point
Menu
Magazines
Browse by year:
July - 2002 - issue > Below the Radar
Has the Semiconductor Industry Stopped Growing?
Monday, July 1, 2002
SINCE THE EARLY EIGHTIES, THE semiconductor industry has been considered to be the biggest driver of the technology economy. This causal relationship is driven by the increasing pervasiveness of semiconductor devices across all major end-markets. Two decades ago, the preponderance of semiconductor devices was targeted towards the PC industry. Today, semiconductors are ubiquitous in applications ranging from cell phones and routers to heart monitors, automobiles, bar code readers and even children's toys. Thus the fate of the semiconductor industry is becoming increasingly intertwined with the general economy.


Therefore, it is not surprising that Wall Street analysts use semiconductor stocks as a predictor of the market, and semiconductor stocks (Intel, Micron, Texas Instruments, Broadcom, Applied Micro, Nvidia, PMC-Sierra, Xilinx, RF Micro, Altera, etc.) are some of the most widely traded and held stocks in the broader market. Based on the last twelve-month trading volume, 15 of the top 100 Nasdaq stocks have been semiconductor stocks. Given the general belief that the economy has reached trough levels and IC consumption should rise dramatically when the economy rebounds, investors are already pricing the expectation of a big upturn in the IC industry in late 2002 and 2003. This is evidenced by the fact that IC stocks today trade at near-peak P/E multiples (see chart 1). However, the interesting point here is that many investors/sell-side analysts are perhaps ignoring a fundamental fact - the semiconductor industry has literally not grown in the past six years (see chart 2).


While it may be unfair to compare 1995 to 2001, since 1995 was a cyclical peak and 2001 was an abysmal trough, most investors today are assuming that post-2002 semiconductors will grow at historic growth rates. While we do agree that we are at the bottom of the semicycle, we remain skeptical that semiconductors will grow at the same rate in the next decade as they have for the past 25 years. In fact, there are several factors that would lead one to conclude that the prospects for the semiconductor industry will be dimmer in the next 10 years.


First, the biggest growth driver for the semi industry for the last twenty years has been the advent of the PC. However, the PC industry has matured in developed countries with high penetration levels in both the consumer and commercial markets (U.S households have over 55% PC penetration according to IDC). Further growth will be difficult without increasing penetration in the third world and lower socioeconomic groups in the OECD countries. There is clearly a demand elasticity issue in penetrating these untapped markets. Additionally, whereas Windows upgrades drove PC sales and consequently IC sales throughout the 1980s and 1990s, today there are very few "killer applications" that require users to purchase new computers. Simply put, the PC has become a discretionary purchase rather than a necessary purchase. While the PC is by no means dead, it is hard to advance a thesis that says that the PC industry will grow at historic rates over the long run. Given that half of all IC sales today still go towards the PC industry, mathematically it is difficult for the IC industry to grow at historic levels (see chart 3).


While similar arguments were made about the PC industry in the late 1990s, most semiconductor industry executives were less worried about the PC phenomenon because the communications industry was driving the growth in the IC sector.


The confluence of the mass adoption of the cell phone (66 million subscribers worldwide in 1995 to 656 million subscribers in 2000 according to Goldman Sachs Research, Global Mobile and IDC), the advent of the Internet (CSFB reports 14 million users worldwide in 1995 to 328M users in 2000) and deregulation throughout the world created a gold rush for telecommunications equipment and communications IC suppliers.


Backed by the currency of unprecedented capital markets, telecom service providers spent capital at a rate that was twice the historic levels with the philosophy of "if you build it, they will come" (see chart 4). When new applications (WAP, Unified Messaging, VoIP etc.) ultimately failed to take off at rates that were predicted and the capital markets collapsed, the result was a communications battlefield that left no one unscathed.


The upshot for the semiconductor manufacturers is perhaps equally poignant. From 1990 to 2000, communications grew from 15 percent of overall IC revenues to approximately 25 percent of revenues, according to Dataquest. Even in late 2000, most pundits expected communications to become the biggest end-market for semis by 2005. Today, as we sit in the midst of the communications maelstrom, it is hard to fathom a case where the communications industry will start growing at robust rates anytime in the near future. The fact that communications represents 26 percent and PCs represent 54 percent of IC revenues implies that over two-thirds of the end-market for IC consumption is growing anemically, particularly in the near-term.


Given the malaise in the communications and PC end-markets, it is hard to have an optimistic view of the IC market. However, we need to remind ourselves that it is always darkest before dawn. The semiconductor industry has witnessed downturns before (the industry did not grow from 1980 to 1982 and from 1989 to 1992) and many experts predicted maturation of the industry during the earlier downturns. Given the manic depressive nature of analysts, it is easy to predict doom and gloom when we are in a trough and eternal ebullience when we are at the top (who can forget semiconductor research analysts predicting a $300 billion IC market in 2000 during the 1995 upturn).


While the PC industry has unquestionably matured, the penetration of semiconductor devices in all electronics is still relatively small at 19 percent, according to IC Knowledge. The communications industry, notwithstanding near-term perturbations, should still be a growth industry in the long-term.
Half of the world has still not made a phone call, broadband penetration is in its infancy, and there are killer applications such as video-on-demand, unified messaging, etc. "still in the garage or labs" that could stimulate demand and lead to a whole new cycle of technology spending. In addition, the pervasiveness of semiconductors in the general economy (housing, medical, automotive and industrial) is in its infancy. Once consumers experience the power of smart cars and homes and realize the benefits of IC-driven medical applications, it will be hard to put the genie in the bottle.


Ultimately, the purview of the upper-middle class and wealthy (voice-activated car phones, navigation systems, automatic sensors, remote activation of household appliances, and numerous medical applications), becomes the necessity of the masses. While many of the aforementioned applications may not be "killer applications" on their own, taken together, these applications should drive industry growth.


In summary, there are two questions that we must answer to ascertain whether the semiconductor industry will grow again. First, is there ultimate end-market demand that will drive IC growth and second, are there any dislocating technologies that could create a platform shift that will obviate the need for semis?


On the first point, we would argue that while it is hard to see many new applications take off and lead to IC industry growth in the near future, some of these productivity enhancers will take off in due time.

In answering the second question, while much has been made of nanotechnology, optoelectronics and other revolutionary technologies that are "around the corner," there is no evidence that any of these technologies will become mainstream any time soon to create dislocations in the semiconductor industry.

Again, as a point of reference, the steel industry grew for more than a 100 years before it matured and the IC industry is merely 30 years old.

The real question to ask then is not whether the semiconductor industry will grow but rather will it grow as fast as in the years past? Given the tumult in the near-term communications industry and the relative maturation of the PC industry, one would conclude that the law of numbers would make it difficult for the semi industry to grow at the 15 percent rate that it was able to achieve from 1976 to 2000.

While we are not sure what the long-term growth rate will be for the IC industry, we would posit that it is likely to be closer to twice the GDP growth rate instead of three times the GDP growth rate as in the past 25 years.

While this still makes the IC industry a tremendous growth industry relative to other "old-world" industries that grow with the economy, the profound impact of valuations are discounting historic growth rates.

Dipanjan Deb is a partner at private equity firm Francisco Partners. Prior to co-founding Francisco Partners, Deb was a Principal with Texas Pacific Group. Prior to that he was the head of semiconductor banking for Robertson, Stephens & Company. Earlier in his career, Mr. Deb was employed by Netro Corporation, a wireless broadband communications company, and by McKinsey & Company.



Ashutosh Agrawal is currently an Associate at Francisco Partners. Prior to joining Francisco Partners, Mr. Agrawal was an Analyst in the Leveraged Finance group at Goldman Sachs & Co.



Twitter
Share on LinkedIn
facebook