Intel Corp (INTC)

[Jerry R]

Craig Barrett interview with SF Chronicle

http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2004/09/26/BUGV88SI8T1.DTL

Q: What was the cause of that inability to deliver on that commitment? Is it the fact that it's such a huge company and you have 80,000 employees?

A: There are probably two ways to look at it. One way is that every time you do a postmortem on one of these projects, you find usually that there's one of three causes for the delay.

You didn't have the right specifications initially for the product. You didn't staff the product properly, either numbers or the numbers with right expertise in terms of employees. Or you weren't carefully managing the project. In 20 years I've looked at projects that have not met the schedule at Intel, it usually falls into one of those three categories.

The other contributing feature is: We build the most complicated things that human beings have ever built.

First of all, you can't see what you're building, and you're building a lot of them. You're building transistors you can't see, and the biggest transistor budget we have is a product that comes out next year called Montecito, from the Itanium processor family. It has 1.75 billion transistors in it.

Sometimes you try to use what worked last year on next year's product. And because the complexity of your products continues to increase, you have to increase the sophistication of the tools and techniques, processes and procedures. Sometimes we fall a little bit behind in that area.

And so we really ask our employees to go back and make sure that from a planning, from a staffing, from a management standpoint, we're putting all of the appropriate emphasis. But also, from the tools standpoint, you're not using last year's tools on next year's products.