Observers of the electronics and computer industries are reporting again and again that Apple has thrown over its microprocessor manufacturing arrangement with Samsung Electronics Co. Ltd. in favor of pure-play foundry supplier Taiwan Semiconductor Manufacturing Co. Ltd.
There are few reliable sources for this, a comment in the Korea Times here, an unnamed supply chain source there, but the general opinion is that this move has been coming for a while and that TSMC has completed the tape out of Apple's A7 processor in 20-nm CMOS in time for volume production in early 2014 – or even in 2013.
This is not sudden. Back in 2011 Apple was reported to be trying to get on to TSMC's 28-nm CMOS manufacturing process but that move apparently hit problems that persuaded Apple and TSMC to aim at the next process generation. One rumor I heard at the time was that the problem may have been related to Samsung peripheral IP used within the A5 and A6 processors that Samsung was not about to hand over to TSMC for the sake of a departing customer's convenience.
It is the case that TSMC with its gate-last bulk CMOS logic process seems to be pulling out a technology lead over Samsung which has a gate-first 28-nm CMOS logic process that has origins in the IBM-inspired Common Platform Alliance. And the logic of Apple moving makes even more sense when one considers Apple and Samsung spent a couple of years in the world's law courts arguing about patents pertaining to tablet computers and smartphones.
However, Apple and Samsung are linked not just by competition in smartphones and
tablets and by a supply arrangement in processors. They are linked by a
common interest in non-volatile memory. So if Apple does go sole-source with TSMC for A7 processors, what would be the consequences for its memory supply, if any?
At some point there has to be an advantage in having captive fabs, otherwise no one would ever build any new ones. Samsung will reap the advantages of being more vertically integrated. Apple can certainly buy enough NAND for its use, but will simply pay more. If they are smart they will have at least two vendors. This situation is where Tim Cook can earn his pay.
at some point there has to be an advantage to not being a dick, the way Apple has been. a company that genuinely has superior products does not destroy their whole supply chain and product pipeline as a defense against follower competition. playing such a defense game is tantamount to admitting your offense (innovation) is failing.
Do you really think it makes sense for Apple to essentially fund their largest competitor?
I do not see them destroying their product pipeline, though perhaps a slight delay. It was going to have to happen eventually, why not now?
Sure there is interdependence in many industries, but I have to imagine for Apple it was almost becoming uncomfortable.
They are also not destroying their whole supply chain, simply realigning it. That actually is good for the whole industry as it means not only can Apple build competitive advantage, but by moving supply out of Samsung which one could argue has far too high a concentration of Smartphone dollars on the supply side if Apple sticks with them, it grows supply chain competitors which helps other smart phone companies have access to top technology.
Sometimes it's shocking how intertwined the supply chain is. But this is far from the first time a situation like this has come up.
If I recall correctly, IBM abruptly stopped buying disks from one specific vendor (Shugart, maybe?) which had a devastating effect on the supplier. That's the position Samsung is in.
The reverse risk to Apple is just as real. Antitrust laws may prohibit restraint of trade, but there are a lot of ways of holding back without going clearly afoul of the regulations.
The real interesting part here is that both Apple and Samsung stand to lose big from any kind of a supply war.
Micron makes about one seventh of the NAND flash market by value.
I am sure they would be eager to supply Apple.
My point is that ALL five NAND flash makers have made efforts not to over-invest in production capacity over the last five years. As such it is possible that in 2013 NAND flash memory could be undersupplied.
Shortages of memory will drive prices up and could even introduce the dreaded A word....allocation.
As a UK citizen and resident I look across the water at the USA, and am always impressed by its ability to invent and innovate to bring useful and world changing products to the market. Apple epitomizes this aspect of the USA's can do attitude.
I am disappointed that the USA has allowed a situation to develop where a major player such as Apple is forced to hunt-the-marketplace to feed itself. When you're in the jungle you're also on the menu, no matter what size you are.
There is a clear advantage, at a cost, in vertical-integration. I think that the USA needs to address the question 'Why can't and we do it here, and why are we not doing it here?'. This is question for both the US Administration and its industry. After all the health of the USA is inextricably linked to the health of its industry. If the USA is proud of Apple and likes the taste, then plant an orchard and grow some more because the future is out-there.
It is very much like when someone says
'but education is so expensive'. My answer is 'If you think education is expensive try ignorance'.
There is a quote to which I often refer,'it is better to strike a candle than curse the dark'.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.