1968: Intel is founded

On 18 July 1968, two engineers who had helped invent the integrated circuit walked out of Fairchild Semiconductor and incorporated a new company in Mountain View, California. Robert Noyce and Gordon Moore initially called it NM Electronics; within months it became Intel—short for Integrated Electronics. Their modest ambition was to produce semiconductor memory chips that would eventually replace magnetic cores in computers.

Success came swiftly. In 1970 Intel introduced the 1103 DRAM, which rapidly became the industry standard for computer memory. The following year brought the boldest leap of all: the Intel 4004, the first microprocessor. Designed originally for a Japanese calculator manufacturer, this single chip held 2,300 transistors and could perform any computation that a room-sized machine had once required. It was the birth of the “computer-on-a-chip.”

The intellectual thread stretches further back. The transistor itself grew out of solid-state physics. Boole’s algebra of logic had shown how reasoning could be reduced to binary operations; Claude Shannon later demonstrated that electrical switching circuits could implement that logic. Noyce and Moore simply completed the miniaturisation that earlier theorising had made conceivable, packing ever denser logic onto silicon.

That same transistor density, later formalised as Moore’s Law by Gordon Moore himself, delivered the raw computational power without which modern artificial intelligence would be impossible. Expert systems of the 1980s still depended largely on specialised workstations; the statistical-learning renaissance of the 1990s needed affordable personal computers; the deep-learning explosion after 2012 required GPU clusters and process technologies that Intel had helped pioneer. Every large language model trained today ultimately runs on chips whose design principles descend from the 4004.

Today’s AI landscape still rests on that foundation. Whether training trillion-parameter models in data centres or deploying tiny neural nets on edge devices, the industry continues to chase density, speed and energy efficiency—the same race that began when two men signed incorporation papers on an ordinary July day in 1968. The hardware substrate of artificial intelligence was, in a very real sense, born that afternoon.