Factory-Proven Physical AI: Why ASUS Is Building a Humanoid Robot

ASUS introduced Gallop, a humanoid robot developed to address a critical question from factory managers: whether a robot can perform specific tasks consistently at a workstation to their quality standards. Unlike robots that merely demonstrate capabilities in videos, Gallop is engineered to operate in real factory environments, handling tasks such as packing motherboard accessories. The robot is designed to work alongside human operators, taking on repetitive or physically demanding roles that are difficult to staff, rather than replacing workers.

Gallop operates through a four-layer system: a planner that breaks down tasks, a supervisor that monitors and adjusts steps, a skills layer for precise movements, and a control layer for real-time adjustments. The robot uses a single NVIDIA Jetson AGX Thor computer and standard robotics software, with ASUS designing its body and dexterous hands in-house. The hand, equipped with a sense of touch, is a particularly challenging component, and ASUS builds it internally due to its scarcity in the market.

The Gallop program began in early 2026, with plans to introduce a wheeled version, enhance its dexterous hand, and showcase the robot at Computex 2027. Customer pilots are scheduled for 2027, with market availability targeted for the second half of 2028. Initially, Gallop will focus on electronics manufacturing, leveraging ASUS’s expertise in the industry, but its skills are expected to transfer to other sectors like hospitals and service sites.

ASUS emphasizes that Gallop is not intended to replace workers but to assist them by handling repetitive tasks, allowing human operators to transition into roles such as setting up, supervising, and improving the robots. The company highlights that the robot’s success will be measured by its ability to perform tasks daily at the same standard as a human worker, with a target of 99.5% success within 20 seconds for its initial task at an ASUS EMS plant.