The recession’s silver lining: a mandate for more efficiency
Jack Harding, eSilicon Corporation, Sunnyvale, CA USA
Semiconductor industry recessions have given launch to many sub industries over the past 40 years. In fact, if one annotates the birth of the many jettisoned segments on the semiconductor industry compound annual growth rate (CAGR) sinus wave that we have all endured, you can see that the package and test segment emerged in 1972; EDA formed in 1982; and foundry, IP, and design services developed in 1992, reaching escape velocity in 1996. The result has been several, multi-billion-dollar segments with honed value propositions, economies of scale and aggregation, and an increased pace of technological innovation.
The mechanism worked like this: demand sunk, fixed costs had to be reduced, and savvy CEOs instructed their operations folks to find a variable cost solution to deal with the weaker economy. The response during the 1972 recession was to outsource to package and test companies ??? and a segment was born. Eighteen months later, demand returned and the operations management said to the CEO, “I guess we should rehire all those packaging and test folks and rebuild the lines.” Of course the CEO said, “Not so fast. The specialized suppliers that we are using are cheaper and better than us, and they remain a variable cost. No more internal packaging, thank you very much.”
The emergence of a commercial EDA industry followed suit, and foundry was not far behind. There were common threads: cost, complexity, and a mandate to focus on one’s core value to the markets you service. Today, virtually no company can or should perform any of the functions available from this outsourced ecosystem ??? and they do not.
There is another outsourced segment coming into its own as this is written: back end operations. Back end operations include the activities from net list or GDS II to end-of-life. “Operations” is the only internal function remaining that has been allowed to exist with gross inefficiency, little to no scale, and technological shortfalls; but not any more.
The recession of 2008 has now caused CEOs to revisit the fixed cost issue and examine the sub-optimal operations contribution. The result is that they are calling upon value chain producers to provide a variable and lower cost solution with aggregated expertise and scalable learning. In short, internal operations are on their way out the same way EDA began to disappear as an internal function in 1982.
Of course, many larger fabless companies will respond with the equivalent of the now infamous, “Real men own fabs.” But we all know what happened to them ??? it just took a little longer. In the meantime, other than the very largest semiconductor firms, the majority of companies doing any kind of chip design will transition from expensive internal ops teams to an outsourced aggregated operations model.
The result of this recession-motivated change in the semiconductor industry will be efficiency, technological advancement, and a refocus of precious resources onto innovation and product development ??? where they belong.
No one likes a recession. But if a byproduct of this one is a purge of waste and a reordering of the ops function, well, it can’t be all bad.
Jack Harding, chairman, CEO, and president, eSilicon Corporation, 501 Macara Ave., Sunnyvale, CA 94085 USA; firstname.lastname@example.org.
Harding wrote on the subject of horizontal specialization during a down economy in The specialization imperative, co-authored by Mark Templeton. Read this archived article online at www.solid-state.com.