There's an old adage in the soft-drink industry: you can spend your way to prosperity or save your way to oblivion.  It seems that leading companies in the semiconductor industry may be applying that adage as well.  In a high-risk, high-reward industry, the semiconductor companies that want to remain leaders have historically had to spend heavily on plant and equipment and research and development to innovate -- to invent and make new products that allow customers to do more at lower cost.

It seems counterintuitive, but it's especially important for semiconductor companies to consider high levels of spending for plant, equipment, and R&D at times like the present, when business is weak.  According to a Standard & Poor's research study, reported in BusinessWeek on June 24, 2002, the semiconductor companies that performed best after past business downturns resisted the normal business reflex to cut the budgets for capital spending and R&D; instead, they poured money into new factories and new generations of semiconductor technology, spending an average of 9% of sales on capital expenditures and 8% of sales on R&D.

Among the big-spending semiconductor leaders today are Intel, Texas Instruments, and IBM, as well as two fast-growing Taiwanese companies that are contract manufacturers, Taiwan Semiconductor Manufacturing and United Microelectronics.  They are leaders in no small part because they have been big enough and rich enough to invest while smaller and less affluent competitors were hunkering down and worrying about surviving.  In a virtuous circle, the big have gotten even bigger and, in our estimation, should continue getting bigger.  As an executive of a competitor that's strapped for cash observed ruefully, "We don't make as much in revenue as Intel spends every year on R&D."

This year Intel is spending $5.5 billion on plant and equipment and at least an equal amount on R&D.  It's been called classic Intel behavior, behavior that has enabled the company to widen its lead over more conservative rivals in the past.  In our view, here's the likely result of such behavior over the next 12 months: if demand turns up in the computing and communications markets, Intel will have six new plants with the latest manufacturing technology producing more powerful chips at a lower cost to help meet that demand.  Indeed, the benefits that Intel's chip innovations will offer to companies may help to create the demand in the first place.

The new plants should permit Intel to lower manufacturing costs by as much as 30% -- a huge competitive advantage.  In effect they will do the work of 2.3 older plants.  On a per-chip basis they are expected to reduce expenses to the levels of 10 years ago.

Intel, Texas Instruments, IBM, Taiwan Semiconductor Manufacturing, and United Microelectronics should reap further cost savings from producing chips on 300-millimeter (12-inch) silicon wafers, which offer 240% more area than the 200-millimeter (eight-inch) wafers that have been the industry standard since the early 1990s.  Most of the semiconductor industry's capital spending over the next five years will likely be earmarked for 300-millimeter manufacturing technology.

The chips made from 300-millimeter wafers will be smaller than ever, squeezing millions of transistors, each less than 0.001 the width of a human hair, onto a component smaller than a thumbnail.  By shrinking the size of the transistors from 0.18 microns to 0.13 microns, the companies will be able to reduce the size of chips by 50%.  (The next technological barrier for the industry to break: transistors of 0.09 microns.)  What's more, the chips will be twice as powerful as, and require less power to operate than, their predecessors.  The transistors on those chips will be able to switch on and off more than 1 trillion times per second.

In fact, the semiconductor industry figures to be a prime example of how the technology sector will generate gee-whiz innovations throughout this decade.  Such a statement may seem almost heretical now, when the tech sector is struggling and, in the view of some pessimists, will never bounce back.  But we think the tech sector ultimately will rebound for this reason: many companies that want to increase their productivity and profitability in the long run will only be able to do so by employing more and more technology.  For those companies, there's no significant alternative.  The current downturn in the tech sector represents a pause -- to be sure, a severe pause -- in the long-term march of technological advancement that promises to continue giving companies the tools needed to run better.

The course of innovation being as rapid and unpredictable as it is, we don't know what the leading semiconductor companies will look like, say, 10 years hence.  But we do suspect that the leaders then will be the ones who spend their way to prosperity -- who invest consistently and intensively in plant, equipment, and R&D in an effort to remain at the forefront of the industry.  That's because spending on plant, equipment, and R&D leads to new semiconductors that deliver higher performance at a lower cost, which in turn leads to further spending on plant, equipment, and R&D, which in turn leads to . . .

*  *  * 

The views expressed represent the opinions of Turner Investment Partners and are not intended as a forecast, a guarantee of future results, or investment recommendations.  The companies and industries mentioned should not be considered investment recommendations by Turner Investment Partners.  There can be no guarantee that Turner will select and hold any particular security for its client portfolios.

This commentary is current as of July 17, 2002. Markets and conditions change, and as a result our opinions on the subjects discussed here may also change. Turner's specific interests, if any, in the companies mentioned here are provided below. Like markets, conditions, and our analysts' outlooks, these holdings are subject to change over time. Turner disclaims any responsibility to inform visitors to this site of any changes to its outlooks and/or holdings.
 
Details on current holdings (most recent quarter-end) of companies mentioned in all recent Turner commentary are here.

As of June 30, 2002, Turner's stock investments owned 1,190 shares of Intel stock, 4 million shares of Texas Instruments stock, 2.7 million shares of Taiwan Semiconductor Manufacturing stock, and 7,290 shares of United Microelectronics stock.  Turner client accounts held no IBM shares.

Turner Investment Partners, founded in 1990 and based in Berwyn, Pennsylvania, is an investment-management firm that's employee-owned.  As of March 31, 2002, we managed about $9 billion in stock, bond, and balanced accounts and mutual funds for institutions and individuals.