How Little We Know

“A trend is a trend is a trend
But the question is: Will it bend?
Will it alter its course through some unforeseen force
And come to a premature end,”

so doodled Sir Alec Cairncross

Nowhere is this more true than in the semiconductor industry.  In  its mid-year forecast for 2000, IC Insights forecast a $250bn market for 2001. It turned out to be worth only $139bn representing a 32% y-o-y drop instead of the forecast 24% rise.

In 2000, IC Insights saw a trend of sunny uplands stretching years ahead – a $272bn market in 2002, $301bn in 2003 and $355bn in 2004 for an overall CAGR of 19% between 1998 and 2004.

The problem for the semiconductor industry in 2000 was, it was being said, keeping up with demand because of  past under-investment. Dataquest were saying the industry would need 25 new fabs in 2001 to keep up with the demand.

The delusion was universal.

In 2000, Hyundai-LG, the forerunner of Hynix and at that time the world’s biggest DRAM supplier, predicted a 100% increase in DRAM bit demand for 2001 but, whereas the 2000 DRAM market was worth $31.5bn, the 2001 market was worth $14bn

Beware of consensus.