Barron's: ARM foundries “struggling” to catch up to Intel’s lead in process technology
The other article is by the Report’s Tom Halfhill, who writes that contract chip manufacturers Taiwan Semiconductor Manufacturing (TSMC) and GlobalFoundries are “struggling” to catch up to Intel’s lead in process technology, as the company was first to market with three-dimensional transistors, which Intel calls “Tri-Gate,” and which the industry generall refers to as “FinFet.”
After reviewing plans laid out by Global and by TSM in February and April, respectively, Halfhill finds the milestones and plans offered leave both trailing Intel’s pace:
All the independent foundries are still scrambling to catch up with Intel. Their first FinFET processes won’t reach volume production until 2015, three years after Intel’s FinFETs made their revolutionary debut. Second, the foundries’ FinFETs will first appear in “14nm” or “16nm” processes that essentially add the three-dimensional transistors to previous-generation 20nm technology, achieving worthwhile gains in speed and power but relatively little improvement in logic density. And third, lithography costs are expected to jump dramatically after the 28nm node, potentially deterring many chip companies from moving beyond 28nm for quite a while.
Adds Halfhill, chip tools are running into their own obstacles that will hamper the move forward:
Furthermore, potential breakthrough technologies that the industry has eagerly anticipated for several years—such as extreme-ultraviolet (EUV) lithography and carbon nanotubes—are still far beyond the reach of today’s chip designers. Even if a new design’s release date is five years out, designers must settle for relatively conventional fabrication technology. Consequently, today’s 28nm processes will probably be longer-lived than previous generations. And chip designers will feel more pressure to extract performance from architectural innovations instead of relying on the heretofore steady accruals of Moore’s Law.
The switch to 450mm will be the largest, most expensive retooling the semiconductor industry has ever experienced. Will you be ready? 450mm fabs, which will give an unbeatable competitive advantage to the largest semiconductor manufacturers, are likely to cost $10 billion and come on-line in 2017, with production ramp in 2018. Unprecedented technical challenges still need to be overcome, but work is well underway at an R&D center in upstate New York, at the Global 450mm Consortium, G450C, and at the imec consortium in Europe. Hear from the G450C General Manager, Paul Farrar Jr., on the current status of activities, key milestones and schedules, and imec’s senior business development director, Lode Lauwers, on why 450mm is important for Europe, and the status of 450mm research on processes and devices.
Pete Singer Editor-in-Chief Solid State Technology