Craig Barrett is spending the last days of his tenure as Intel chairman the same way he spent his previous 35 years at the corporation: moving at a superhuman pace that leaves exhausted subordinates in his wake.

Mr. Barrett has maintained this lifestyle since he replaced Andrew Grove as CEO of Intel in 1998. "Was it hard to follow a legend?" he asks himself in his typical blunt way, adding, "What do you think?" Mr. Barrett barely broke pace when he became chairman in 2005, and shows no sign of slowing even now, at age 69, as he faces retirement.

Mr. Barrett is legendary for a working week that begins at the Intel facility near Phoenix, where he has lived for a quarter-century ("I haven't been home for five days in a row in 20 years," he says) and moves Tuesday mornings to Santa Clara, Calif., and Intel headquarters -- with stops around the world (30 nations per year on average) in between.

Mr. Barrett began his career at Intel during a market crash, led the company during the dot-com crash, and now finds himself retiring during yet another market crash.

He waves his hand in a sine wave through the air, "That's how it is in this business. The endless boom-bust cycle -- I've been through eight of them. You just deal with what you're given . . . and develop a thick skin."

Still, if no catchphrase or law is likely to attach to Craig Barrett's name, his tenure at Intel, leaving the semiconductor industry's leading company even more dominant that when he arrived, offers a collection of important lessons. Call them Barrett's Rules.

- The business is bigger than the business. Mr. Barrett has long recognized that, with a company as large and influential as Intel, governments and cultural forces loom as large as any competitor. That's why, when Intel was being battered by Japanese competitors in the early 1980s, he took his fellow managers to Japan to visit everything from manufacturing plants to corner markets to show them how to build in quality.

- Don't mess with Moore's Law. Inevitably, Mr. Barrett says, every few years "some company will say, 'What's with the rush to improve our technology every two years? Let's slow down to say, four years, and only have to invest half as much capital.' It always sounds like a cool idea, and it always ends up with that company losing market share."

- Invest during hard times.The two most controversial decisions Mr. Barrett made as Intel CEO were: 1) to take the company, largely by acquisition, into the communications business; and 2) to maintain the company's traditional level of capital investment right through the darkest days of the dot-com bust. Of the former, he just shrugs, "I bought high and sold low. But at least money was cheap in those days." But history has shown that the latter decision -- for which Mr. Barrett took a lot of flak (he uses an earthier term) -- may have been the most brilliant of his career. Intel not only came out of the downturn faster and stronger than its competitors but still had a capacity shortage, which could have been devastating had Mr. Barrett not made the investment.

- Consensus is good -- except when it isn't. "I remember being in a meeting at one of our plants with 21 of our manufacturing managers. We started talking about changing our factory model and one of the veteran managers -- one of those guys who kind of ran his own little kingdom like a prince, said, 'Are we going to discuss this? Are we going to get a vote on this?' And I said, 'Yeah, we'll vote -- the only problem for you is that I get 22 votes.' In other words, there's a time to let everyone twist the knobs and a time to make a decision."

- Follow the business, not Wall Street. "The job of the CEO is not to reward the short-term speculator of your stock," Mr. Barrett says, "but to do a good job long-term for your shareholders, employees and customers. You don't invest for 'let's have a 20% lay-off tomorrow to prop up our stock' or 'let's cut R&D to get a positive response from Wall Street.' Thank God for Moore's Law, because it won't let us think like that; because if we do we get hammered."

- When something works, don't re-invent it, reproduce it. Perhaps Mr. Barrett's greatest contribution to the semiconductor industry was the concept of "Copy Exactly," the absolutely exact reproduction of successful existing practices and facilities in other locations. Copy Exactly has been the key to Intel and other chip companies actually improving yield rates (the ratio of chips that actually work) even as the products themselves have become thousands of times more complex and miniaturized and fabricated by the millions. The decision not to reinvent the wheel every time was, in fact, the subject of that contentious meeting where Mr. Barrett outvoted his managers. "I got the idea from McDonald's," he says. "I asked myself why McDonald's french fries tasted the same wherever I went. That's what I told my guys, "We're going to be the McDonald's of semiconductors."

- It pays to have good competitors. Despite AMD's victory with the EU, Mr. Barrett has nothing but praise for Intel's competitors, which have ranged over the years from Motorola to NEC to Samsung to, always, AMD. "It's like athletes: To be a great company you need great competitors," he says. "It's what keeps you alive and keeps you honest."