To: Gottfried who wrote (5197 ) 1/29/2003 8:44:34 AM From: Proud_Infidel Read Replies (1) | Respond to of 25522 Is the semiconductor industry maturing? By Jim Feldhan Semiconductor Business News (01/28/03 09:24 p.m. EST) The following is a regular column provided to SBN by analysts at Semico Research Corp., a market research firm in Phoenix. Jim Feldhan is president of Semico. At the most recent SEMI (Semiconductor Equipment and Materials International) Industry Strategy Symposium, there was a debate amongst market analysts as to whether the semiconductor industry is in the mature phase of its life cycle. From where I'm sitting it certainly doesn't appear that this is an industry we can call mature. What does it mean to be mature? When an industry is in the mature stage, companies begin to see only small incremental operational and product improvements. Additionally, major strides in product upgrades are no longer occurring. Due to the reduction in innovation and lack of changes in manufacturing processes, companies exit the industry as dominant players take over the market. Cooperation among firms on price, product definitions and lack of product differentiation allow the big to get bigger. The small innovative companies have no new offerings for the market and are driven out of business. This may sound a little like the SRAM or DRAM business but certainly not characteristic of the semiconductor industry. Those who are proponents of the 'mature' theory have grabbed a lot of headlines lately but must not fully understand the semiconductor industry dynamics. I could go as far as to say that they lack insight, and foresight. If this industry is mature, do they believe Moore's law is dead? The Samuel Clemens (aka Mark Twain) phrase certainly applies here, “Reports of my death have been greatly exaggerated.” Have you seen the semiconductor technology roadmap? The state of wafer process technology and the roadmap for the next 5 years can safely say that improvements in manufacturing will allow Moore's law to continue on. In terms of barriers to entry, the cost of a new wafer fab has escalated to where new manufacturing plants are clearly a significant barrier. However, the success of the foundry model has effectively eliminated this barrier and continues to be an active catalyst for continued innovation. The recent move by the foundries over the last few years to develop and offer state-of-the-art technology will insure the continuation of innovative products from fabless companies. A few, select IDMs, such as IBM, see the advantage of offering foundry services, and are providing a new competitive dynamic to the industry. Recently there has been concern over the reduced capital expenditure budgets signaling a declining industry. The downturn over the last two years has caused manufacturers to focus on a return to profitability, and Semico believes this is a good thing for the long term. Additionally many of the large producers such as Intel, TSMC, UMC, TI, and Infineon, made substantial investments over the last three years laying the foundation for 130 and 90 nanometer capacity. It should be no surprise that capital expenditures will decline in 2003 as these facilities come down the learning curve, increasing wafer yields. This was forecasted in a Semico report “Wafer Demand: Rockin Ramp-up” dated April 2002. In the longer term, the foundation of a new capacity shortage is being laid. The industry over invested in 2000 and early 2001. But averaging the investments between 2000 and 2002, the industry is slightly below historic averages. The conservative attitude for 2003 is resulting in another year of under investment. Only companies such as IBM, Intel, TSMC, and TI, fall in line with historic norms. The biggest fallacy to the mature issue is the belief that applications are drying up. Yes, the PC and cell phones have been and continue to be a large driver for the semiconductor market. However, Semico believes that the emerging consumer market will be the source for the next set of “killer” applications. Products such as wireless networking, home automation, home health care, broadband to the home, security, GPS, home entertainment, telematics and many more are expected to drive the market. Additionally, I'm positive there is something out there that will certainly surprise us; an application that in 5 years, we'll look back and say how did I live without it! So what are the life cycle phases of an industry and if the semiconductor market is not mature, then what is it? The commonly accepted industry life cycle phases include fragmentation, shakeout, maturity and decline. Semico would define the industry in the classical shakeout phase. This is when efficiencies are realized and dominant players begin to take shape. Industry volumes begin to rise. During this phase there can be a number of false starts. Over the last couple hundred years there have been only a few inventions that changed the world infrastructure. The steam engine enabled the industrial revolution, refined petroleum enabled the internal combustion engine and electricity changed the business and home infrastructure. Semiconductors started the latest revolution, the information age. On the horizon are fuel cells that will enable and take the information age and technology platforms to the next level of productivity and innovation. Although some of us are personally showing signs of age, fortunately, the semiconductor industry was just issued its drivers license.