Military Electronics
Military electronics is the family of radar, guidance, communication, fuzing and computing equipment built for armed forces, and in the 1950s and early 1960s it was the most important early customer and funder of transistors and integrated circuits. Civilian markets for transistors, such as hearing aids and pocket radios, were real, but they were small and price-sensitive. Defense agencies wanted devices that were small, light, rugged, low-power and reliable, and they were willing to pay a large premium and to underwrite factory development to get them, which is why so much early semiconductor history passes through military contracts.
The military interest began at the start: Bell Labs shared the transistor with defense users soon after its announcement. The US Army Signal Corps, concerned with battlefield communications, supported transistor development and manufacturing improvements, and Bell Labs held a symposium in 1951 for military and academic representatives, ahead of the better known April 1952 meeting at which commercial licenses were offered for 25,000 dollars. Military sponsors contributed funding for production engineering, which helped Western Electric and licensees improve yields. Defense agencies also funded the move from germanium to silicon, since silicon devices could withstand the temperatures found in aircraft and missiles.
Reliability, not just size, was the driving requirement. A missile cannot be repaired after it is launched, and a guidance computer containing thousands of components multiplies the chance that one of them fails. Military agencies therefore developed qualification regimes: standardized part types with JAN designations, specifications such as MIL-S-19500 for semiconductors, screening and burn-in tests, and later MIL-STD-883 test methods for microcircuits. These rules forced manufacturers to measure and control failures statistically. They also raised costs, which is one reason that military electronics were expensive relative to civilian parts, and why commercial firms later adapted, rather than copied, the standards.
Several specific programs turned that demand into manufacturing capacity. Missile guidance for Atlas and then Minuteman required compact, high-reliability digital electronics, and Autonetics bought silicon transistors and later integrated circuits in volume. The Army's Diamond Ordnance Fuze Laboratories, a proximity-fuze research group, produced Jay Lathrop and James Nall's photolithographic technique for printing transistor patterns around 1957-58, and military micromodule and molecular-electronics programs funded competing routes toward miniaturization. When Texas Instruments and Fairchild produced integrated circuits at the end of the 1950s, government and aerospace purchasers were the first buyers, which made it possible to build production lines before prices fell far enough for civilian computers to use the chips.
A common misconception is that the military invented the transistor or the integrated circuit. Both were invented by company engineers, Bell Labs for the transistor and Texas Instruments and Fairchild for the integrated circuit, and the government's role was that of a patient, demanding and well-funded customer. The opposite exaggeration is also wrong: Silicon Valley did not grow only from defense money, because the private firms soon sold to computers, instruments and consumers, and venture capital took over much of the funding. Both sides contributed, with defense purchases providing early volume and civilian growth providing long-term scale.
The pattern persists in modern technology. Defense research agencies still fund early work that markets are not yet ready to pay for, and standards for high-reliability parts continue to differ from commercial ones. The idea that a demanding early customer can fund a technology down its learning curve until it becomes cheap enough for everyone has become a standard model in industrial policy.
Read military electronics as the first great customer of the transistor: its demand for tiny, rugged, dependable parts paid for the learning that later made chips cheap enough for ordinary products.