The transistor at Bell Labs: three physicists and a mutiny
Bardeen, Brattain, and Shockley invent the switch that replaced tubes — then Shockley's company drives the Traitorous Eight to found Fairchild and, eventually, Intel.
In December 1947, in a laboratory at Bell Labs in New Jersey, three physicists watched a tiny germanium device amplify a signal. John Bardeen, Walter Brattain, and William Shockley had built the first working transistor — a solid-state switch that would replace fragile vacuum tubes. They demonstrated it to executives a week later. In 1956 they shared a Nobel Prize.
Then Shockley left, founded a company in California, alienated his best staff, and those staff founded Fairchild Semiconductor — which spun out Intel and a dozen other firms. The transistor's invention is a physics story. The transistor's industry is a soap opera with stock options.
Why did vacuum tubes have to die?
Early computers and radios ran on tubes: glass bulbs that controlled current with heat and voltage. They worked. They also burned out, drew huge power, and took up a room when you needed thousands of them. ENIAC's eighteen thousand tubes failed regularly. A machine that could switch without the glass and the heat was the dream.
The point-contact transistor Bardeen and Brattain built was crude by later standards, but it proved the principle. Shockley's junction transistor improved the design. Bell Labs patented and licensed widely — a decision that seeded an industry instead of bottling it inside AT&T.
How did Shockley's personality invent Silicon Valley?
Shockley Semiconductor opened in Mountain View in 1956, chasing the new physics. Shockley was brilliant and, by many accounts, impossible — paranoid, harsh, obsessed with loyalty tests. In 1957, eight of his best engineers resigned as a group. They became known as the Traitorous Eight. Fairchild Camera and Instrument backed them to form Fairchild Semiconductor.
Fairchild made silicon transistors practical and cheap. Employees left to start their own firms — the "Fairchildren" — including Gordon Moore and Robert Noyce, who founded Intel in 1968. Venture capital, stock options for engineers, and the culture of spinning out startups crystallized around this exodus. Shockley never built a lasting company. His former employees built the landscape.
What did Moore's Law actually say?
In 1965, Gordon Moore noticed that the number of components on a chip roughly doubled on a regular schedule, dropping the cost per function. The observation became a target the industry aimed at for decades — not a law of physics, but a self-fulfilling roadmap. Smaller transistors, denser chips, cheaper computers, then phones, then the AI accelerators running today's models.
Every smartphone is a grandchild of that Bell Labs bench and that Fairchild mutiny. So is every data center GPU. The physics started with a few men and a crystal. The scale came from an industrial race to shrink switches until they became invisible and everywhere.
Why tell the transistor story now?
Because modern tech loves software myths — garages, apps, overnight unicorns — and forgets that software sits on switches someone had to invent. The transistor turned electronics from a specialist craft into a mass industry. Without it, there is no personal computer, no internet hardware, no AI chip.
The other reason is cultural. Bell Labs shows what patient industrial research can do. Shockley's failure shows that brilliance without collaboration can still lose the future to the people who walk out the door. Silicon Valley's founding legend is often told as destiny. It is also a reminder that institutions and egos shape who gets to build the next switch.