Leibson's Law

AMD’s founder Jerry Sanders once chauvinistically opined that “Real men have fabs.” That’s a really expensive proposition today in a world of diminishing returns for semiconductor companies. The cost of “real men” fabs is on the order of several billion dollars and rising in the nanometer era. At the same time, the semiconductor industry is now so large that it can no longer muster double-digit growth rates. However, even single-digit growth rates are attractive when annual industry revenues are several hundred billion dollars a year.

IBM, Chartered Semiconductor, and Samsung looked at the curves and concluded that they’d rather hang together than separately in the future world of nanometer silicon. The three started Common Platform to share development investment and spread the risk in semiconductor-manufacturing research. Add Freescale and STMicroelectronics to the growing roster of semiconductor companies in this collaboration and there’s a growing list of ecosystem partners as well.

Today, I attended the Common Platform Technology Forum. It’s being held at the Santa Clara Convention Center. I spoke with EDA industry analyst Gabe Moretti just before the opening keynote, who told me that he’d heard 1500 people had signed up for the event. It gave me a warm feeling. So many annual semiconductor-related events here in Silicon Valley are in decline. It was good to attend an event with strong attendance. And attendance was strong. By the time the keynote started, the auditorium was SRO with more than 1000 people listening.

I stayed for the introduction and keynote speech (because I have to leave for Savant’s International SOC Conference being held in Newport Beach tomorrow). Adalio Sanchez, GM of IBM’s Global Engineering Solutions, gave the keynote introduction. I liked the way he set things up, so I’m reproducing his comments here. Sanchez enumerated four forces affecting business decisions in the semiconductor industry:

1. Growth is not what it used to be. It’s harder to find areas of rapid growth.
2. The technology integration rate is exploding at an unprecedented rate.
3. The shrinkage rate in time to market is dramatically quickening.
4. More and faster innovation in semiconductor manufacturing is required while IC-manufacturing research and development costs are skyrocketing.

Sanchez said that these factors led IBM, Chartered, and Samsung to pool their R&D talents and funds.

Then Sanchez introduced Michel Mayer, Freescale’s Chairman and CEO. Freescale joined the Common Platform collaborators early in 2007. Freescale (formerly Motorola Semiconductor) has a very long—and up-and-down—collaborative history with IBM on the PowerPC.

Mayer started by echoing Sanchez. “Making money in this industry is an issue, now more than ever,” he said. He then referred to the current market as the “post-PC” world, which is sort of wishful thinking from a chip vendor that Mayer admits is no longer a major player in the still-large PC and server processor business. I just think it’s disingenuous to call this a “post-PC” world. Nevertheless, Freescale’s focus on embedded silicon makes for a very large available market. Think “light fixtures, cars, “friges,” and toasters” said Mayer. And he’s right. Those sorts of embedded applications will soak up a lot of chips.

Mayer forecasts three major trends that will bend the semiconductor business in the near future:

1. The Green Market. “Green is going mainline and will quickly become a business mandate,” said Mayer.
2. The Silver Lining. An aging population of boomers.
3. The Net Effect. Broadband, pervasive connectivity.

Mayer claims that going green is the industry’s most daunting challenge. Energy management is becoming a major factor in both consumer and business buying decisions both for economic reasons and for matters of social conscience. Mayer mentioned smart thermostats and processor-controlled fluorescent-light dimmers of two examples of green-semiconductor applications.

The aging boomer population (people born between 1946 and 1964) will spur demand for new types of goods and services such as in health care and safer transportation, predicts Mayer. For health care, Mayer suggested that an infusion of semiconductor intelligence into home-care products will allow aging boomers (who simply refuse to age) to deal with the vagaries of age more privately, more easily, and with more dignity. As a mid-bracket boomer myself, that sounds really good to me. At the same time, boomers refuse to acknowledge their waning driving skills. I know I don’t see as well as I did 20 years ago. Mayer predicts that the technologies being developed for autonomous vehicles will help improve vehicle safety, for both autonomous and assisted driving.

Mayer referred to the rapid adoption of social networking as the driver for increased connectivity demand. A population raised on MySpace and FaceBook takes constant connectivity for granted and drives the need for new lifestyle products that supply such connectivity.

These three trends push the pervasiveness of “semiconductor solutions” into people’s lives, says Mayer. He claims that the use of semiconductors across a broad front means that the market is no longer reliant on a single segment, which should smooth out the up-and-down cycles that have characterized the semiconductor market. I believe we’ve heard that story once or twice before. Semiconductor vendors believe this myth at their peril.