2004 Has Been Great, But What Comes Next?


Well, it's holiday season already. From a personal view, many are looking forward to a few days off, relaxation, great food, and a wonderful time with friends and family. From a business perspective, however, it may not be as cheerful. While the numbers indicate that 2004 has been a very good year for the semiconductor industry, what will 2005 bring?

The year 2004 will be among one of the best years for the capital equipment markets. Having moved to the upside, latest estimates show the semiconductor equipment market will expand 66% this year. In more detail, the wafer fab equipment market is estimated to rise 72%, while the packaging and assembly equipment market should grow 49%. The automated test equipment market will rise by 52% in 2004.

The semiconductor equipment markets have shined bright as 2004 comes to a close. But worries about year-end semiconductor demand and 2005 prospects decrease the luster. Concerns about oil prices, inflation, interest rate hikes, elections, Iraq, and terrorism seem to have slowed final product demand. A renewed uncertainty in the macro-economic environment is depressing consumer confidence at the corporate and individual level. Consequently, end-user spending for semiconductors has become selective. Demand growth is slowing. Excess inventories have emerged. Meanwhile, capital equipment spending has started to slacken. Back-end equipment orders have slowed. In the front end, there is word of push-outs and delays before year-end. Even though this uncertainty period tarnishes the prospects for strong growth in the ending months of this year, it still cannot derail the entire year's strong recovery.

For 2005, it is a different story. Gartner Dataquest's indicators for the semiconductor market reflect this change. 2005 may be the next transition year for the industry. The outcome will be determined by several factors: the temporal expanse of the current macroeconomic uncertainty period, strength of end-user demand in 2005, and the vigilance of the industry to deal with and control excess buildup in inventories and capacity. Regardless of how these factors play out, 2005 will likely be a slow year for capital spending — harboring either a spending pause or the start of a renewed industry downcycle. The bigger impact and divergence is for 2006 prospects.

Several possibilities abound. Being conservative and cautious, a possible downcycle could occur, entailing issues on both the supply and demand sides. We do not see a purely supply-driven downcycle, but one with a demand component to it. Still, there is considerable probability for a pause and second spending leg in the current upcycle. One could also take an optimistic view and believe that the industry (having learned from the disaster in 2000-2001) has been acting cautious and cognizant of its risks and business fundamentals. Over-exuberance has not been part of this cycle, despite strong capital spending growth in 2004.

This belief is underlined by the current atypical inventory burn. In prior cycles, the industry eliminated excess inventories only after overcapacity emerged (outside typical seasonal patterns). This is not the case this time. Capacity additions remained controlled and matched to the current demand ramp in the current cycle. Supply has remained relatively tight and, despite a tight supply-demand situation, the industry moved to liquidate excess inventories as they emerged. The industry is acting from a position of strength, closely balanced supply and demand, to tackle excess inventories.

The packaging/test cycle of investment is farther along than for wafer fab. Following last year's strong gains, we estimate growth at approximately 50% for the combined back-end equipment segments (packaging/assembly and ATE) in 2004. Back-end equipment orders began to exhibit slower growth in the last half of 2004. In prior cycles, an overshoot in the back end and consequent bookings declines were a signal for the next industry downcycle. But a back-end investment overshoot does not necessarily force an industry downcycle; it is merely an indicator.

While the global economic recovery remains on a solid path, it has become less lofty and more uncertain. Both elements cast a shadow over the strength of device demand for 2005. Demand growth will be weaker than in 2004, and not as strong as hoped for. Gartner Dataquest's indicator models show an increasing number of downward indicating factors for next year. However, continued growth in device unit shipments, moderate investments during the upcycle, and early reaction by device makers could combine to make it a shallow downcycle. In fact, should demand in 2005 surprise to the upside, the start of the downcycle may be delayed until 2006.

Click here to enlarge image

JIM WALKER, VP research, Semiconductor Packaging and Assembly, may be contacted at Gartner Dataquest, 251 River Oaks Parkway, San Jose, CA 95134; (408) 468-8483; e-mail: jim.walker@