Global Manufacturing

The ECTC plenary Session was held the evening of June 1. Torsten Wipiejewski of ASTRI was the chair.

Rolf Plieninger of Infineon pointed out that his company was now #4 in the world semiconductor manufacturing after Intel, Samsung, and TI. He pointed out that of the three cycles that face manufacturers (industry-wide economy, technology, products) that an individual company could only influence "products" strongly and "technology" moderately. He listed the big drivers that occur at about 10 year intervals: 1985-PCs, 1995-communication, 2005-convergence, 2015-?bio/health/safety. In each driver the percent value due to semiconductor chip has dropped in time from 60% for the first PCs to 40% of today's cameras.
Despite the many negative trends (shorter product lifetimes, relentless cost pressure, 20% of cash flow going to capital) it is a great business as long as you are a global player and take advantage of local skill sets and can spread exchange rate risks. He saw design as becoming balanced throughout the world. He sees custom SIP as beating ASIC except in huge production runs.

Picture from left: Rolf, Ping, and Dongkai

-------Ping Wang of Freescale Semiconductor described some challenges of going global. They have moved part of their R&D to Asia. However, most fresh student graduates in Asia have had no exposure to processing equipment as well as having no "work place skills". Much training must take place in the first year with Westernized teachers dominating conversation to the detriment of building a company culture. Once trained, many new workers then try to find another job in one of the large cities. She remains optimistic.

-------Dongkai Shangguan of Flextronics showed charts indicating that IC production followed GDP growth trends although GDP grows at 2.5% compared to a 10% annual semiconductor growth. He predicted the trends of America's 38% ownership of manufacturing in 2002 dropping to 28% by 2020. Part of this is due to China growing over the same period from 7% to 34%. Although mass production will still be the method, personalization will become the overlay. Part of this will see the global market as being the sum of coordinated regional markets.
-------Historically the companies have seen that captive manufacturing leads to hierarchical costs. Thus, EMS providers out-source globally. However, the downside is that the supply chains are long and less able to anticipate any economic downturn with the chance of bigger losses to suppliers in a recession. EMS companies are taking on some design and product development (called ODM). In the future he sees design as a commodity. One of the big challenges for a global company will be to develop win-win relationships with Chinese OEMs. Another challenge is for a global EMS to compete with emerging regional EMSs. A problem across the board is lack of experienced electronics manufacturing engineers.
--------He recommends the following strategic directions: (1) Ubiquitous manufacturing to gain from every regional advantage, (2) expand in scale and global reach with a deep set of skills and capabilities, (3) have innovation, productivity, efficiency, and lean manufacturing, (4) provide your customers with green services to get more visibility and loyalty.

--------Michael Lebby of OIDA described the history and projection of the opto-electronics marketplace starting with LEDs in the 60s with lots of fundamental R&D at IBM and AT&T; the discrete communication lasers of the 70s; the improvement in quality in the 80s with Bellcore, Nortel, Siemens taking the R&D lead; the material breakthrough of the 90s (InGaAlP/InGaN); and this decade perhaps being measured by the manufacturability improvements in communication electro-optics and the huge growth of entertainment rather than communications sectors.
--------Despite the sustained 12% growth in market, industrial engineers are still 'recovering' from the communication bubble pop of 2000. The flat display panel, CD/DVD reader heads, and high brightness LEDs represent more than 80% of the market today. In contrast, the communication sector is still overbuilt with InP fabs still mostly empty. There should be consolidation, but each company thinks their IP mix is better than the other guys. In particular, there are 30 different ways to make a High Brightness LED and thus lots of IP and trade secrets. However, markets have a way of overcoming egos and Michael predicts reorganization and emergence of a strong InP industry by 2009.