Its the Chips, Stupid!

 
 
By eweek  |  Posted 2001-03-12
 
 
 

Its a little like the United States building a working space-based missile defense and then being successfully invaded by Mexico. Such is the condition of microprocessor titan Intel. For some years, the chip maker has pursued a vast and complex strategy of investments, acquisitions and forays into new markets, all the while churning out a steady cycle of upgrades to its bread-and-butter line of Pentium processors. But now, the grandness of its vision has become perhaps too complex, as the double whammy of receding PC demand and its own slowness to ship 64-bit processors has caused the company to miss its initial estimates for two consecutive quarters. Meanwhile, the once-derided AMD continues to make substantial inroads into territory where Intel should be the most invulnerable: the desktop. AMD has been shipping the fastest 32-bit X86 CPUs for some time, in greater volume than the much larger Intel can manage. Could it be the company that convinced the business world that paranoia is a virtue has run a little short? If its paranoia that works, then fine, but maybe a better trait would be focus.

Maybe Intels leadership thinks it doesnt matter how laggard the company is in shipping strategically important chips like Itanium and McKinley, as long as it buys enough companies and maintains an interesting mix of investments in its portfolio. We think it does matter, and, apparently, PC makers and IT managers think so, too. What other reason could there be for AMDs increase in PC market share—to the point of assuming technological leadership?

What the onetime paragon of technology companies needs to do is focus on one or two key technologies—foremost of which is the 64-bit Itanium. This must be brought to market, recall-free, ASAP. Defect-free 32-bit chips must also become routine. The 1.1GHz Pentium III recall is still a recent memory, and a Pentium 4 line that, despite its top 1.5GHz clock speed, proves to be barely equal to a 1.0GHz Pentium III in benchmark tests isnt a step forward for Intel customers.

The fact is, in the history of computing, 2001 is very late to be introducing a commercially viable 64-bit processor. And even though Intel now says Itanium will ship by midyear, most buyers consider the McKinley follow-on, due next year, to be the chip that is really worth running.

Intel executives at the companys recent Developers Forum said many of the right things, including repeated references to "focus." But the kind of focus thats needed to put the company back on track will mean, above all, the engineering excellence weve seen Intel demonstrate before coupled with much-lacking timely execution.

Rocket Fuel