Very telling comment by Intel CEO Barrett...
Silicon Valley Pioneer Sees Potential Future In Silicon-Less Chips
Monday April 19, 10:55 am ET
By James Detar
When you talk about founders of the computer revolution, you only need a few names. One is William Shockley.
In 1957, a group of Silicon Valley-based Shockley Semiconductor's key engineers left "the father of the transistor" to form Fairchild Semiconductor, earning the nickname the Fairchild Eight or the Traitorous Eight. And in 1968, two of the eight - Robert Noyce and Gordon Moore - left Fairchild to start Intel. (NasdaqNM:INTC - News) Their focus was memory chips, but they shifted to microprocessors, the brains of computers.
By the time IBD started in 1984, Intel was already a giant, with its chips powering most IBM-compatible PCs. The next year, it released the 386, the first mainstream 32-bit chip. It had 100 times more transistors than its first microprocessor, the 4004, circa 1971.
Ask Intel Chief Executive Craig Barrett what the future of the chip industry is, and you get a surprising answer. Chips of the future might be without that staple material called silicon.
This comes from the head of the company that helped put the silicon in Silicon Valley.
Barrett, a 30-year Intel veteran, left a post as a professor at Stanford University to join the startup in 1974. Based on Intel's bylaws, Barrett must give up the CEO reins next year when he turns 65.
He recently spoke with IBD.
IBD: Where do you see the chip industry in the next 20 years?
Barrett: Looking out 10 to 20 years, one exciting prospect is the question: How far does our silicon technology take us? What comes next? I'm not predicting the demise of silicon, but we're going to a transition. Mechanical relays went to vacuum tubes, which went to transistors (on silicon), which will go to something else. Intellectually, there's this great challenge as to what's next.
These transitions are an exciting time. Typically, new leaders emerge and new markets emerge. It's something more than just a philosophical interest here at Intel. I can't name a vacuum tube company that made it into the transistor era. The question is: Who will make the transition?
IBD: What were Intel's best and worst moves the last two decades?
Barrett: Probably if you went back a little over 20 years, one of our not so smart moves was to ignore the possibility there might be a market for PCs. Microprocessors were in first-generation PCs. But not enough people thought there was a market for that. We left that alone. That was back in the mid-'70s.
But then getting out of DRAMs (dynamic RAM, the memory chip that quickly became a commodity) and focusing on microprocessors was a good move. We had the technology and we caught the wave - the PC wave.
IBD: What were the chip industry's best and worst moves in that period?
Barrett: Probably the best move was collectively making Moore's Law (which says the number of transistors that can be put on a chip will double about every 18 months) continue to work. That allows us increasingly to get more functionality at lower cost. It's an industrywide phenomenon.
We were slow in getting to this convergence between computers and communication.
IBD: Is chip technology keeping up with user needs?
Barrett: I think so. Usually people twist that argument the other way: Is technology outrunning user needs? You have to assume some degree of balance there.
That's true especially this last year or two; we've seen a great influx of digital cameras and mobile handhelds and laptop computers, and the growth of Wi-Fi and wireless technology. Those are advancing fast to meet end-user needs.
IBD: What are the hottest markets for chips today?
Barrett: Wireless capability is a hot market. Wireless will be everywhere from handhelds to laptops to the digital home. <font color=red>The next stage will be effectively the software-configurable wireless capability. It means having the device self-configure itself to the best connectivity available to it.</font>
IBD: How long before these devices that connect themselves to the Web come to market?
Barrett: The basics are there today. It's a matter of integration. Our industry's purpose is to innovate and integrate. The innovation is there. Now it's a matter of integration to get into a cost-efficient mode.
IBD: What trends and issues are shaping the chip industry?
Barrett: Infrastructure cost is one: how much it costs to build a factory (billions of dollars) and the next set of masks (chip design blueprints). A mask set may be $1 million-plus.
IBD: What can chipmakers do to keep costs down?
Barrett: Volume. The industry has two approaches. Independent device makers like Intel have the wherewithal to invest front to back. At the other extreme are the fabless companies that go to silicon factories for their manufacturing. They built the foundry industry in order to service companies that have subcritical mass.
IBD: What new business uses will tomorrow's chips enable?
Barrett: We will probably continue to see an extension of trends we've watched many years. We'll bring mainframe capability to the desktop. The health sciences and computer worlds will converge. Health science has been one of the last to use computers for record keeping and database management. That would make that segment more productive.
IBD: What must U.S. chipmakers do to retain world leadership?
Barrett: We need to have substantial investment in research and development. And to hire the best and brightest in the world.
IBD: Is it harder to recruit the best and brightest when firms in China and India are hiring engineers that used to migrate to the U.S.?
Barrett: Nobody said it would be easy. It will be more competitive. Things that are exact parallels are basketball and semiconductors. If you want to have a world champion basketball team, you have to do the training and make the investment. If you want the leading semiconductor industry, you'd better have the training, which is where our universities come in. And the platform to compete is the R&D that we, the companies, do. |
