(6 of 13)
He was also in love. His wife Eva, a refugee herself, recalls their first meeting at a New Hampshire resort where they both worked in the summer of 1957--he as a busboy, she as a waitress. Eva recalls the encounter ("He had a bad accent, even though he doesn't think so!") as a lightning bolt: "I walked into this room, and there were a bunch of guys. One shook my hand, and it was, you know, like shaking a limp fish. But then there was this really good-looking guy who shook my hand, and I was just like, wow!" She still smiles at the memory, rolls her blue eyes and swallows a giggle. In June 1958 they were married.
The two moved out to California, where Grove entered the Ph.D. program at the University of California, Berkeley. Again he was a star. When he graduated, he had the pick of American research corporations. Grove narrowed his choices: prestigious Bell Laboratories or Fairchild Semiconductor, a start-up staffed by a handful of brilliant engineers. Grove, who says he has "excellent antennae," listened to the Berkeley buzz and came back with a sense of the future: Fairchild.
In the early 1960s, the computer industry was in the midst of a benign revolution--and Fairchild was a breeding ground for revolutionaries. Early computers were fast, but attempts to make them faster were running into a thermodynamic wall: every time you asked the computer to think harder, it got hotter, like a grad student sweating his orals. The heat came from vacuum tubes, which acted as giant on-off switches, holding and releasing electrical charges. (A central "computer" tallied up all the on-off signals as ones and zeroes, and translated the results into real mathematics.) But the tubes, which sucked up huge amounts of energy, represented a limit on the power of these early computers.
The logical solution was to replace the tubes: build a device that performed the same role--storing electrical charges--but that was less temperamental. The device was an electrical "switch" called a transistor, essentially a tiny electrical gate that controlled the flow of electrons that computers needed to do their math. Yet wrangling infinitesimally small electrons into place demanded phenomenally pure chemical surfaces. In the 1950s and '60s this was an act of near alchemy, certainly beyond the capabilities of most scientists. What the world needed was a reliable base for these circuits. What would it be?
The answer, of course, turned out to be what gave Silicon Valley its name. Gordon Moore (who ran Fairchild's research arm and later became Grove's mentor as CEO of Intel) believed you could store those charges with an integrated circuit made by sandwiching metal oxide and silicon into an electrical circuit called an MOS transistor. Unlike trickier semiconductors, silicon is both a wonderful conductor of electrical charges and a nearly bottomless sink for heat, meaning it doesn't melt down as you push electrons under its surface at nearly light speed. Because it is made from refined sand, silicon is abundant as the earth.
And, in MOS, unstable as hell. One day you'd run a voltage through a sample and see one thing; the next day you could run the same voltage through the same sample and get a different reading. It was a nightmare. Of course, if you could fix that little problem, you'd be onto something big.