This year could see stronger growth than 2003, particularly in hot areas such as networking chips and consumer electronics. For example, several leading network-processor and switch-fabric vendors expect their revenue to roughly double in 2004. These forecasts are based on existing design wins ramping into production, not idle speculation. Disruptive Technologies Many new and disruptive technologies were just beginning to emerge when the downturn hit three years ago. In 2000, first-generation network processors (NPU) became available and showed promise despite missing some key features. Before NPU vendors could ship second-generation devices, however, the downturn hit, freezing system development at most networking equipment vendors. NPU revenue failed to grow during this period, despite improved products. But today, equipment vendors are restarting system designs. In the past, many of these systems were based largely on ASICs designed in-house. But these equipment vendors laid off many of their ASIC designers during the downturn and now must use third-party chips, such as NPUs and switch fabrics, to complete their designs. Thus, we expect these emerging networking technologies finally to gain ground in 2004. IP storage networking is another disruptive technology. This approach uses Ethernet and TCP/IP to replace the traditional Fibre Channel storage area network (SAN), reducing cost and allowing storage data to utilize existing LAN and WAN connections. Although initial IP storage products were available in early 2003, at that time information technology (IT) managers had no budget for any type of new technology. With IT spending expected to rise in 2004 for the first time in three years, we should see some pilot testing of IP storage in 2004, but given the conservative nature of most IT managers, volume deployment probably won't happen in 2005. At that point, however, the transition should happen quickly; by 2007, we expect IP storage devices to outsell Fibre Channel devices. For startups in these areas, these delays have been difficult, to say the least. Many of these startups were founded around the end of the boom and had business plans that assumed rapid adoption of new technologies. Now they are waiting and waiting for design wins to turn into revenue.
One problem is that most chip vendors did not adjust their spending in 2003 in response to rising revenue. In fact, many continued to layoff engineers, cancel projects, and reduce investment in an attempt to regain profitability. In response to customer requests, these companies often canceled projects to develop next-generation performance levels, concentrating instead on reducing cost at current performance levels. Unfortunately, customers are not good at projecting their own needs. During a downturn, they barely want to talk to you. Then one day, a new system development project is approved, and the customer suddenly needs your chip right away. But they don't want your current device, because their new system won't ship until 2005. They need a next-generation chip, and they may not want to wait 12-18 months for you to design one if they can get it from another chip vendor right away. Foundries and other manufacturers have a similar problem. During the downturn, many of these companies did not invest in new fab capacity; after all, the existing fabs were running at 50% utilization. But today, some fabs are already approaching full capacity, and the upturn has barely begun. Since it takes 12-24 months to add capacity, foundries need to get started soon. The semiconductor industry always goes through cycles. The gloomy downturn has already set the stage for a new period of shortages and rising prices. This period could start as early as mid-2004. The companies that will prosper during this period will be those that planned ahead during the downturn. Vendors with new technologies and next-generation performance will have a competitive advantage. Companies with added manufacturing capacity can take advantage of higher prices and still sell out their capacity. If your company is still focused on cutting costs, it is time to start preparing for the recovery. Originally published in Nikkei Electronics Asia, February 2004
© 2002-2004
The Linley Group |
