July 2009

With reports of Q2 2009 foundry sales doubling compared to Q1 2009, foundries are increasing capital expenditures, beefing up design enablement tools and keeping their fingers crossed that improved sales will last for more than just two quarters.
The big news is that GLOBALFOUNDRIES is already making progress on their goal to win new sales from major semiconductor manufacturers with their announcement of STMicroelectronics as a strategic customer. Not only is STMicroelectronics a significant addition, their product focus with GLOBALFOUNDRIES will start with low power 40nm products. That’s a very nice win as the big three foundries are also courting new customers at that technology. But will GLOBALFOUNDRIES be able to keep up with all the design enablement offerings that are being developed by their competitors?
The Common Platform has taken another step toward a full service design to manufacturing service with the addition of the Synopsys design flow system Lynx into their ecosystem. The focus of this ARM-Synopsys-Common Platform partnership is on ARM’s high performance low power processor architecture combined with IBM’s high-k metal-gate 32/28nm low power process and the Synopsys design flow. Common Platform customers will have access to all this expertise with the potential benefit of getting 32nm/28nm products to market much earlier than originally planned.

At a time when most companies are cutting back on capital investment and capacity expansions, the challenge of developing a high technology industrial park appears to be an impossible task.  But Silicon Border continues to seize that challenge. 

Silicon Border is a 10,000 acre science park located in Mexicali, capital of Baja California.  The park hopes to attract semiconductor, flat panel display, photovoltaic and other technology-driven manufacturing tenants. 

Since breaking ground in 2005, Silicon Border has invested $20 million in preparing the infrastructure to support the needs of high technology manufacturing such as access to power generation (1600 MW), onsite water and waste treatment facility, advanced data and voice communication systems. 

In late-May 2009, solar-cell manufacturer Q-Cells, a manufacturer of solar cells,  announced plans to locate its North American manufacturing facility in the Silicon Border area with an investment of up to $3.5 billion.  This certainly makes sense.  Tenants of the Silicon Border Park enjoy free trade with 43 different countries, strong IP protection laws, a young, educated workforce and location incentives from the Mexican government. 

For selfish reasons, I am partial to locations in Arizona, however, this Park certainly provides an attractive option for companies that need easy transport access to U.S. markets while still enjoying low cost of operations. 

In July Semico published a blog, "GM should kill the Corvette.  That blog generated a heavy reader response, much of it defending the Corvette and its pushrod technology.  Having read the responses, I still think the Corvette uses highly-developed but archaic technology; but I also believe that the reasons why GM should kill the Corvette have to do with marketing, not technology.

One of the things contributing to GM’s troubles is that its divisions, which had been separate car companies, lost their brand identity.  The GM car you owned, Chevrolet, Pontiac, Oldsmobile, Buick or Cadillac, once made a statement about you that everyone from eight to eighty, male or female, understood.  You wore that identity like a suit of clothes.  That brand identity was destroyed.  

Chevrolet’s mainstream cars became nearly as big and heavy as Cadillacs:  same size engines, same feature sets.   Smaller Chevrolets were badge engineered to sell as Oldsmobiles, Pontiacs, Buicks or Cadillacs,  cheapening those brands.  Executives decided that it would be more economical to make one V-8 for several different divisions.  A Pontiac might actually have a Chevrolet engine; but, in the minds of the decision makers, the Chevrolet engine was just as good as a Pontiac engine would have been.  Who would care? 

The race to mass produce the next innovation in nonvolatile semiconductor memory appears to have a new contender.   Phase change memory (PCM) has been touted by industry leaders for years.  As far back as December of 2006 Wilhelm Beignvogl, SVP, technical innovation, Qimonda AG, stated, “phase change memories have the clear potential to play an important role in future memory systems”.  This statement came as researchers from IBM, Macronix, and Qimonda were prototyping the PCM memory device nearly four years ago.

Samsung Electronics Co., Ltd. and Numonyx BV recently announced they are jointly developing market specifications for PCM.  Could this be the key, finally, for wide scale commercial acceptance of PCM?  Most in the industry agree that developing package specifications and ensuring pin for pin compatibility will help speed up the adoption of PCM.  The two companies working together may also encourage future adoption as OEM’s are weary of contracting with only one manufacturer of new, unproven, technology.