Chip change as PC tide ebbs…The dire state of the semiconductor industry is not being helped by the sea-change that will happen when consumers get fast access to the Internet, and powerful PCs will not be necessary. David Manners reports The state of the semiconductor industry is so dire that something has to snap. “The industry is completely out of money this year,” says Jean-Philippe Dauvin, president of World Semiconductor Trade Statistics. “The coffers are dry – the reserves built up in the good times are exhausted,” agrees Malcolm Penn, chairman of Future Horizons. The market is expected to contract again this year to $134bn compared to 1995’s $150bn. The 1998 DRAM market is expected to decline to $14.5bn compared to 1995’s $40bn. Even the microprocessor market is expected to be flat this year. Watching the WebTV… With the PC market slowing and a perception that Internet-type applications may not require the steeply escalating DRAM content designed to speed up the use of larger and larger programmes, DRAM may never recover its former importance in the world semiconductor industry. But whatever snaps, it won’t be in Europe, reckons Dauvin. “The Asian Crisis has had a disastrous effect,” he says, “in the first half of 1998 there was a 27 per cent drop in demand for equipment. There’s a disastrous export trend from Japan to Asia and a collapse of the Japanese market.” Japan’s semiconductor market – which represented 40 per cent of the world market ten years ago – was running at only 20 per cent of the world market in Q1 98. America will also suffer. “My guess is that the Americans and Intel – running at 150mph – have just reached the exit. We don’t need their stuff for the Internet,” says Dauvin. Others think the same. “The consumer would rather see his modem double in speed than his processor,” says Bob Metcalf, founder of 3Com. When consumers get fast access to the Internet, they will no longer need a powerful PC because they’ll be able to use other peoples’ memory and processing power. This will drive a sea-change in the semiconductor market, reckons Dauvin. The PC tide that ran so strongly is already ebbing and the next tide will be digital consumer – mobile phones, automotive, DVD, set-top box, energy management and smartcard. As the equipment market transforms, massive adjustments will have to be made within the semiconductor industry. Last month the Japanese electronics majors announced the most horrendous year-end results with profits down by between 80 and 96 percent and with Mitsubishi Electric in loss. They blame the collapsed DRAM price caused by the overcapacity which followed the slowing PC market. “People built capacity for DRAM based on 40 per cent market growth a year,” says Dauvin, “with 20 per cent of that coming from growth in the PC market and the other 20 per cent coming from the increased requirement for memory within the PC.” With the PC market slowing and a perception that Internet-type applications may not require the steeply escalating DRAM content designed to speed up the use of larger and larger programmes, DRAM may never recover its former importance in the world semiconductor industry. From representing around 25 per cent of the industry in 1995, DRAM will drop to little over 10 per cent this year. “People want faster access, not more complicated desktop applications,” says 3Com’s Metcalf. Bad news for the DRAM industry. Hard on the heels of the awful Japanese financial results, the Americans have been showing pain. Motorola has announced 15,000 job losses with more to follow. Intel has announced 3,000 losses. National Semiconductor announced 1,400 job losses and a financial loss for the year of almost $100m. Leading PC companies Packard-Bell, Compaq and HP have given profits warnings. All of them blame the same thing – the softening PC market. “We are moving from a PC-driven market to a mass digital consumer market,” says Dauvin. That move does not favour Intel, the Americans, the Japanese, the Koreans or the Taiwanese. But it does favour the Europeans. Mobile communications, automotive electronics, smartcards, energy management, set-top boxes and DVDs – the full panoply of the digital consumer boom – are European strengths. “Intel is inside, but life outside the motherboard will be the major driving force for the next five years,” says Dauvin. So what’s going to happen to the semiconductor companies? “There will be an extremely severe consolidation process,” predicts Dauvin. He does not think this will happen during the slump, but when the market turns. “The first big consolidation will happen when the market turns up in 1999/2000,” says Dauvin, “because when the market turns up it moves quickly and that’s when you get investments. But, if you do not have the technology and the people, you start to lose market share. And if you lose market share when the market is moving upwards then you cannot continue. Remember the crash of 1985? The market bounced in 1986, but the consolidations didn’t start until 1987”. That was the year National took over Fairchild. The following year Harris took over General Electric’s semiconductor operations including RCA and Intersil. “Managements are already starting to cut major assets,” says Dauvin,”when you’re obliged to lose people you are on the floor because, when the business comes back, where are the people?” So the casualties of the consolidations of 2001/2 are going to be those companies without the technologies and the people in place to take advantage of the market upturn of 1999/2000. And some companies have already started shedding their people.