What is a transistor, and why was it the most important invention of the 20th century?
A transistor is a switch with no moving parts that can also amplify. Once you have one, you can build any logic function, any memory cell, any signal processor — and you can put a billion of them on a fingernail. Vacuum tubes did all of this first; transistors did it small, cheap, and reliable enough that civilization could afford to put them in every object.
The transistor is not impressive in isolation. The bottleneck shift is that for the first time, a switching element became cheaper than the labor required to wire it up — which made discrete electronics obsolete and the integrated circuit inevitable.
What the transistor does
A transistor has three terminals. A small signal applied to one terminal — the gate, or in older designs the base — controls a much larger current flowing between the other two. That is all it does. It is a faucet for electrons, and the handle is electrically activated.
This single behavior covers two distinct uses. As a switch, you drive the gate hard enough to turn the device fully on or fully off, and the transistor becomes a controlled bridge in a circuit. Stack switches in the right pattern and you get NAND, NOR, NOT, and from those you can build any Boolean function, any adder, any register, any memory cell. As an amplifier, you let the gate signal swing gently, and the larger current responds in proportion — turning a weak microphone signal into a speaker-driving signal, or a faint antenna signal into a usable receiver output.
Both uses require the same device. The same piece of silicon runs your CPU and your guitar amp, with only the circuit around it different. That generality is why the transistor showed up everywhere at once.
December 1947
John Bardeen and Walter Brattain were two physicists working at Bell Labs under William Shockley, who ran the solid-state physics group. The team had been trying for years to build a solid-state replacement for the vacuum tube — a device that could amplify signals without the heat, fragility, and power draw of a glass bulb full of hot wire.
On December 16, 1947, Bardeen and Brattain pressed two thin gold contacts into a sliver of germanium, applied voltages, and watched the output signal swing larger than the input. They had built the first working point-contact transistor. Shockley, who had been pushing a different design and was not present for the breakthrough, was reportedly furious about being scooped on his own team. Within a month he had designed the more robust bipolar junction transistor, which became the version that actually went into production.
All three shared the 1956 Nobel Prize in Physics. The patent on the device, US 2,524,035, lists Bardeen and Brattain as inventors. The corporate value, which is harder to measure but is somewhere in the trillions of dollars, has not been collected by anyone in particular — Bell Labs licensed the technology widely.
Source: Bell Labs records; Bardeen-Brattain-Shockley 1956 Nobel Lecture.
Why this was bigger than the lightbulb
The vacuum tube already did everything the transistor does — switching and amplifying. The 1940s ENIAC computer was built from 17,000 vacuum tubes. The first transatlantic submarine telephone cable was full of tube amplifiers. Radio, television, radar, hearing aids, and the entire early electronics industry ran on tubes.
The problem with tubes was not their function but their cost structure. A tube is a fragile glass envelope around a heated filament; it draws watts of power just to warm up, it fails on a Poisson schedule, and it is large enough that putting more than a few thousand of them in one machine becomes a serious building-design problem. ENIAC needed its own room and a dedicated team to replace burned-out tubes each day.
Transistors removed every one of these constraints at once. They were solid (no glass to break), used millivolts and microamps of control power, lasted decades, fit in your palm, and could be mass-produced photolithographically. Within ten years they were cheaper than the brass sockets that held vacuum tubes in place. Within twenty years they were cheaper than the solder joints that connected them to circuit boards. The cost crossover is what set the rest of the industry in motion.
The integrated circuit follows logically
Once a single transistor cost less than the labor to wire it up, the obvious next move was to put many transistors on the same piece of silicon and wire them up at the silicon level rather than the workbench level. Jack Kilby at Texas Instruments and Robert Noyce at Fairchild Semiconductor independently arrived at this idea within months of each other in 1958-59. Kilby used germanium and wires bonded by hand; Noyce used silicon and planar deposition, which was the version that scaled.
The integrated circuit was less an invention than a consequence. If you accept the transistor and you accept that wiring labor dominates the assembly cost, the integrated circuit is what you do next. The interesting question becomes not whether to make ICs but how many transistors you can fit on one — which is the question Gordon Moore would answer six years later, and which still governs the industry today.
Strategic read
The transistor matters not because it does something a vacuum tube could not do, but because it changed the cost curve of a function society already wanted in unlimited quantities. The lesson generalizes: a new technology rarely succeeds by doing something fundamentally new. It succeeds by making the cost of an old function low enough that the function gets applied to a thousand places nobody bothered with before.
AI is the current example. Large language models do not do anything human writers, illustrators, programmers, or analysts cannot do. They make those functions cheap enough that they get applied at scales nobody could previously justify. The right comparison is not "AI vs. human" but "AI vs. the unit economics that gated each of these jobs to high-stakes, high-budget situations." When the unit economics flip, the function shows up everywhere — exactly the way transistorized radios, hearing aids, and pocket calculators showed up in the 1950s and 60s.