When the chips are down

Moore’s Law drove the PC market for years. But Rob Enderle argues that it no longer matters and the focus needs to shift away from cost reductions and back to revenue generation.


Last month, Intel confirmed Moore’s Law once again by extending to 45nm a manufacturing process that was thought to run out of steam at 65nm. Gordon Moore, the co-founder of Intel after whom the law is named, even provided a quote for the announcement.

Moore’s Law is the empirical observation made in 1965 that ‘the number of transistors on an integrated circuit for minimum component cost doubles every 24 months’. While it doesn’t mean that performance will increase by a factor of two, it certainly implies densities, power consumption and heat probably will.

In Silicon Valley, people care about Moore’s Law because there is a widely-held belief that it drives the industry. However, I am convinced that it hasn’t driven the industry since the 90s.

Right now, there are a number of problems with the PC market that need to be addressed. The first, as represented by Apple and Google, is that much of the real money is currently being made with embedded devices like the iPod and web-based, advertising-funded services like search engines. The second is that people aren’t as excited about getting new computers as they are about getting new TVs. The third is that most people have more performance than they seem to need — it is adequate network speed and content access they lack. Finally, for new product opportunities like Origami/UMPC and Media Center PC, the user experience is far from the appliance experience the consumer demands.

The combination of all these things is crippling the growth of the PC industry. One of the reasons that Intel’s and AMD’s margins are so slim is because neither has realised they should probably be building their market rather than working so hard to drive each other out of business.

It certainly doesn’t help that IBM and Apple, who co-founded the industry, both seem to think the PC isn’t important. IBM left the PC business two years ago, selling it to Lenovo; Apple, at its last PC conference, announced a phone, a wireless access point, and a set-top box, none of which had much to do with PCs.

Even Microsoft seems to agree. If you were at CES, you would have seen Microsoft on stage giving far more time to the Xbox than to its PC partners, something that clearly wasn’t missed by many of them.

This doesn’t mean that Intel’s discovery of a way to shrink the processor without substantially increasing costs wasn’t amazing. Correcting the design problem that prevented the move from 65nm to 45nm was a huge piece of work and gives it a competitive advantage.

In addition, we certainly can applaud Intel for the progress it has made with 45nm, beating IBM Microelectronics and AMD to the punch by nearly a year. The fact that it came from behind and now appears incredibly hard to beat is a far from trivial feather in its cap.

But this development won’t make any difference if the market moves to set-top boxes, game systems and smart phones: Intel isn’t in the majority of those segments, so a future market dominated by these devices probably has Intel gone from the physical world and into the history books.

The real advancement needed is for all the major players to put aside their differences and come together to develop a plan to return the PC market to its historic growth rate and profitability. The focus needs to shift away from cost reductions and back to revenue generation.

That would be an advancement that the employees, stockholders, partners and even customers (because they would be getting the product they have never known they always wanted) could stand behind.



Rob Enderle, founder of the Enderle Group, is a leading IT industry analyst