Semiconductor Slowdown? Invest!

History shows that you can’t save your way out of a downturn. Just look at Intel vs. Sun.

popularity

By Kurt Shuler
Samsung’s announcement that it is investing a company-record $42B in 2012 for technology development came as a bit of a shock to many in the financial and technology press. The size of this investment dwarfs that of any other company I know of, and even exceeds the expected technology investments of entire nations, such as the combined investments of the Japanese semiconductor vendors.

Intel Invests In Server Processors During “Dot-Bomb” Recession
Samsung is honoring the time-tested advice of one of Intel’s founders, Gordon Moore, when he stated, “You can’t save your way out of a recession.” I experienced the wonders of this wisdom personally during the “dot bomb” recession of 2000 to 2002, when I was a product marketing manager working on the Xeon IA32 server processors.

intel_recession

Back in 2000, Intel had little market share in high performance servers. But we targeted this market with the first Pentium 4-based server processors code named “Foster.” It also was the first Intel processor with what is now called “Hyper-threading” and was our first “toe dip” into the world of multiprocessing.

In addition to hardware innovation, we spent millions of dollars on software innovation. In 2000, Intel helped found the Open Source Development Lab (OSDL) to enable enterprise Linux development. In addition, Intel spent tens of millions of dollars in co-marketing funds to encourage existing enterprise software developers to port their existing Sun SPARC software to x86 Linux and Microsoft Windows.

Results: Intel #1, Sun SPARC Is Dead
I couldn’t find any openly available market share data to show how successful Intel was versus Sun by investing in technology during that recession, but I have something even better that shows that Intel’s results were from better technology and not just fancy marketing: The chart above (reprinted from The Register) tracks the processor architectures used in the Top 500 high performance computing supercomputers by year. (Source data is here.) Notice how SPARC was number 2 in 2000, with IBM’s Power architecture servers in second. Now look how Intel in 2001 begins to build share, with a breakout in 2003 where it began to surpass SPARC. From data at www.top500.org, you can confirm that Intel’s growth in share was primarily from Xeon x86 and not the newer 64-bit Itanium.

Lesson Learned: Invest, Don’t Retrench
Every company does not have tens or hundreds of millions of dollars to invest like Intel did in the 2000 to 2002 time frame. But we all have the capability to research our markets, determine what key technologies might be important after a recession, and then focus our R&D spend on creating these new technologies.

What a technology company should not do is randomly cut costs, hoping to keep enough cash in the till to keep the company alive until the recession is over. That company may be alive, but it will be trounced in the marketplace by other companies that chose to invest in technology.

Sources:
Venturebeat, 17 January 2012.
Reuters, 17 January 2012.
The Register, 14 November 2011.
Top 500 Supercomputer Sites.

–Kurt Shuler is director of marketing at Arteris


Tags: