a local cutting-edge manufacturer even when its capacities are redundant or too small to be economical then the pot of subsidies would have to increase yearly and not be spent wastefully. Is there a precedent for that?
We're going to see a repeat of these events 20 years from now when Vietnam's economy grows, thanks to manufacturing relocation, and we're just as dependent on them for various things. Right now the world is hoping and praying that Vietnam will liberalize when money flows in, but it's more likely they'll follow a similar path as China, albeit less powerful.
It took several iterations of territorial expansion for Germany to be considered an actual fightable foe. Austria, Memel, Sudetenland, the rest of the Czech lands, Poland. Even after the invasion of Poland, when war was declared, it was a Phony War in the beginning. The fighting in the West only really started in 1940.
There is a very simple way of telling a country to revert course: deem their passports no longer valid. Fundamentally, the right to move between countries is provided by the International Covenant on Civil and Political Rights Article 12 which says "Everyone shall be free to leave any country, including his own" but then continues "The above-mentioned rights shall not be subject to any restrictions except those which are provided by law, are necessary to protect national security, public order (ordre public), public health or morals or the rights and freedoms of others, and are consistent with the other rights recognized in the present Covenant".
For China, it'd be even easier since they signed but didn't ratify this treaty despite it has been signed in 1966 and in force since 1976.
Although arguably NAND and DRAM are commodities, and US already has a back up strategy ( Micron ).
But what make you think China will "cut off the world" IF that happens?
That's not how this works, that's not how any of this works.
But we're not talking about a corporate takeover. We're talking about a naval invasion of a very densely populated and well defended island. That might involve heavy artillery bombardment and prolonged urban combat, maybe a guerilla war and sabotage actions.
There is no guarantee that the fabs are still operational after all that.
I mean, we're on different political sides, not good vs evil, please acknowledge that.
One of the sides runs a series of concentration camps. Sells organs from political prisoners, etc.
But yes, I suppose it's just politics. Personally, I think genocide is slightly evil, but feel free to disagree.
If, I mean If, China is evil, wouldn't that make any country not declaring war to China evil too?
Independent media across the globe and across political spectrums have reported that China has locked away >= 1M people in concentration camps.
Claiming that the Uyghur genocide is made up is just trolling. These things have been documented for years.
And the Uyghur is not the only minority to be persecuted.
Tell me who is the Dalai Lama working for?
> wouldn't that make any country not declaring war to China evil too?
I think you'll have to ask a philosopher, if evil is a transitive property :)
But my two cents is
(1) NO!
(2) War is not the solution.
Honestly, war is very rarely a good solution. Sanctions a political pressure might be better tools.
"Independent media"
The CCP and Xi in particular have turned "reunification" into a matter of national and personal pride.
More substantially, they see Taiwanese independance as a threat because it could serve as an example to independance movements in Hong Kong, Xinjiang, Inner Mongolia and probably a few more.
And what's wrong with national pride? And in what way do you think waging a war could boost pride? I thought we learnt that after two WWs. You might want to check China's current economy and civilian living status, we are not like some middle east country with nothing to lose, no offense.
An actual peaceful reunification is just not going to happen anytime soon, especially given how things are developing in Hong Kong. And if it involves the military, it's conquest, not a reunification.
> And what's wrong with national pride?
It is all too easily used by populists to gather support based on an "us vs them" mindset. And (like personal pride) it can compel you to do self-destructive things to avoid losing face.
> And in what way do you think waging a war could boost pride?
"We beat them, so we're better than them". Dumb, but people do think like that.
> You might want to check China's current economy and civilian living status, we are not like some middle east country with nothing to lose, no offense.
Then why all the fuss about Taiwan?
A truly great nation could say "look, how about you strike down that silly idea about retaking the mainland, give back our national treasures, and in return we'll reconize you as an independant country?"
That's your speculation, are we at war now?
What fuss about Taiwan? we always declare Taiwan is part of China.
By all means what you're saying is "this man wants war, let's fight him".
> "We beat them, so we're better than them". Dumb, but people do think like that.
Maybe because you think like that and assumes people are like you.
A conditional statement is not speculation. Also:
https://www.theguardian.com/world/2021/apr/12/largest-chines...
https://www.france24.com/en/live-news/20210105-taiwan-says-c...
https://www.businessinsider.com/chinas-military-buildup-atti...
> What fuss about Taiwan? we always declare Taiwan is part of China.
Status quo is that it's not part of China, and hasn't been for well over 100 years. This doesn't hurt China in any way. So why not simply accept that status quo?
> By all means what you're saying is "this man wants war, let's fight him".
I have no idea what you're trying to say here.
> Maybe because you think like that and assumes people are like you.
Quite exactly the opposite.
But I guess you've just arrived at the point where you can't argue your position in good faith anymore.
You think leading US chip designers would send their designs to China?
The well organized state of industrial espionage by China, is a pretty good reason this won't happen.
Furthermore, the rest of the world would be forced to embargo anything related to Taiwan.
The reason advanced chip factories are located in China is that they are expensive and Chinese authorities can't be trusted.
Adding to that is that sending your chip designs to China, is a sure way to get them stolen.
Lots of companies don't sell the latest tech in China. And don't manufacture in china because their designs would be stolen. This is not a problem for cheap manufacturing, but for high tech stuff it might be...
https://www.laserfocusworld.com/blogs/article/14039015/how-d...
And China has about as much right to govern Taiwan as the Habsburg-Lothringen dynasty to govern Czechia.
But you do not seem to deny that China acts in a hostile way towards Taiwan. You switched to "someone else has a dirty conscience too" defense.
"you do not seem to deny that ...", I didn't deny it immediately since it's a bit far fetched to argue comparing to the previous paragraph, we even bombed a small island with no people for twenty years(keyword: Second Taiwan Strait Crisis), but that's history, does Xi want Taiwan back? of course, does Xi want to go down as a tyrant? of course not, look at history, we got HK and Macau back, peacefully, remember that.
On the other hand, recently, Taiwan government refused vaccines from China, I think they really went too far.
While not 100% watching it I feel as though China<-->US relations have whole strategies around Taiwan.
Of course, if war over Taiwan breaks out, I plan to hit up my local computer shop, preferably within the hour.
You will hit up your local draft station instead.
Even if an immediate retaliation would not not happen, it will within years.
But the threat of invasion will be continued to played up in US politics in order to support increased military spending and aggressive industrial policies towards US allies (in particular Taiwan but also South Korea).
It's dangerous exactly because we don't know what will happen.
Then I thought about where I would build one, given this profitable circumstances: 1. Cannot be in (most) of Europe, due to high taxes and advanced and advancing work regulations. Standards are too strict to allow for innovation. Most of low tax Europe are tax heavens unsuitable for high scale industries. Estonia and Georgia give some breathing room, but are friendlier to services. 2. I cannot build it in Africa: high taxes, extreme corruption and high risk of politics affecting the company. 3. I might be able to build it in the US. Being close to the client is not as important as having access to sea. US politics are turbulent and if corporate tax is raised, then there's no competition to be made. 4. South America is a bag. Almost all SA countries have corrupt governments at every layer, but there are a few business friendly ones. Most countries cannot be considered. Panama is already focusing in other things, Uruguay (25% corporate) and Brazil (~34%) cannot compete. 5. I'm unsure about Middle East, I don't see anything there 6. Asia is massive: 6.a. China could happen. 6.b. South East Asia is too erratic compared to Taiwan, investors will go to Taiwan instead, or invest in something else. Singapore is too small. 6.c. Japan is an stagnant economy and there is too much bureaucracy and regulation involved. 7. Australia ~26% corporate tax rate is not going to make the cut. I'm not knowledgeable about regulations.
I probably have missed something, I tried to put onto words my thinking process on why Taiwan might still be the most profitable place to make these industries. The whole world is in the middle of a shortage of this material, yet alternatives to Taiwan don't seem to arise.
Secondly, TSMC is so far ahead of the curve that any competitor will take years and billion of dollars to get to where they are right now. ASML develops the technology in close collaboration with TSMC and I suspect TSMC owns many of the patents required to operate EUV successfully.
The industry is happy with the status quo because TSMC offers the best technology at the best prices.
The (first) Foundation was set up on a small, insignificant planet, surrounded by belligerent powers. Its continued existence was made possible only by playing those powers off against one another.
And the Foundation itself: highly advanced technology that those very powers came to rely on...
I mean to me this sounds as if TSMC could easily raise prices and increase profits and their customers will still buy the exact same amount... Because they have to.
Nothing is “priced in” and prices really highly on historical volatility
One way to tell is to look at how nearer term options are priced than LEAPS are. LEAP pricing is telling you nobody knows a damn thing and they are consistently undervalued.
TLDR: not there, at least yet.
I'm not sure how true that is anymore. Companies trying to manufacture elsewhere have to cope with 101 issues that increase costs. Everything from JIT logistics not working to lack or workers to uncooperative governments. China is getting more expensive and some low skill work is going elsewhere. But for most companies, setting up elsewhere means prices much higher than the market will tolerate.
China is also, again by far, the largest e-commerce market and the largest mobile payments market on Earth.
I apologize for sounding like an infomercial, but these are basic economic facts that one would expect a reasonably well-read individual to be aware of. I would advise that in future you get better informed before coming here to advertise your ignorance.