The Intel 4004 Microprocessor and the Silicon Gate Technology
A testimonial from Federico Faggin, designer of the 4004 and developer of its enabling technology
Busicom was a Japanese manufacturer of electronic calculators. The company was one of many international contenders in the nascent electronic calculator business. The availability of MOS LSI (large-scale-integration) technology, had just started a fundamental change in the calculator business from electromechanical to electronic. In 1969, Busicom had developed a modular design for a family of high-end calculators based on a special-purpose, macro-instruction computer that operated on shift-register memory. Looking for a competitive advantage and having heard that Intel had the highest density and fastest technology available, Busicom went to Intel with the purpose to have their logic design converted into seven chips. Ted Hoff, the manager of the Application Department of Intel evaluated the proposal and felt that the Busicom design could be both simplified and improved by replacing the special-purpose, ROM and shift-register-based computer with a general-purpose, ROM and RAM-based architecture. Working with Busicom engineers and later with Stan Mazor, Hoff defined the basic architecture of the computer, inspired by the architecture of other CPU's of the time, and was able to convince both Intel and Busicom managements that his approach was superior to the previous one.

Hoff’s proposal consisted of four different chips: a 2048-bit ROM with a 4-bit programmable input-output port (4001); a 4-registers x 20-locations x 4-bit RAM data memory with a 4-bit output port (4002); an input-output expansion chip, consisting of a static shift register with serial input and serial/parallel output (4003); and the 4-bit CPU chip (4004). Busicom entered into an exclusive contract with Intel for the development of the chip set in October 1969. Intel agreed to start the design immediately and produce working samples one year later. The proposal, however, lay dormant for six months until Faggin, hired to lead the project, joined Intel in April 1970 and started to work furiously to contain the schedule slippage.