ASUS tests its Gallop humanoid on a real factory packing line
ASUS is training its Gallop humanoid on a motherboard packing station, aiming for 99.5% success in 20 seconds, with customer pilots in 2027 and sales targeted for 2H 2028.

ASUS has described Gallop, its first humanoid robot, which is already being taught a real job at an ASUS group electronics factory. The task is packing motherboard accessories: 15 small parts, many in thin clear bags, handled in about 20 seconds per board. The target is a 99.5% success rate within that time, the same bar applied to a human worker.
Gallop learns mainly by watching experienced staff, who wear head cameras with their consent. Simulation then creates variations of those examples, and an operator in a VR headset and gloves guides the robot on the actual parts. A planner AI splits the job into steps, and a second AI checks each one. Everything runs on one NVIDIA Jetson AGX Thor computer, and better AI models can be swapped in without rebuilding the robot. For now it works from a fixed base with two arms.
Why it matters
Many humanoid demos show a robot doing something once. ASUS is instead measuring Gallop against a station's daily quality rules, which is the test factory buyers care about. Next come ASUS's own dexterous hand with a sense of touch, a wheeled version, and a showing at Computex 2027, then customer pilots in 2027 and market availability targeted for the second half of 2028. ASUS has not yet shared current success rates, speeds, or pricing, so the 99.5% goal remains unproven in public.