Share This

Thursday, 4 February 2010

Apple Takes On Intel

Apple Takes On Intel
Lee Gomes, 02.04.10, 06:00 AM EST

Why chipmakers might not be necessary anymore.

BURLINGAME, CALIF. -- Besides Apple's stock prices and Steve Jobs' reputation for visionary entrepreneurship, something else is riding on the success or failure of the new iPad: The future of the semiconductor industry.

The chip inside the new iPad is a microprocessor called the A4 that was designed in-house by Apple ( AAPL - news - people ), most likely using the expertise it acquired via its 2008 acquisition of PA Semi, a Silicon Valley start-up. Selection of the A4 was described as a blow to both Intel ( INTC - news - people ) and Qualcomm ( QCOM - news - people ), since products from those companies were spurned in the process.

It certainly was that, but it also suggested that semiconductor technology has matured to the point where for many applications, the Intels of the world might not be necessary anymore.

Everyone knows about Moore's Law, which describes the tendency for electronics to regularly double in capacity with no accompanying change in price. What is often forgotten is that Moore's Law isn't some force of nature like gravity, one that occurs independent of human intervention. To the contrary, it is an extraordinarily expensive process, requiring billions of dollars a year in R&D. Spend the money and your chips can keep packing in the extra circuits. Skimp, and they stop improving.

Intel has skillfully taken advantage of this dynamic over the last 20-odd years. It had a dominant position in Pentium-style processors, one that made it billions in profits, which it alone was able to invest to design and manufacture the next generation of even-better products. Advanced Micro Devices ( AMD - news - people ) proved to be a valiant rival to Intel, but it was an expensive fight for AMD. And the fact that Intel has a market cap nearly 20 times that of AMD suggests that investors are unsure if the smaller chipmaker has the staying power to keep at it.

One of the fundamental rules of technology is that things that start out hard and complicated, able to be tackled by only a few people, eventually get easier to do, allowing more people to handle them. Dell ( DELL - news - people ) grew enormously during the 1990s because it figured out the complex art and science involved in running an efficient PC manufacturing process. Once it cracked the code, though, others were able to do the same thing.

Reader Comments

Surely you're not suggesting that Apple has designed its own completely new CPU from scratch? This makes no sense at all! The time and resources required would be considerable. The fact that the i

Currently, only mammoth companies like Google ( GOOG - news - people ) and Microsoft ( MSFT - news - people ) know how to run big data centers. But that knowledge, too, will one day cease being a black art known only to a few and instead emerge as a straightforward bit of engineering that anyone can master.

It's clear that this "democratization" process is occurring right now in semiconductor design. The iPad is a relatively high-end device, yet Apple believed it didn't need to look outside its own walls for a CPU, and thus could forgo paying any form of "Intel tax." By contrast, the iPhones and iPods tend to use a chip called ARM that Apple, like many other ARM customers, need to license.

Of course, Apple is a very big company and, especially for the sort of high-volume product it hopes the iPad to be, it can afford the sorts of up-front engineering expenses that would make smaller companies reel. But if it can afford to make an in-house chip good enough for the iPad in 2010, might it not also be able to make one good enough for the Macintosh in 2013? And if it can do so by then, why couldn't Hewlett-Packard ( HPQ - news - people ) and Dell also?

To be sure, a walk through recent tech history involved stepping over the cadavers of any number of would-be Intel competitors. They might have been simply ahead of their time. Making chips seems to be getting easy enough that a company best known for marketing and branding is able to take it on.

To read more of Lee Gomes' stories, click here. Contact the writer at lgomes@forbes.com.

1 comment:

  1. Right, "engineering knowledge will one day cease being a black art known only to a few and instead emerge as a straightforward bit of engineering that anyone can master".

    ReplyDelete