Will Intel Figure Out the Device Market?

NEW YORK ( TheStreet) -- A decade ago I did an in-depth study of Intel ( INTC).

I talked to some Chinese OEMs, and asked how the company could win their business. Don't give us computer chips, they told me. Give us complete designs. Give us brands that will buy what we make and market it.

At the time, Broadcom ( BRCM) was just starting to eat Intel's lunch at these OEMs. That situation hasn't really changed, even as feature phones have become smartphones and smartphones have become tablets.

As I wrote at Seeking Alpha, Intel has a device problem.

I know the company will insist that's not the case. They may point to the 20 Windows 8-based tablets that are now in the pipeline, or to the new Ultrabooks, as they did to TheVerge.

It's true that Windows 8 is going to keep Intel's results strong in the near term. As James Rogers noted recently here at TheStreet, Intel still earns eye-popping margins on its semiconductors. It still dominates in the data center and in the "enterprise" market.

I still own the 600 shares of Intel that I acquired in the 1990s, after selling on splits several times so my basis in them is zero dollars. But I'm not adding to my position.

Because just as Moore's Law holds that better-and-better gets faster-and-faster, with capabilities rising and prices falling steadily, year-after-year, the rate at which the market for devices changes is also increasing.

Here is one example: Amazon.com ( AMZN) announced its Kindle Fire tablet, priced at $199, last October, but that product is now obsolete. The whole market is on watch for Amazon's next announcement, which has to include a better screen, more storage, a faster processor, a better browser and an even lower price.

Phones change even faster than tablets. Many companies refresh their product lines twice a year. Phones get announced, delivered and trashed faster than primetime comedies. The turnover is faster than for reality TV stars. (Whatever happened to Snooki, or Nokia ( Nokia) for that matter?)

To keep up with that pace of change, you have to get inside the product design process. You have to design your computer chip around the device, not have the device designed around the chip.

Because Intel owns its own fabrication plants -- it's one of the very few outfits that still does -- it has a tough time handling this pace of change. It takes time to ramp-up production and to increase yield so it can earn those fat margins.

The device market doesn't give you that time.

I think this is why companies like ARM Holdings ( ARMH), which merely licenses designs to other chip companies, who then order production, have become so powerful. You can customize an ARM design, creating several generations of hardware while tweaking the chipset in small ways. You can't do that with an Intel Atom processor. It is what it is.

You can see the problem clearly by looking at Samsung.

Samsung licenses chip designs, as GSMArena reports. It builds those designs in its own plants, as Ubergizmo notes. It assembles devices and brands those devices. It gains many of the advantages Intel claims for itself, but also takes more of the money from the finished product.

It has also adapted itself to the faster pace of the device business and now dominates the phone market alongside Apple, as Betanews reports. That's a market approaching 2 billion units a year.

Intel, they're eating your kimchee. Before I put more money into you, what are you doing about it?

At the time of publication, the author was long INTC.

This article is commentary by an independent contributor, separate from TheStreet's regular news coverage.