Rabbit Hole MapsAbout

Technology & Medicine › Digital age

Microprocessor (Intel 4004)

November 1971· Invention or discovery· Communication & computing

Well documentedSolid records, physical evidence or scientific consensus.

A computer's entire central processor on a single chip, first sold by Intel as the 4004 in 1971.

See it on the Technology & Medicine map →

How it works

Japanese calculator maker Busicom asked Intel for a set of custom chips. Engineer Ted Hoff proposed instead one general-purpose, programmable processor; Federico Faggin designed it, with Stanley Mazor and Busicom's Masatoshi Shima. The 4004 held 2,300 transistors and was roughly as powerful as the room-sized ENIAC, on a chip smaller than a fingernail. Intel bought back the rights and advertised it in November 1971. The 8080 (1974) and 8086 (1978) followed.

Why it mattered

Programmable processors became cheap parts, spreading computing into calculators, cars, appliances, personal computers and phones; the 8086's descendants still run most PCs.

Think of it like

Like replacing a building's separate hard-wired controls for lights, heating and elevators with one programmable building-management system.

How we know

Intel's design records, Faggin's and Shima's oral histories, Busicom calculators using the chip, patents, and Intel's advertisement in Electronic News on November 15, 1971.

Still unknown or debated

Other claimants exist: Garrett AiResearch's processor for the F-14 fighter (1970, secret until 1998) and Texas Instruments' TMS1000; which counts as first depends on definitions.

Inside this (2)

Connections (5)

▸Integrated circuit (microchip) led to the Microprocessor (Intel 4004)Well documented

The connection

As chips held more transistors, roughly doubling every two years, engineers could fit an entire processor, with its arithmetic, logic and control, onto one chip. Intel's 4004 packed 2,300 transistors, made possible by the silicon-gate method that Federico Faggin had developed at Fairchild in 1968.

When and how

Intel in Mountain View, California, 1969–1971, designed by Federico Faggin, Ted Hoff and Stanley Mazor with Masatoshi Shima of the Japanese calculator firm Busicom.

How we know

Intel's design records and Faggin's accounts show the 4004 depended on denser chip-making; Gordon Moore's 1965 article had predicted such integration.

Debates and open questions

How long the doubling can continue is debated: features are now only a few dozen atoms wide, and gains increasingly come from stacking chips and specialised designs.

▸Microprocessor (Intel 4004) led to the Personal computerWell documented

The connection

The microprocessor made a computer cheap enough for one person. The Altair 8800 used Intel's 8080, a successor to the 4004; the Apple II used the MOS 6502, which sold for about $25; the IBM PC used Intel's 8088. That x86 family still runs most desktop and laptop PCs.

When and how

Albuquerque, New Mexico (Altair, 1975); Cupertino, California (Apple II, 1977); Boca Raton, Florida (IBM PC, 1981).

How we know

Each machine's specifications name its processor; the January 1975 Popular Electronics cover story on the Altair featured the 8080 chip.

Debates and open questions

Intel itself did not foresee home computers; how much the PC owed to chipmakers versus hobbyist groups such as the Homebrew Computer Club is debated.

▸Microprocessor (Intel 4004) powered the Mobile phoneWell documented

The connection

A cellular phone must constantly report to towers, switch towers mid-call and turn voice into digital code. Early handsets needed many boards of chips; processors and specialised chips shrank all this to pocket size. From the 1990s, power-saving processor designs from Britain's ARM let phones run for days on a charge.

When and how

Motorola's labs in the 1970s and 1980s; ARM designs, first created at Acorn in Cambridge, England, in 1985, adopted by Nokia and other phone makers in the 1990s.

How we know

Handset teardowns and chip-industry data show processors at the heart of every phone; ARM reports its designs in almost all smartphones.

Debates and open questions

Little is disputed; an open engineering question is how phones will keep gaining speed and battery life as transistor shrinking slows.

▸Microprocessor (Intel 4004) translates surgeons' hand movements in Robotic surgeryMostly agreed

The connection

A surgical robot is a computer-controlled machine: processors read the surgeon's hand movements at the console thousands of times a second, scale them down, filter out tremor and drive the motors in the instrument arms. Without cheap, fast microprocessors, descendants of the 1971 Intel 4004, this real-time control would be impossible.

When and how

Intel in Santa Clara, California, 1971; SRI International and Intuitive Surgical in Silicon Valley, from the late 1980s to da Vinci's approval in 2000.

How sure are we?

Engineering descriptions of da Vinci and earlier systems document their reliance on digital processors for motion scaling and tremor filtering.

Debates and open questions

The link is general rather than specific: robotic surgery depends on all of modern computing, not on the 4004 itself.

▸Microprocessor (Intel 4004) evolved into the GPU chips driving Deep learning (modern AI)Well documented

The connection

Microprocessors put a computer's logic on one chip and grew millions of times more complex under Moore's law. Graphics chips (GPUs), designed for video games, pack thousands of simple processors doing math in parallel, exactly what neural networks need. Around 2009–12 researchers found GPUs sped training dozens of times; Nvidia's chips now power most AI.

When and how

Intel's 4004, 1971; Nvidia's GeForce 256, marketed as the first GPU, 1999, and its CUDA programming toolkit, 2006; AlexNet trained on two Nvidia gaming cards, 2012.

How we know

Papers such as Stanford's 2009 GPU study and AlexNet (2012) report the speedups; Nvidia's data-center revenue shows the scale today.

Debates and open questions

Whether specialized AI chips, like Google's TPUs, will overtake GPUs, and whether chip supply and electricity can keep pace with AI's appetite, is debated.