Sony AI's Ace robot won 3 of 5 matches against high-level amateurs and rallied competitively with professionals under official table tennis rules.
Sony AI researchers published a Nature study detailing Ace, a table tennis robot that combines ball-spin perception, real-time AI decision-making, and an 8-jointed high-speed robotic arm. Ace won 3 out of 5 matches against high-level amateur players and held competitive rallies against two professionals. The robot operates under official table tennis rules, marking a significant milestone in physical-world AI robotics. This is the first documented case of a robot competing at this level in an uncontrolled, real-world physical sport.
Ace's core technical stack — high-frequency ball-spin detection, sub-millisecond decision loops, and closed-loop robotic control — is the blueprint for the next generation of real-world AI agents. The perception-to-action pipeline Sony built here solves problems that simulation-trained models still fail at: unpredictable physical dynamics, sensor noise, and latency-constrained inference. If you're building anything that involves vision + real-time actuation, this Nature paper is a systems architecture reference, not a sports story.
Pull the Sony AI Nature paper on Ace this week and map their perception-decision-action latency pipeline against your own real-time inference stack — specifically where they handle ball-spin classification under 10ms.
Tags