The Unitree G1 humanoid robot has just gone viral, and it’s not for walking or dancing. This time, it’s playing table tennis with a level of finesse that looks almost human. In a video posted on X by project lead Zhi Su, the G1 can be seen rallying with a human opponent, exchanging over 100 consecutive shots without missing a beat.
Unitree G1 powered by HITTER control system
Behind the robot’s performance is a new control system called HITTER (Humanoid Table Tennis Robot), developed by researchers at UC Berkeley. HITTER allows the G1 to observe, plan, and react to fast-paced play ball speeds reaching over 5 meters per second with barely a fraction of a second to respond.
That’s a major leap for humanoid robots, which have historically struggled in real-time, fast-reaction environments like ball sports. HITTER bridges that gap using a multi-layered control system.
Live photos of the AYN Thor reveal its dual-screen design, retro styling, and powerful Snapdragon hardware ahead of its October release.
Dual-layer control enables fluid motion
The G1 uses 9 cameras to track the ball in real time. A high-level model-based planner functions as the robot’s brain, calculating where the ball will land and deciding how to return it. A low-level controller then takes that plan and executes it using full-body coordination.
What makes the Unitree G1’s performance so impressive:
- Tracks fast-moving balls with 9 synchronized cameras
- Executes real-time motion planning and prediction
- Uses human motion data for realistic movement
- Balances, side-steps, and swings with natural timing
- Fully autonomous, no remote control or external aid
Viral moment signals shift in robotics expectations
The video of the Unitree G1 playing table tennis has sparked a wave of enthusiasm and some disbelief across tech circles online. It’s not just that a robot can hit a ball; it’s that it does so with balance, rhythm, and a sense of purpose that feels eerily human.
This demonstration isn’t about novelty. It’s a clear signal that humanoid robots are beginning to close the gap between mechanical efficiency and lifelike agility, and the next wave of general-purpose robots might be far more capable than we thought.
