SAN JOSE, Calif. - Despite a sudden and disturbing lull in the IC market, GlobalFoundries Inc. is moving full speed ahead with its aggressive silicon foundry strategy.
Will they succeed or fail? Here's what various analysts think:
Jim Feldhan, president of Semico Research Corp., said: "They have a very aggressive roadmap, but now have to execute" to overtake some of the competition.
Doug Freedman, an analyst with Gleacher & Co., said: ''GlobalFoundries is not a pretender. I have always said GlobalFoundries' business model has the potenial of being successful.''
Jim Walker, an analyst with Gartner Inc., said: ''GlobalFoundries (will give) TSMC a run for the money.''
Steven Pelayo, an analyst with HSBC, said: ''We continue to believe the real battle for foundry market share will begin in 2H 2011 as leading-edge, 28-nm production begins in earnest.''
G. Dan Hutcheson, president of VLSI Research Inc., said: ''You could hear a tear ripping through the fabric of the foundry industry at GlobalFoundries' Global Technology Conference. The ripping sound was the applause given to Lip-bu Tan’s comment that 'Clearly the industry needs a strong alternative.' Coming from the lips of Cadence’s CEO was a very strong message that you seldom hear from someone at this level. More powerful was the mid-speech applause it received. Techies applaud before and after, but seldom interrupt a speech with them.
It was a clear message of user frustration with the current virtual monopoly structure. But customers always hate virtual monopolies and if the incumbent continues to deliver value, they are seldom unseated. That said, the history of technology riddled with the bodies of virtual monopolists who got too arrogant, allowed their value to diminish, and were unseated by an up-start. This invariably happens when a market is at a strategic inflection point. So is this the case for foundries?
The strategic inflection point is not high-k/metal-gate nor yields. Those are some of the features of a strategic inflection point I predicted years ago: that is, the transition from micro-chips to nano-chips. Back then, I wrote that it would be ever more interesting and boy is that the case now.
The classical foundries came largely unprepared to the sub-100 nanometer era. They had not needed R&D because equipment suppliers provided working unit processes, while integration and scaling was fairly simple. A foundry spending 3 percent on R&D is equivalent to an IDM spending just over 1 percent. That was a difference that made foundries extremely profitable. That period has ended.
The fact that it has ended is exemplified by TSMC’s 40-nm defect density slope contrasted against the cliffs seen at GlobalFoundries and not surprisingly, Intel. Another feature of the nano-chip era is the loss of design portability across foundries, as Jean-Marc Chery of STMicroelectronics put it. This is extremely true at 32/38-nm because of high-k/metal-gate. That means that the customer’s ability to lever pricing goes away. In other words, it’s a winner-take-all game for TSMC. All they have to do is execute, while the others fail. So far, that has not been happening, which is why there’s a ripping tear.
This means that there is another strategic feature of this battleground. It’s not so much as a battle between the TSMC and GF as it is one between TSMC and the Common Platform. It is a fundamental principle that the easiest and most profitable way to take a market from a leader is by doing something they are unwilling to do. In this case it’s fab sync. Not only can GlobalFoundries offer a virtual second source between fabs in in Germany, Singapore, and in the future New York, fab sync gives the customer the ability to run the same wafers at IBM and Samsung. That’s important because the customer is not nearly as interested in a second source as they are in maintaining the balance of pricing power.
Lack of design portability means that TSMC will have a pricing power slope that is far more in their favor than ever before (as my Dad used to say, 'everyone in business wants a level playing field … just one that slopes down and away from them'). While customers will hate that it is highly unlikely that TSMC would ever allow fab sync of its fabs to others. TSMC has earned more than 100 percent of the profits ever earned by the foundry industry (meaning the rest have lost money). There is no way that they will give up that. Therein lies GlobalFoundries strategic advantage.''