- Start with a real job. Gallop is learning a real workstation in an ASUS group factory, not a staged demo.
- Learn from people. Gallop learns mostly by watching skilled workers do the job, then practicing on the same parts.
- Built to keep improving. Its modular design lets us swap in better AI models without rebuilding the robot.
- Prove it at home first. Customer pilots are planned for 2027, with market availability targeted for 2H 2028.
Videos of humanoid robots are everywhere right now. They dance, sort blocks, and fold laundry, and they have helped many people picture what is coming. A factory manager watching those videos asks a different question: can it do this station, for a full shift, to my quality standard? That is the question we built Gallop to answer.
S.Y. Hsu, ASUS Co-CEO, recently introduced Gallop on CNBC. Following on from this, I would like to take the opportunity to talk further about what we are building, how it learns, and what we are still working on.
Why a humanoid, and why ASUS
Until recently, AI lived on screens. Now it can see, decide, and move a machine. The parts are largely ready. What is still missing is proof that a humanoid robot can work at a real workstation, every day.
Most factories are built around people, with stations, tools, and trays sized for human hands. Traditional automation is excellent for fixed, high-volume tasks, but it struggles with work that changes from one production run to the next. A robot with two arms and dexterous hands can work at the same stations, with the same tools, so a factory doesn’t have to be rebuilt around it. Not every task needs a humanoid, and we choose the ones that make sense.
ASUS has spent decades integrating hardware and software, from motherboards to PCs to servers. We already build the kind of computing a robot depends on. And we have our own electronics manufacturing (EMS) factories to test in, on real parts, under real quality rules. Most robotics companies have to rent that.
Starting with one real station
Gallop’s first job is at an ASUS group EMS plant: packing motherboard accessories. The station has 15 small parts, many in thin, clear bags, and takes about 20 seconds per board. Our target is 99.5% success within 20 seconds, the same standard we hold a person to.
We chose this station because it is hard. Small, thin, and soft parts are exactly what a simple gripper can’t handle. It also tests the rule we judge ourselves by: a robot has to take on most of a job. If a person still has to stand beside it the whole time, nobody gains anything.
The full loop now works: a person shows the task, the robot learns it, and we test it on the line.
How Gallop learns
Teaching Gallop works much like training a new colleague.
- General knowledge. We don’t build our own giant AI model. We start from leading open models that already understand how hands and objects behave.
- Watching people work. Experienced workers wear a small head camera, with their consent, so Gallop can learn how the job is really done.
- Practice variations. Simulation generates variations of each example, so small changes on the line don’t surprise the robot.
- Hands-on practice. An operator in a VR headset and gloves guides the robot through the task, on the same parts and with the same cameras as the real station.
- On-the-job learning. Once deployed, what the robot records feeds back into training.
The lesson for us has been that our edge isn’t more data. It is the right data: a real station, real parts, and real quality rules. Our workers are the teachers, and their know-how is what Gallop is built on.
Designed to keep getting better
Inside, Gallop works in four layers. A “planner” looks at the scene, understands the job, and breaks it into steps. A “supervisor”, a second AI, checks every step and re-plans when something goes wrong. A skills layer handles reaching, grasping, and placing. And a control layer moves the arms and hands precisely and stops instantly if anything is wrong.
The layers talk to each other through fixed interfaces. New AI models now arrive every few months, and this design lets us swap a better one in without rebuilding the robot.
All of this runs on a single NVIDIA Jetson AGX Thor computer inside the robot, using standard robotics software. We design the robot body and the hand ourselves. The hand is the hardest part of any humanoid and is still scarce worldwide, so we build ours in-house, with a sense of touch. Motors and other standard components come from established partners.
Today, Gallop works from a fixed base with two arms; a wheeled version comes next. Its strength is pick-and-place work at a station.
Where we are, and what comes next
The Gallop program started in early 2026. Over the coming year, we will bring our own dexterous hand into the robot, put Gallop on wheels, and show it off to customers at Computex 2027. Customer pilots are planned for 2027, and we’re targeting the second half of 2028 for market availability.
Electronics manufacturing comes first, because it is the industry we know best. Because many jobs share the same pick-and-place skills, what Gallop learns in one place can carry over to the next. That is how we see it moving into hospitals and service sites over time.
People often ask whether Gallop is meant to replace workers. It isn’t. It's designed for the repetitive, physically demanding stations that are hardest to staff. The people who teach it move into setting up, supervising, and improving the robots, and their experience becomes more valuable, not less.
A demo shows that a robot can do something once. A factory needs a robot that does the job every day. That is the standard we have set for Gallop, and we will keep sharing our progress against it.
Gallop at a glance
What it is
ASUS’s first humanoid robot, from the Humanoid Robotics BU of the Physical AI Solutions Business Group
Where it works today
A motherboard accessory packing station at an ASUS group EMS factory
Platform
Built on NVIDIA’s robotics platform (Jetson Thor)
Key design choices
In-house dexterous hand with a sense of touch; swappable AI models; works at existing workstations
First markets
Electronics manufacturing, then hospitals and service sites
Timeline
Customer pilots in 2027; market availability targeted for 2H 2028
ASUS (2357.TW) is an innovative technology leader delivering the world’s most comprehensive AI solutions across Infrastructure, Physical AI, AI PC/Devices, and Engineering the AI Advantage. Guided by its “Ubiquitous AI. Incredible Possibilities” strategy, ASUS is bringing enterprise-to-edge AI to life. The ASUS portfolio also includes the ROG and ProArt sub-brands, which serve gamers and creators worldwide.
https://asus.com
