As many times as Intel has reorganized lately—four times since 1997 according to published accounts—I have to wonder whether it really matters how the company is structured. About the time the old order is out and the new order is in the new order becomes the old order and the process starts anew.
This years model is an Intel based around market opportunities rather than technologies, which is about as clear an admission as youre going to get that we have plenty of technology running around but not enough interesting uses for it.
This reorganization is supposed to better connect Intel with the marketplace. The goal is finding new uses for as many MIPS as possible. The perfect innovation, from Intels perspective, is something that causes customers to buy lots of new processors, either as upgrades or in completely new types of devices. But Intel is also extending its silicon offerings to include features not directly related to the CPU, such as wireless technology.
Touted as an example of Intel “done right” is the supposed “success” of the Centrino mobile processor set, a new version of which has just been introduced. I have not looked at the numbers, but its hard for me to get excited over Centrino, best-known for forcing 802.11b wireless on customers just as 802.11g became affordable.
In my view, if Centrino really had been successful, customers would have dumped their old notebooks just to get a new Centrino-based machine. I havent seen a lot of that happening.
If Centrino kept hardware OEMs buying Intel instead of AMD that alone would make it a success by some measures. But that doesnt fulfill the mission of convincing customers to buy new CPUs they wouldnt have bought anyway.
The reorg is a tacit admission that Intel is going to have to do more product R&D if chip demand is to increase. I am talking about real products, such as home entertainment or medical devices that could be built around specialized Intel chip sets. Already Intel has been shifting its focus away from the “speeds and feeds” of old toward feature sets enabled by additional Intel silicon.
The company has been doing this sort of work almost forever, but its latest output has been kind of depressing. Think Media Center PCs and iPod-ish video players as examples of Intel innovation. And Centrino, of course.
None have ignited a marketplace sensation. The next opportunity seems to be 64-bit processing, though its barely on most customers radar. The challenge of getting customers to shorten their PC and server replacement cycles has gone unmet.
This problem isnt Intels alone. Microsoft, as Ive commented before, faces the same issue: Theres no shortage of better ways to do things, but getting from here to there presents a tremendous challenge. Especially while maintaining significant backward compatibility and reducing risk to acceptable levels.
Many seem to have a vision of “computing tomorrow” but little idea of how to get there. Intels reorganization seems a short-term play that focuses attention on selling technology products rather than real innovation. Of course, weve had lots of innovation but little to show for it, so Intel seems to be doing the right thing—doing something with what weve got rather in concentrating on new stuff nobody has a use for.
But if you dont like Intels latest reorg, dont worry. Based on the companys history it the next reorg is never very far away.