As the half-year mark came and went, it was clear that in 2001 the industry would see the largest year-over-year percentage decline ever recorded. IC Insights has calculated that in a worst-case scenario, the IC market will post a 28 percent decline for the year, and in a best-case scenario, it will suffer a 15 percent drop.
The most likely scenario-a 21 percent decline-makes the 2001 IC market drop the largest on record, surpassing 1985's 20 percent slide.
Still, within the stark figures are nuggets of promise. For example, assuming that the United States does not descend into an outright recession, IC Insights expects a 6 percent sequential increase in the IC market in the third quarter, followed by an 11 percent sequential jump for the fourth quarter of this year. If the market plays out in this way, it would bring a moderate 4 percent increase to the second-half IC market as compared with the first half of 2001.
We believe the IC industry encountered "the perfect storm" this year. In other words, the three main negative forces that affect the IC industry-weak worldwide growth in gross domestic product, an IC inventory burn and IC industry overcapacity-all hit at once in the first half. The record decline in IC unit and dollar volumes only served to reinforce that trend.
However, we believe the IC market has the potential to display strong sequential quarterly growth in the second half, even without a significant increase in electronic-system sales. The result will be what we call a "structural rebound."
The 36 percent IC market and 27 percent IC unit-volume increases the industry en-joyed in 2000 were excessive compared with long-term industry averages. Likewise, the huge quarterly IC market and unit-volume declines in the first half of 2001 must also be considered excessive.
One example of how a structural rebound can occur can be given using a company like Cisco Systems Inc. as an example. The rumor in the second quarter was that the giant networking-equipment OEM had so much IC inventory at its disposal that it was not ordering any components in its quarter ending April 28. This was probably an exaggeration, but nevertheless, the fact remains that Cisco's quarterly IC orders did not match the company's quarterly IC usage for its ongoing electronic-system sales.
Cisco has recently satisfied the difference between its IC orders (low) and IC usage (still relatively high) by using ICs from its inventory. When a significant amount of that excess IC inventory is depleted, even assuming a flat system sales rate, Cisco-and companies like it-will need to come back into the IC market. That will probably take place sometime late in the third quarter, when OEMs will begin to order more parts and trigger a structural rebound.
Cisco reported approximately $4.7 billion in electronic-system sales in its quarter ending April 28. After reading the newspaper headlines, you would have thought its sales went to zero. Assuming about 20 percent of Cisco's equipment sales value is derived from semiconductor content, the company probably used about $1 billion worth of semiconductors in its second quarter. If half or more of that semiconductor usage was coming from inventory (which we believe is a realistic assumption), the math shows that Cisco's semiconductor orders could jump by $500 million in one quarter after that inventory is depleted. The rise in orders could occur even if the company's system sales level does not pick up.
A few points need to be made about a potential structural rebound. First, even sequential growth in the IC market in the third and fourth quarters of 2001 would not be enough to prevent that market from registering a significant decline for the year. Second, the occurrence of a structural rebound does not mean that the next boom period has begun. Overall, annual semiconductor market growth rates of 20 percent or greater will most likely require electronic-system sales growth rates at or above the long-term average of 8 percent. IC Insights does not expect this situation to occur until 2003.
When we first developed our latest IC industry cycle model, IC Insights had no misconceptions that this would be "the last word" on cycles in the chip business. Still, the model can be a useful tool for understanding the industry.
Since 1970, the IC industry has encountered six cycles. IC Insights believes that the average growth rate for the industry has steadily declined from Cycle 1, in the early 1970s, through Cycle 5, in the mid- to late 1990s. This is a change from our previous cycle history chart, which kept a 17 percent average constant throughout each of the six cycles.
Overall, the adjustments are minor. We peg the IC industry's average growth rate during Cycle 1 (1971-77) at 17 to 18 percent, and during Cycle 5 (1993-98) at 16 to 17 percent. However, IC Insights believes that the average growth rate for the current cycle-Cycle 6, 1999-2001-has plummeted, to 11 percent or less.
What caused the dramatic drop? The bottom chart on page 67 lists some general observations about the semiconductor industry as well as some possible clues as to why the dip has been so steep. During the 26-year span from 1970 through 1995 the IC market experienced only two years of negative growth. However, in only a six-year period, from 1996 through 2001, the IC market will register three negative-growth years. This is not business as usual.
There's another frightening piece of data. From 1995 through 2001, the semiconductor industry will have spent about $250 billion in capital expenditures to achieve only about $50 billion in additional semiconductor sales. IC Insights believes that recent overspending by the IC producers has resulted in a 1995-2001 IC market average annual growth rate of only 2 percent.
Overall then, IC Insights believes the DRAM market "problem," and now the IC industry "problem," is still one of oversupply.
Bill McClean is president of IC Insights Inc. (Scottsdale, Ariz.), a market research firm.
Return to 2001 Midyear Forcast
RSS
RSS| 
Digital| 
Mobile
