By Dana Blankenhorn

Nearly a decade ago I had the chance to talk with Chinese original equipment manufacturers about Intel (INTC).

Why aren't you buying its communications chips, I asked? Why are you buying chips from Broadcom (BRCM), for instance?

The answer came back clearly. Broadcom gives us designs and orders. Intel just gives us chips and an "ecosystem" of software suppliers. We can't afford to design products. We can't market them to consumers. We need complete designs from mass-market customers.

Since then, the situation has gotten much worse. Back in the previous decade, Intel was worrying about peripheral lines of business, like communications chips. Now it's losing the plot of the microprocessor market, its bread and butter.

A recent CNET story about Apple's process in creating the new iPhone is illustrative. Apple (AAPL) licensed a design from ARM Holdings (ARMH), then had its own design shop, formerly called PA Semi, push that design to its limits.

The result was a design called the A6, a proprietary product manufactured by Samsung, against which the company was competing directly both in the consumer market and in court. But we can't let business get in the way of business.

In the PC world, a product was designed around the chip. Intel dominated that market because it designed great chips and could manufacture them at low cost. In the device world, the chip is designed around the product, foundries can be rented to create a myriad of designs. Intel has no answer for this business model.

Instead, Intel has designed a new Atom processor called Clover Trail, reports Toms Hardware, which supports Windows 8 but not Linux. That's a step in the right direction, but it provides little differentiation for all the Windows OEMs out there, struggling to compete not only with Apple but with one another.

The PC era in which every machine was nearly identical is ending. The device era is one where you need to be different, and that starts at the level of the processor. Intel is the only company big enough to manufacture its own designs. What I call Moore's Second Law -- as designs get more complex, they cost exponentially more to get into production -- means Intel has won the mass market race.

But within each generation of chip-making technology there is infinite variety, and an infinite number of choices that can be made. Hardware, in short, has become software. Device makers demand the freedom to make those software choices for themselves. There's not enough room in any phone to design "around" a basic chip -- the differences have to be inside the chip or they won't fit.

Intel is betting that tablets running Windows 8 will gain a big piece of the device market, and that OEMs will compete fiercely for a growing business in a unitary space.

But that's not where things are heading. To compete with Apple, companies like Amazon (AMZN) and Samsung are taking more control over what their products do, and demanding more control over what their products are. Just as, on the retail end, that control presses software, content and after-market companies, so on the production side it presses companies that make, and think they should control, chips and what they do.

If Intel can't crack the problem I saw them facing a decade ago, it is in more trouble than anyone knows.

This article is commentary by an independent contributor, separate from TheStreet's regular news coverage. At the time of publication, Blankenhorn was long INTC but no longer knows why. He knows why he also owns AAPL.

More from TheStreet.com

• Netflix should raise prices again
• Apple, Amazon are 2 best companies
• Does HP deserve a do-over?