IDF: 14nm Production on Track for Late 2013; Paths to 10nm and Below
Here's my question: Will they have enough customers to have those 5 factories rocking? Let's say they don't ever put chips in Apple phone or Tablets..........how are those 5 factories going to make full use of the Technology advance?
Wally, I just don't understand how ARM will "collapse" as you say when they have a dominant architecture. It's like saying Burger King will wipe McDonalds out because they can crank out more burgers, even though the quality isn't as good. My thinking is that both Intel and ARM will succeed, just as both Burger King and McDonalds succeed. Your stance seems to be that Intel will destroy everything in its path. Isn't that very misleading?
"Can anyone answer my question? How in the world are 5 factories going to be pumping out chips ? Where's all the demand gonna come from?"
The demand for the high end already exists. It's why all of TSMC customers have been so angry. And why some ARM customers are begging Intel to do their fabrication. But design takes time. It's not an instant thing. If Intel does have excess capacity, which we haven't heard anyone say yet, they will have the ability to do foundry work. I don't see it. Intel is re-tooling for mobility and is addressing a lot of new markets. And there are a lot of embedded markets as well. Intel doesn't make PCs - Intel makes processors and they are going into all aspects of society. Including medical, automotive, signage, and industrial, in addition to PCs, servers, smartphones, tablets and other mobility applications.
Intel is also a master at adjusting costs when confronted with lower demand. If that happens they will just use the lull to move faster to 14nm where demand will be better and better.
Intel will have huge 22nm capacity and then huge 14nm capacity. Why would anyone want 28nm ARM processors, given a choice? As it becomes more and more obvious that Intel will stay on the cutting edge as ARMs fabrication fades, there will be much more of an incentive to switch. We are in the midst of that change where it is becoming obvious. 14nm will be available in a year. That means a whole lot of discussion and planning is taking place right now for those who want to be on the cutting edge.
Intel has made the decision to own mobility. And they have the fabrication, manufacturing prowess and capacity to make it happen. All of their focus for the last 18 months has been on changes necessary to address the mobility market. These efforts are coming to fruition and will substantially arrive in 2013.
So, where is the demand going to come from? It will come from those who realize that ARMs fabrication is crashing just as Intel's is moving steadily ahead with its every eight quarter pace without pause - to 14, then to 10, then to 7 and then to 5nm. ARM is going to stall out after 20nm for at least three years. Don't think any major manufacturer is going to wait three years for them to catch up. And how will they catch up with shrinking revenues and market share?
The demand will come primarily from mobility. Intel started late. But Intel's newly defined assault on mobility has been well planned and well executed. You have people like Mike Bell backed up by Intel's R&D, superior fabrication, world class manufacturing.
It is possible that Intel will be the only major IDM (integrated device manufacturer) left in the world. As processors become more complex this is more and more important. The ARM fabless model is essentially broken because it doesn't have the tight integration of design and manufacturing that is required to have a successful experience with today's state-of-the-art processors and fabrication.
Let us not forget the economics. Each new processor cycle brings improved economics. This is why Intel keeps telling us it is necessary for them to continue the every eight quarter pace. Those that can't keep up (the entire ARM world) will have lessor economics, worse margins and problems with cash flow. You don't have to be smarter than a fifth grader to see where this is going.
Intel will provide the most bang for the buck. The best performance, the least power consumption and the best price. If you build it, they will come. Intel is building it. And there is nothing the competition can do about it...
The PC market is a reliable and steady market that Intel has tight control over. All the customers there have been customer for several generations of product. While there is some slowdown there, it's a huge market which will account for much of Intel's chip supply. The tablet and smartphone market on the other hand is a new market where Intel is a new entrant. Their process lead does give them an edge over the competition, but they are clearly behind on architecture. Besides, the customers decision to go with Intel chips over competing alternatives is not based only on the technology competitiveness but things such as a cost, an area where ARM based solution have a distinct advantage. On the tablet/mobile side therefore, while Intel may get some volumes, they would be at much lower ASP than on the PC side. Based on past performance (Medfield) things are not looking good for Intel on this side.