Sony released a robot that for the first time defeats professional table‑tennis players.

Sony has created a professional table‑tennis robot

The company Sony announced the creation of a robot named Ace, which became the first machine player capable of competing with top athletes and even winning official matches under the rules of the International Table Tennis Federation (ITTF). The information came from Reuters based on research published in *Nature*.

What Makes Ace Unique
- Physical sport vs. AI

Unlike computer games where algorithms already surpass humans, table tennis remains a challenging task for artificial intelligence due to the need for instant reaction and precise interaction with objects in a confined space.

- Project Goal

To develop a robot that not only knows how to play but also demonstrates human‑like speed and accuracy in perception, planning, and execution of movements in a dynamic environment.

Technical Foundation
System Description Robot Platform 8 joints: 3 for racket positioning, 2 for its orientation, 3 for swing speed and force. Vision Nine traditional cameras around the table + three “gaze‑control systems” measuring angular velocity and ball spin. This allows precise determination of trajectory in 3‑D space.

Test Results
- In April 2025 Ace played five matches:

- Won three times against players with more than ten years of experience.

- Lost two matches to professionals who regularly compete in leagues.

- Professional tennis player Mayuka Taira noted:

> “His actions are hard to predict, he shows no emotions and has superhuman reaction to ball spin. He trains in simulation, so he reacts differently from humans, creating unexpected situations.”

What’s Next
Peter Dürr, director of Sony AI Zurich and project lead for Ace, emphasized:

> “We’re working on helping professional athletes better adapt to such an opponent and find its weak points. This is a key area of our research.”

Thus, the robot Ace has already proven its competitiveness at the level of official tournaments and opens new prospects for the development of AI in real physical interaction.