21 ms·
So, as someone who used to be an engineer in semiconductor manufacturing facilities, this reminds me a lot of Sematech. That was a joint venture between Americ
by rossdavidh 6y ago
So, as someone who used to be an engineer in semiconductor manufacturing facilities, this reminds me a lot of Sematech. That was a joint venture between American companies in the 90's, when there was a concern that we were falling behind Japan. I'm sure there will be people who try to claim otherwise, but my view from inside AMD (at that time still doing its own manufacturing) was that nothing much ever came out of Sematech. It was a lot of hype, but not much result.
If, at the time of the founding of TSMC, you had tried to get a consortium of companies, much less nations, to agree on the idea of a semiconductor foundry, you would not have been able to get agreement to do it. Consortiums can help to pool resources when everyone knows what needs doing, but even Intel has had a problem getting the latest generation of semi manufacturing going, and their problem sure wasn't a lack of resources. Sometimes the problem is the speed of innovation, and the more parties you have involved the harder it is to get agreement to try any given thing, especially if it is innovative (i.e. not certain to work).
I wish them good fortune, it would be good for the stability of the world economy not to have all the world's advanced semiconductors made in one region, but I am skeptical of whether or not this can make a big difference, no matter how much money is spent.
- andy_ppp 6y agoI agree, their problem wasn't lack of resources it was they stopped believing in Moore's law. They started to believe that 5 and 3nm advancements were not coming and didn't make the hundreds/thousands of new inventions needed to get them there. At least that's what I get reading between the lines here... Jim Keller talks about this in this in the Lex Fridman podcast: https://youtu.be/Nb2tebYAaOA?t=1892 https://youtu.be/Nb2tebYAaOA?t=1892
- guai888 6y agoWith the success of Apple M1, TSMC and Samsung will grow faster than Intel. Taiwan,Korea,Japan and US will control the semiconductor industry in the near future unless a game changing technology is introduced. Taiwan semiconductor companies's stocks are going banana after US slap sanction on SMIC. UMC stock has not changed much in years jumped 250% in last few weeks. (UMC invented foundry as service, then TSMC copy them)
- thanksforthe42 6y agoI'm only half kidding when I say Apple doesn't consider themselves a US company. They consider them Apple. It's not like they are cooperative like US automakers, and they have done China's will.
- guai888 6y agoChina CCP is a snake. Get in bed with them and they will bite you one day. It is just their nature. Apple will learn one day.
- birdsbirdsbirds 6y agoIf you look at Apple's apple, you will notice that it is already bitten. It was not the snake, but Eve.
- BlueTemplar 6y agoWasn't it Alan Turing ? (Though I guess he's kind of an 'Eve' of computing...)
- Daho0n 6y agoNo more so that any other big country. At least they don't bomb thousands of people a month.
- MangoCoffee 6y ago>If, at the time of the founding of TSMC, you had tried to get a consortium of companies, much less nations, to agree on the idea of a semiconductor foundry, you would not have been able to get agreement to do it. Consortiums can help to pool resources when everyone knows what needs doing TSMC or its ecosystem is already like a consortium. this semiwiki article point that out. https://semiwiki.com/semiconductor-manufacturers/intel/293645-how-intel-stumbled-a-perspective-from-the-trenches/ https://semiwiki.com/semiconductor-manufacturers/intel/29364... "TSMC, for example, has a massive ecosystem of partners and customers who together spend trillions of dollars on research and development for the greater good of the fabless semiconductor ecosystem. There is also an inner circle of partners and customers that TSMC intimately collaborates with on new process development and deployment. This includes Apple of course, AMD, Arm, Applied Materials, ASML, Cadence, and Synopsys just to name a few." TSMC hand out their kit to their partners on their node process. bottom line: fabless no longer just simply toss design over to the foundry.
- guai888 6y agoWhen TSMC was founded, they were not in foundry as services business. UMC actually invented it and TSMC joined the game later. Some people said UMC made the wrong decision to invest in China too early. That effort took too much of UMC's resource. TSMC focused their resource in Taiwan and was able to move forward faster in technology when compare to UMC. When UMC made the move to China, Taiwan government were not happy, so I believe government direct resource to TSMC in retaliation to ensure TSMC has the edge. it is no wonder TSMC won the technology race.
- cromka 6y agoJust wondering if in the light of this announcement and the China's attempts at taking over the TMSC control, wouldn't it be better if TMSC relocated their HQ to Europe? Not saying they should move completely, but to strategize to limit their operational risks. I am sure they'd receive some nice grants from EU.
- FooBarWidget 6y agoChina taking over TSMC control. How, exactly? China isn't looking to invade Taiwan any time soon, unless America puts bases in Taiwan first. And an invasion of Taiwan would immediately result in the destruction of TSMC, thus invasion is not a viable strategy to gain TSMC.
- eloisant 6y agoIt can work if it's countries who build a company together, like European countries created Airbus for example.
- mrtksn 6y agoEurope has some successful international projects like CERN or Airbus. Besides, what are the other options? Europe is in this strange place where collectively account for enormous wealth and power but individually no country is a match to the big boys like USA/China when simultaneously are super unoptimised due to cultural/lingual differences. The younger Europeans I met were all very eager to learn English, so who knows, maybe as the UK is out of the EU the English language can become the lingua franca with less resistance as it is no ones language(Well, there are the Irish but it's a small country). At the end of the day, those large Eurpoean projects bring a lot of good. Be it political or simply keeping some talent in Europe. I am also hopeful for some competition from the UK :) Since the collapse of the USSR, finally there's a competition.
- aledalgrande 6y agoTotally agree. Want to add another European project to the list: ESA.
- tcgv 6y agoWhich, among other projects, created / operates GALILEO, the European GPS :)
- TacticalCoder 6y ago> Europe has some successful international projects like CERN or Airbus. Besides, what are the other options? If we have to go back to mentioning Airbus (1970) and CERN (1954) to find successful European projects, I think this show how big of a serious problem there is in Europe no!? P.S: a lot of the young in Europe are learning english because they want to get the f--- out the EU as soon as they can. I'm european and I'm raising my kid in a british college since years for that very reason.
- norwegianwhale 6y agoIf there ever is a promising project from europe (a.1) it bootstraps / gets venture capital (a.2) Leaves after 2 years for the US / gets acquired and the fiscal headquarter moves to Delaware (Stripe) (b) Gets absorbed by an old blue chip, like Siemens, Telefonica, and dies a slow death. (c) Becomes highly stymied by bureaucracy, and the initial momentum wanes without winning much market share (Zalando, Deliveroo, Idealista). It's called the market of lemons problem. Only the least productive projects want to have their hands tied and financed by the government.
- thunkshift1 6y agoLets not forget that some of the core tech that sits at the heart of semiconductor manufacturing (asml) is already European. It will be daunting, but something just might come out of it.
- stiray 6y agoLinux kernel? KDE? QT?
- ISL 6y agoI suspect that a long-term path out of this sort of thing has two key ingredients: 1) The commoditization of foundry technology. 2) The hardware equivalent of the reproducible build. The first is necessary for nation-states to feel comfortable building and maintaining national-security infrastructure. The second is necessary for the rest of us. If we are able to purchase an open-hardware CPU and verify its integrity through hardware/metrology checksums in addition to a software checksum, society will have made a step forward. The hardware checksums might be expressed through moire-patterning visible in x-rays, for example. The CPUs on which such technology might be developed need not be fancy - if a proof-of-concept works on cheap chips, the extension to the high end is apt to be feasible.
- rossdavidh 6y agoThe commoditization of foundry technology, at least for the smallest geometry (i.e. highest performance) semiconductors, is going in reverse. There are fewer and fewer companies that can perform at the leading edge, and now you could probably count them on one hand without needing all the digits.
- jacobwilliamroy 6y agoSematech would have succeeded if the U.S. government had any interest at all in protecting its domestic manufacturing, or if Japan had better labor protections. Now everything is offshored and I can't fly my drone on military bases because all computer chips come from China. Pretty bad move in hindsight.
- sangnoir 6y agoEurope already has successful semiconductor companies, including: Philips, Nordic Semiconductor, NXP, and the crown-jewel the entire world depends on: ASML. Aside: how did the Netherlands become such a semiconductor powerhouse?
- bsder 6y ago> So, as someone who used to be an engineer in semiconductor manufacturing facilities, this reminds me a lot of Sematech. However, the prior initiative was VHSIC and that basically produced the modern semiconductor industry (including the software tools!) in the US. I was always under the impression that the difference between VHSIC and Sematech was simply that VHSIC threw around a LOT more money.
- cabalamat 6y ago> nothing much ever came out of Sematech. It was a lot of hype, but not much result It may be that nothing much will come out of this initiative, but what choice does Europe have? Europe simply has to have control over its computing/communications infrastructure, if it wants to be truly independent. As Macron recently said ( https://geopolitique.eu/en/macron-grand-continent/ https://geopolitique.eu/en/macron-grand-continent/ ): > In concrete terms, this means that when it comes to technology, Europe needs to build its own solutions in order not to depend on American or Chinese technologies. If we are dependent on them, for example in telecommunications, we cannot guarantee European citizens the secrecy of information and the security of their private data, because we do not have this technology. [...] > The same goes for the extraterritoriality of the dollar, which is a fact and not a new one. Less than ten years ago, several French companies were penalised to the tune of billions of euros because they had operated in countries that were banned under American law. In concrete terms, this means that our companies can be condemned by foreign powers when they operate in a third country : that is a deprivation of sovereignty, of the possibility of deciding for ourselves, it weakens our position immensely. > Unfortunately, we grasped the full implications of this when it came to the discussion on Iran. We Europeans wanted to remain within the framework of what we call the JCPOA. With America’s withdrawal, no European company was able to continue doing business with Iran for fear of sanctions from the United States. Hence, when I talk about sovereignty or strategic autonomy, I am bringing together all these issues, which at first glance seem very different.
- megaframe 6y agoI wish they'd just pushed for TSMC to open a fab in Europe, with incentives and protections. Micron has one in Italy that is basically shut down. I was once told that to avoid payouts, required by Italian Law, for letting employees go the factory had been put on a sort of "pause" everyone sent home but not let go. If things like this still go on I see no hope for a European electronics/Semi manufacturing industry to ever take hold.
- cma 6y agoDoesn't AMD mainly exist as it is because US military contracts had second source requirements imposed on Intel? There were lots of lawsuits and stuff as well hat went into their IP sharing, but that military stuff in some ways made Intel+AMD into a partially government-coordinated consortium. The EU doesn't have a unified military in the same way as the states.
- athms 6y agoIBM made AMD the second source for Intel's CPUs used in the IBM PC line of computers.