Three mavericks, part of the Traitorous Eight, help found Silicon Valley and the semiconductor industry.
Where were we? Oh yes -- the three guys at AT&T Bell Labs who tackle the problem of burned-out vacuum tubes and wind up creating Silicon Valley. Let's start with William Shockley.
Shockley hates New Jersey (where Tony Soprano is the first to say "Garbage in, garbage out -- do you have any problem with that?"), goes home to Palo Alto, Calif., and starts Shockley Semiconductor in an old prune orchard. An aside: Would the place have had as much allure if we called it Prune Valley instead of Silicon Valley?
Shockley is a terrible manager but has a first-rate mind, so he knows that if his semiconductor company is going to succeed he will need Very Bright People, and he begins to recruit them. A few years later, these Very Bright People -- Gordon Moore, Robert Noyce, Eugene Kleiner, Jean Hoerni, Victor Grinich, Julius Blank, Jay Last and Sheldon Roberts -- have had enough of Shockley and his dysfunctional company and decide to bolt.
From left to right: Gordon Moore, C. Sheldon Roberts, Eugene Kleiner, Robert Noyce, Victor Grinich, Julias Blank, Jean Hoerni and Jay Last (1960).
This group, known in history as the "Traitorous Eight," turn to Arthur Rock, an investment banker, for financing. With additional financial support from Stephen Fairchild, they form Fairchild Semiconductor.
Transistors are hot, hot, hot -- and sell for $100 each! --mainly because they aren't vacuum tubes, are cheaper to build and are super reliable. Somebody figures out that silicon is kind of like germanium and is as cheap as sand. Hell, it is sand.
Fairchild Semiconductor sees what Texas Instruments is up to by putting more than one transistor on a device and figures it can do better. It uses optical lithography and a mask, which allows it to put eight (eight!) devices on one chip. TI and Fairchild sue each other for a decade, then they agree to joint ownership of the integrated circuit. Customers want even more and more elements on a chip, so Moore, Fairchild's R&D chief, postulates that the number of transistors would double every year ... and the costs would drop correspondingly.
Moore is able to dine out on that one theorem for the next 50 years. It's a theorem as powerful as E = MC2. All of a sudden, costs not only don't double when you double the number of components; they actually fall and markets explode. Fairchild and TI are on the edge with better technology in a smaller size at half the price.
By 1968, it's time for Moore and Noyce to say goodbye to Fairchild. They call their old friend Rock, the guy who brought Stephen Fairchild in the last time, to raise money for a new company, Intel. (Rock also helped finance another little company: Apple). They have no trouble raising money because Moore and Noyce are known quantities and the suspicion is that something remarkable is happening in Prune Valley. Intel gets started by making transistors like everyone else, and demand exceeds supply, but Intel is looking for Act II because its founders know that sooner or later basic transistors will become a commodity.
Remember that great deal that AT&T offered for licensing transistor technology for $25,000? One of the companies that was interested was Tokyo Tsushin Kogyo, which was struggling in the tape recorder business and searching for new products to make. It applied for a transistor license to MITI, the Japanese Ministry of International Trade and Industry, which originally turned the company down but eventually relented, letting it get a license for the AT&T patents. Tokyo Tsushin Kogyo started making little radios that fit into your pocket and changed its name to Sony. By that same action, MITI catalyzed the Japanese semiconductor industry, which by the 1980s was starting to give Intel fits.
Then there's this other company outside Chicago, called Galvin Manufacturing, which in the early 1930s was trying to get into new businesses. While one of its engineers, Bill Lear, was on a hot date, his lady friend suggested that she could be in a more romantic mood if his car could play music. Lear and his buddy invent the car radio, and Galvin changes its name to Motor-ola. Motorola understood cars and radios, but it never jumped on the consumer electronics market, so it missed what Sony saw. Motorola went into making chips itself; the early Apple computers ran on them exclusively. And, yes, Motorola did invent the cellular phone, in 1973; a cellular phone is just a special-purpose radio and will go on to become the ultimate piece of consumer electronics.
Lear goes on to get 107 patents and then builds ... the Lear Jet! Which is the plane of choice for Silicon Valley until the guys start buying G650s (Gulfstream), which you can buy today for around $70 million.
But we're getting ahead of ourselves. In the 1970s the Japanese play a critical role in getting Intel off the ground.
Enter Busicom, a Japanese calculator company, which shows up and wants 12 "custom" integrated circuits. Poor Intel is now faced with the problem that befuddles many a tech company: demanding customers that want their own solution. Intel assigns a guy named Ted Hoff to meet this demand, and he figures that the rest of his life will be spent designing custom solutions for every demanding client. Maybe a better solution is a processor that can be "reprogrammed" for each demanding customer. In other words, a generic solution for all of these specialized needs. Busicom offers to pay for this development, but it doesn't ask for ownership of the underlying intellectual property.
It's a trillion-dollar mistake. Hoff had created an entire computer, on a chip. No software yet, but some kid at Harvard, Billy The Kid Gates, along with his buddy Paul Allen, later figure out that with some instructions they could actually get a home/hobby computer to operate.
By the early 1980s, Intel is heavy into memory chips and is getting its rear handed to it by the Japanese semiconductor industry, which is cutting prices left and right. Moore asks Noyce what a new management team would do to straighten out Intel. "Why, they would get us out of memory chips," was the answer. Moore responds by suggesting that they do that themselves ... and they go heavily into computer chips and save the company.
Meanwhile, there's a thirsting need for money for all these new-fangled, tech-heavy startups, not the sort of money that Rock was able to raise from well-heeled individuals, but serious risk capital in the form of venture capital, a subject we'll cover next.
Big iron is under relentless pressure to prove its value in an x86 world. We look at five ways to bring it into the distributed data center. Also in the new, all-digital Mainframes In The Age Of The Cloud special issue of InformationWeek: Five steps to a better wireless LAN. (Free registration required.)
InformationWeek Tech Digest August 03, 2015The networking industry agrees that software-defined networking is the way of the future. So where are all the deployments? We take a look at where SDN is being deployed and what's getting in the way of deployments.