Collaborative Robots
Collaborative robots (cobots) are robots designed to work alongside humans in shared workspaces without the safety cages, fencing, and exclusion zones that traditional industrial robots require. Cobots achieve this through force-limiting: if the robot contacts a human, it detects the collision and stops or reduces force before causing injury. The cobot market (≈$2.3 billion in 2025) represents a bridge technology between traditional industrial automation and the general-purpose humanoid robots now entering production.
How Cobots Differ from Industrial Robots
Traditional industrial robots (from Fanuc, ABB, KUKA) are fast, powerful, and dangerous. They operate behind safety fences at speeds and forces that would seriously injure a human. Cobots invert these priorities: they sacrifice some speed and payload capacity for the ability to share space with human workers. This means cobots can be deployed on existing production lines without facility redesign — a worker can hand a part to the cobot, the cobot performs a repetitive subtask (screwing, gluing, inspection), and the worker handles the next step. No cage, no exclusion zone, no facility rebuild.
Universal Robots (UR) pioneered the modern cobot market and remains the leader with its UR3e, UR5e, UR10e, and UR20 arms. Fanuc's CRX series brings traditional industrial robot reliability to the cobot form factor. ABB's GoFa and YuMi target precision assembly. Doosan Robotics offers cobots with higher payload capacities. The common thread: 6-axis arms with force/torque sensing at every joint, designed for tasks like assembly, machine tending, pick-and-place, quality inspection, and palletizing.
Cobots and AI
The integration of AI is transforming cobots from programmable arms into adaptive workers. Traditional cobot programming requires manually specifying waypoints and trajectories — essentially teaching the robot a fixed sequence of movements. AI-powered cobots can use computer vision to locate parts in varying positions, VLA models to understand natural language instructions, and imitation learning to acquire new skills from human demonstrations rather than manual programming. This shifts cobots from "programmable" to "trainable" — the same paradigm shift happening in humanoid robotics, but applied to the simpler arm form factor.
Cobots vs. Humanoids
The relationship between cobots and humanoids is complementary, not purely competitive. Cobots excel at tasks requiring a single arm operating from a fixed or rail-mounted position: repetitive assembly, machine tending, quality inspection. Humanoids are needed when the task requires mobility (walking between workstations), bimanual manipulation (two arms coordinating), or working in spaces designed for human bodies. Many facilities will deploy both: cobots for fixed-position repetitive tasks, humanoids for mobile and flexible tasks. The cobot market also serves as a training ground for the AI and control technologies that humanoids need — sim-to-real transfer, force-sensitive control, and human-robot interaction patterns developed for cobots transfer directly to humanoid development.