The 3 Reasons Semiconductor Experience Revenue Cycles – Forbes
The semiconductor business is notorious for its “Boom-Bust” cycles. For a couple of years the industry will be highly profitable, then for the next couple of years profits will elude the entire industry.
Why is that?
When you boil it down, there are only three events that ever cause semiconductor cycles:
- Capital spending ups and downs
- Process migration glitches
- Demand slowdowns
Let’s briefly look at each of these.
Capital spending is the big one and accounts for more than half of all cycles, or for more cycles than the other two combined. In a nutshell it’s a predictable sequence: When semiconductor makers are profitable they invest in new manufacturing capacity. This new capacity ramps in about two years to full output, causing an oversupply which unravels pricing. When profits vanish manufacturers shut off capital spending, which later on causes a shortage.
This has been the subject of a prior post, so I won’t go into more detail here. The dynamics of the cycle are presented in this post’s graphic.
Next Next on the list is process migration glitches. These are rarer events, and tie into how semiconductor manufacturers plan their capacity. These companies determine how much product their customers are likely to purchase and decide how they are going to produce it. Since demand grows nearly constantly (more on this later) these companies plan to satisfy a portion of this growth by adding production capacity (processing more wafers) and to satisfy the other portion by improving efficiency through process shrinks. A process shrink allows you to squeeze more transistors onto a wafer by printing finer-resolution lines.
A company that stumbles and fails to follow its process migration roadmap will produce fewer transistors than planned and will not be able to keep up with ramping demand. When the industry as a whole has this problem it causes a sizable shortage.
The last two industry-wide process stalls were in 1994-5, and in 2005. Both extended existing CapEx-driven shortages, and both of these shortages were followed by unusually spectacular collapses.
There are indications that we are facing another process-driven shortage in coming years. I’ll deal with that in future posts.
Finally we have demand slowdowns. I have mentioned in an earlier post that these are exceedingly rare. On average they occur every 15 years. I have not found a way of predicting these, and would probably be very well off if I could. The last two were the Internet Bubble Burst of 2001 and the Global Financial Meltdown of 2008, which occurred in surprisingly rapid succession.
A demand slowdown creates similar results to an overcapacity – chip makers cut prices to try and stimulate demand but instead cause a downward price spiral.
So there you have the three events that trigger semiconductor cycles: CapEx cycles, process stalls, and demand slowdowns. Only the first of these is easy to predict – the other two are trickier.
Objective Analysis helps our clients to understand the semiconductor cycle and to anticipate triggering events, identifying when they occur. If you would like to learn how we can help you please visit our website.