Apr 30 2014, 9:18pm CDT | by Forbes
I want to be the first to say it: The Semiconductor Cycle is dead!
This is something that is brought up every time the semiconductor cycle has an extended upturn. It’s also said every time the industry is at the end of a downturn. Just when the next change is about to happen, then more and more industry leaders start to pronounce the end of semiconductor cycles until everyone starts to believe it.
I figure, who better to start that ball rolling than a semiconductor industry analyst.
But do I believe it? Not for a second!
So why bring it up now? Because we are at the onset of a period that has all the earmarks of an extended upturn, and that will bring people out of the woodwork prognosticating the end of the semiconductor cycle. Rather than just joining these folks, I figured I would be both the first to proclaim the end of the semiconductor cycle as well as the first to debunk that line of thought.
This post’s graphic shows how prices for both DRAM and NAND flash have been relatively stable for over 18 months. With this stability comes profitability for chip makers and a general sense of well-being in the semiconductor industry.
The semiconductor industry largely expands through organic growth. Chip makers typically re-invest their profits more aggressively during this phase of the semiconductor cycle. A couple of years later this aggressive investment blossoms into an oversupply, and a price collapse ensues. You can almost set your watch to the rhythm: Two years up, two years down.
Since a sizable portion of the market consists of undifferentiated commodities, and since production is capital intensive, producers are highly motivated to run their factories at full bore at all times. When an overcapacity strikes, they take advantage of the lack of differentiation and cut prices to take away enough market share to find homes for their factories’ output.
The competitor who lost this market share responds by cutting prices even lower, and a race to the bottom ensues, causing a collapse.
Investment analysts often ask me if it will be different this time, if the recent consolidation in DRAM (and certain smaller sectors) won’t cause suppliers to become “Rational” and invest more moderately. Some point to other industries that have undergone similar changes: US railroads and automotive aftermarket products.
Although I have heard these arguments a few times, I have been unable to locate studies that show these markets to have changed. Perhaps a reader can point me in their direction. For now I must be skeptical that this even occurred. It has the hallmarks of a myth.
But say that this did happen in these other industries, and say that somehow, without collusion, semiconductor makers could pace their factory upgrades and expansion to perfectly match the market’s growth. There would ensue an extended period of profitability. Life would be good.
Life would be so good, in fact, that outside investors would place bets on a new entrant to the market. That would throw a wrench into the works, creating an oversupply, and the market would undergo another collapse. This is not conjecture – it’s what happened when Japan entered semiconductors in the late 1970s, Korea in the late 1980s, and Taiwan in the middle 1990s.
In other words, it’s certain that there will be a collapse. The only question is when it will hit.
Although a lot of very good analysts dedicate a significant effort towards predicting chip consumption, and produce very reasonable demand forecasts, my company has had a lot of success tying the timing of semiconductor cycles to prior years’ capital investments. This usually produces very good results, but things look a little different this time.
Memory processes have reached a threshold which promises to be tough to cross. In particular NAND flash looks like it will cause a lot of grief to process engineers and fab managers. Two transitions are taking place today, both of which require the use of new materials and processes that have never before been successfully brought into production. Without going into a lot of detail, allow me to say that Objective Analysis is warning our clients that budding shortages could extend into 2016 or even 2017.
In summary, though, the industry hasn’t changed. It’s a capital-intensive undifferentiated market, with a lag time between investments and production volume, and this is the recipe for industry cycles.
Investors should plan for continuing semiconductor growth but keep their eyes wide open for signs of an oversupply. It may not occur until 2017, but it will happen, and when it does it is likely to surprise all those folks who confidently stated that the semiconductor cycle had died.
Forbes is among the most trusted resources for the world's business and investment leaders, providing them the uncompromising commentary, concise analysis, relevant tools and real-time reporting they need to succeed at work, profit from investing and have fun with the rewards of winning.
blog comments powered by Disqus