今天在EETimes上看到一篇该刊总编辑专访英飞凌 CEO Wolfgang Ziebart 的文章(英文原文附后)，觉得有点意思。
 
一年以来，Wolfgang Ziebart 俨然半导体行业的思想者，经常看到他指点江山。他在这里强调的两点：一是半导体行业整合势在必行，二是不会大者“更”大。要理解他说的话必须注意他说的整合是指垂直市场内的整合，"The consolidation will take place in market segments"，而不是象之前一些大公司到处狂买小公司搞大而全那样的横向整合。我的理解，之所以会这样，主要还是半导体领域研发成本的巨大压力使然。当每一家公司都拿出营业额的20~30%去做RD而且几乎都在做同样的事情、开发同样的产品的时候，合并就不可避免了。当然，它的一个前提就是摩尔定律已经实际上不再是事实上的定律，人们对先进节点工艺的追求不再象以前那样激进，至少不是所有产品领域都是那样迫切地需要32nm。现在甚至一些人在喊“技术过剩”，就是这个道理。如果用房地产市场的一句行话说，那就是“有泡沫”。即便没有泡沫，从另一角度来解读，半导体公司控制成本、追求更高的投资收益比也是解释得通的。
 
在这种情况下，对研发成本的投入开始变得更理性而精明，因此进行一些同类产品范围内的整并以避免研发和销售资源的重复投入。作为佐证，我们可以看到，在特别强调工艺或者说工艺总是走在最前面的市场领域，如FPGA、CPU、DSP和Flash，哪一个不是一两家寡头睥睨天下而甚少出现群起而攻的状况？
 
模拟领域是一个例外，因为这玩意一般都是定制的而且十分有利可图，门槛又低，只要几个有经验的人就可以玩了，所以暂时鱼龙混杂，十分的热闹。

一家之言，欢迎讨论。山外看山，难免肤浅。呵呵。

========
IC industry in midst of fundamental change, says Infineon CEO
 
Brian Fuller  
EE Times
(01/10/2007 11:18 H EST) 
 
SAN FRANCISCO — The semiconductor industry is experiencing consolidation changes that are fundamentally different from any that have gone before, which will result in fewer vendors that are not necessarily larger than they are today, Infineon CEO Wolfgang Ziebart said. "The industry is really undergoing fundamental change . . . and previous key factors of success [will] be valid only in certain areas of the market," Ziebart said in an interview with EE Times. A portion of the complete interview is posted at EET.com.
 
The disaggregation of the past 15 years created a more efficient ecosystem for electronics design and more companies. Those companies were founded in venture-backed startups that could avoid manufacturing costs by outsourcing to foundries. That drove a fundamental shift: Twenty years ago, some 70 percent of products were made on the two most recent process nodes, Ziebart argued. Today, because of the high cost of research and development, that percentage has dropped to a range of 40 to 50 percent, he added.
 
"For quite a majority of the companies, Moore's Law is not the driver anymore," he said. With pricing pressures and R&D traditionally budgeted at around 20 to even 30 percent of total sales, the industry is in for a major consolidation. "It will not take place in that companies will get bigger and bigger," he said. "The consolidation will take place in market segments. In those segments of the market, you will have fewer players but stronger players. Those players will not necessarily become bigger than they are." Ziebart's company has taken this to heart in the past three years by spinning out its memory operations as publicly traded Qimonda focusing on the communications and automotive sectors.