The hidden history of industrial robotics

01The inventor who read Asimov

George Devol was not a roboticist. He was an inventor with patents in radar, microwaves, and industrial controls. In 1954 he filed a patent for a programmed article transfer device — a mechanical arm that could repeat a sequence of movements. The patent, issued in 1961, is the foundational document of industrial robotics.

Devol was inspired partly by science fiction. Isaac Asimov’s stories of obedient machines had shaped the public imagination, and Devol saw a real-world application: moving hot, heavy parts in factories. He built a prototype using hydraulic actuators and magnetic drum memory. It was crude by today’s standards, but it was the first machine that could be programmed to perform a physical task and repeat it.

02The salesman who made it real

Joseph Engelberger met Devol at a cocktail party in 1956. Engelberger was an engineer and entrepreneur who had worked on jet engines and nuclear handling systems. He recognised that Devol’s patent could become a product. Together they founded Unimation, the world’s first robotics company.

Engelberger was the evangelist. He spent years trying to sell the Unimate to manufacturers who had never seen a robot and did not understand why they needed one. The breakthrough came when General Motors agreed to install a Unimate on its die-casting line in 1961. The robot removed hot parts from a die-casting machine — a job that was dangerous, repetitive, and unpleasant for human workers.

03Nuclear roots and the arms race

Before Unimation, the technologies that would enable robotics were developed for a different purpose: handling radioactive materials. The Argonne National Laboratory built master-slave manipulators in the 1940s and 1950s, where a human operator moved a master arm and a slave arm in a hot cell mirrored the motion. This was teleoperation, not automation, but it established the mechanical and control principles.

Cold War research funding accelerated the development of servomechanisms, encoders, and feedback control. The same technologies that guided missiles and handled nuclear fuel were repurposed for factory automation. The lineage from military research to industrial robotics is direct and underappreciated.

04The automotive adoption wave

Through the 1970s and 1980s, automotive manufacturing became the dominant market for industrial robots. Spot welding, spray painting, and assembly were repetitive, physically demanding tasks that robots could perform consistently. GM, Ford, Chrysler, and eventually Toyota and Honda all adopted robots on their lines.

The Japanese robot industry emerged as a formidable competitor. Companies like FANUC, Kawasaki, and Yaskawa licensed or adapted Unimation’s technology and then improved it. By the 1980s, Japan was the world’s largest producer and user of industrial robots. Unimation, which had created the industry, was sold to Westinghouse in 1983 and later to Stäubli.

05The computer revolution and the Stanford Arm

In 1969, Victor Scheinman at Stanford University built the Stanford Arm, the first all-electric, computer-controlled robotic arm. Earlier robots like the Unimate used hydraulics and magnetic drum storage; the Stanford Arm used DC servos and a minicomputer. This is the architecture that all modern industrial robots inherit.

The move to electric actuators and digital control enabled finer precision, cleaner operation (important for electronics and food), and eventually the ability to program complex trajectories in software rather than by teach pendant alone. The Stanford Arm design was commercialised as the PUMA (Programmable Universal Machine for Assembly) by Unimation in 1978, a robot that became a standard in research labs and factories for decades.

06Collaborative robots: the next frontier

The concept of a collaborative robot — a robot safe enough to work alongside humans without guarding — emerged in the mid-1990s through research at Northwestern University by J. Edward Colgate and Michael Peshkin. Their insight was to limit force and speed rather than rely on sensors and software for safety.

The first commercially successful cobot was the UR5, launched by Universal Robots of Denmark in 2008. It was small, lightweight, easy to program, and affordable compared to traditional industrial robots. The cobot market has grown rapidly, bringing robotics to small and medium manufacturers who could never justify the cost and complexity of a traditional robot cell.

07What the history reveals

The hidden history of industrial robotics reveals that innovation is not a single leap but a chain of contributions across decades. Devol patented the concept. Engelberger built the business. Scheinman made it electric and digital. Japanese manufacturers made it reliable and affordable. Danish engineers made it accessible to small firms. Each generation stood on the work of the previous one.

The history also reveals that the driving force was not curiosity but necessity: handling nuclear fuel, removing workers from dangerous die-casting lines, improving automotive quality, and reducing labour costs. Robotics did not descend from the laboratory to the factory. It was born in the factory, shaped by the needs of production, and only later migrated into research, service, and consumer applications.