Sony AI unveiled Ace, the first robot to achieve expert-level performance in a competitive physical sport, defeating multiple professional table tennis players. The findings, published in Nature on April 22, 2026, mark a longstanding milestone for AI and robotics research.

Ace uses nine cameras to track the ball, measures spin by reading the logo on the ball, and has eight joints directing movement. Trained via reinforcement learning, the robot defeated four high-skill players in December 2025 and scored wins against Japanese pros Minami Ando, Kakeru Sone, and world top-25 player Miyuu Kihara by March 2026.

Kinjiro Nakamura, a 1992 Barcelona Olympian who observed Ace play, said no one else would have been able to execute the shot and he did not think it was possible. Sony engineered Ace to match the speed and reach of athletes training 20+ hours weekly to keep matches fair.