Integrated circuit

In 1958 Jack Kilby at Texas Instruments builds an integrated circuit of germanium, several components on one piece. In 1959 Robert Noyce at Fairchild builds a silicon version with a flat process that can be manufactured. The chip, as industry, is Noyce's approach. The invention is shared by both.

The contribution is density. Wiring thousands of separate transistors by hand does not scale. Printing them on one wafer does. That is the difference between a 1960s computer and a processor with billions of switches. Phones, factory robots, and neural nets are all limited by how many switches you can afford. This is the step that makes the number huge.

A chip still needs the transistor. It does not replace it. It is a way of making many of them at once, and of connecting them with metal traces instead of solder.

The 1959 date on the card leans on the silicon planar circuit, which is the one that scaled.