@Tamza2- a good point. I would have thought I had learned by now that its dangerous to say "ever." As things stand now is a bit safer, so I'll stick to that. I just think (though I know it could change) that with the cost of fabs today, we probably aren't going to see a lot more of them going into China, especially with the IP issues. But as Bill McClean said, let's see what happens with the Samsung fab. That might dictate what happens in the future.
" ... doesn't seem like more than 10 percent of the world's chips will ever be built there. At least, as things stand now."
We used to say the same about Japan and (electronics) (cars) (etc) in the 1950's 1960's even in the 1970's. Some decades from now China WILL be manufacturing at least as much of just about everything (as % of WW production) as their population is % of WW population. Sure there will be variations around that ... but perhaps you didnt mean to say "EVER", just "as things stand now."
That's true, and I think that is what China's indigineous chip manufacturing foundries will continue to focus on for the most part. With China's huge internal market, that's a good business. But I don't see a lot more new fabs being built there, and it doesn't seem like more than 10 percent of the world's chips will ever be built there. At least, as things stand now.
I am not sure that's true, it's all about networks IMO. If you try to do it all, you will fail. SK's current strengths (in the form of Samsung mainly) is transient, sooner or later they will have to change their strategy.
Chip manufacturing prowess is becoming irrelevant. As the success of Samsung shows, vertical integration of everything in the supply chain from chip to product is the key. So SK is the only Vertical Integration Power in the world right now.
SMIC is trying to carve out market share in selected differentiated technologies such as embedded non-volatile memory (eNVM), CMOS image sensors (CIS) and Power Management IC (PMIC).
That's probably sensible. I don't think SMIC can compete head tio head with TSMC and GlobalFoundries. Instead, they are looking to serve the indigenous Chinese market with specialized products meeting the needs of OEM customers, and products which can be produced without requiring bleeding edge process technologies SNIC simply can't deploy.
I think they're making a bet on the future size and growth of the local market, but it's not clear they have another choice.
However, of the data I have seen, it has been increasing in cost exponentially and is now just as scary as the costs of new fabs.
Thanks. That was the impression I got reading the industry news, especially the exponential increases. It's nice to get confirmation from someone better placed to observe it than I.
Semiconductor manufacturing might just be the ultimate capital intensive industry, with the amount of capital required to remain competitive truly enormous.
The number of joint ventures and consortiums is no surprise, as hardly anyone can go it alone. It reminds me a bit of the early days of mainframes when you had IBM and the BUNCH (Burroughs, Univac, NCR, Control Data, and Honeywell) with IBM and Unisys the last men standing. The rest, including earlier competitors like GE and RCA dropped out because of the enormous R&D costs and difficulty of obtaining a commensurate return. (A senior RCA executive quoted in Computerworld decades ago piously proclaimed that RCA could have matched IBM in R&D, but chose not to. Of course, where is RCA now?)
I suspect we'll see further consolidation in the chip industry as costs rise further.
South Korea is a lot closer to China, with a strong incentive to cultivate economic ties. They see an emerging Chinese market they want to serve, they can best do that manufacturing in China, and the Chinese government will weclome the substantial investment as well as the technology transfer.
Thus far, the moves Samsung has been making have all been smart ones, which is how they've attained the position they have.
US politicians and regulators are still wearing "Yellow Peril" blinders, with notions of China dominated by the Communist government of Mao qand a perception of China as a threat. The Chinese government still calls itself Communist, but China's economy looks a lot like free market capitalism these days.
It reminds me a bit of Alice in Wonderland again, with Humpty Dumpty proclaiming that words mean what he wants them to mean. Some years back, a senior Chinese government official stated in an interview published in the Wall Street Journal, that if it worked, it was a triumph of the People's Revolution, and made pretty clear it didn't much matter what "it" was.
It wasn't all that long ago that that the notion of a thriving stock market in Beijing would have been unthinkable, but things have changed. The US is still catching up with the changes.
Do you know if anyone has tracked the increases in costs at progressively smaller geometries?
It's difficult to track this information because it depends on what you include. In addition with consortiums, it becomes even more challenging because there is a joint investment by the consortium and adaptations and optimizations by each of the partner companies. However, of the data I have seen, it has been increasing in cost exponentially and is now just as scary as the costs of new fabs. TSMC's CEO once said that process development costs were increasing so fast that the company would not be able to continue process development alone, and that was several generations ago.
Blog Make a Frequency Plan Tom Burke 17 comments When designing a printed circuit board, you should develop a frequency plan, something that can be easily overlooked. A frequency plan should be one of your first steps ...