Game, Set, Bot: Sony AI’s ‘Ace’ Serves Up a Defeat to Table Tennis Pros

2026-04-22 · Source: Big Data & AI News - EE Times · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Advanced, medium

Summary

Sony AI, established in April 2020, has developed "Ace," an autonomous robot designed to compete with human table tennis players, claiming it is the first robot to achieve expert-level performance in a competitive physical sport. Building on its Gran Turismo Sophy AI agent, Sony AI utilized gaming environments as benchmarks for training. The robot employs a reinforcement learning approach, trained through thousands of hours of simulated table tennis, to learn optimal ball return strategies. Ace integrates a perception system with nine active pixel sensor cameras and three gaze control systems for 3D ball triangulation and spin measurement, alongside an asymmetric actor-critic deep reinforcement learning policy for control. Its custom hardware features eight degrees of freedom for agility. In April 2025, Ace won three out of five matches against elite players, achieving a return rate above 75%, but lost to professional players.

Key takeaway

For robotics engineers developing physical AI agents for real-time interactive tasks, consider leveraging virtual gaming environments for large-scale reinforcement learning. Your team should prioritize robust perception systems capable of high-speed, precise tracking and integrate them with custom hardware to bridge the simulation-to-reality gap, especially for applications requiring rapid human interaction or high-stakes physical control.

Key insights

Gaming environments serve as safe, scalable testbeds for developing and benchmarking advanced physical AI agents.

Principles

• Reinforcement learning excels in complex, dynamic physical tasks.
• Simulation-to-reality transfer requires robust perception and control.
• High-speed physical interaction demands precise real-time sensing.

Method

Sony AI trains robots using large-scale reinforcement learning in virtual gaming environments, then integrates this with advanced perception systems (APS cameras, GCSs) and custom robotic hardware for real-world deployment.

In practice

• Use virtual environments for initial AI agent training.
• Implement event-based vision for high-speed object tracking.
• Combine learning policies with model predictive control.

Topics

• Sony AI
• Ace Robot
• Table Tennis AI
• Reinforcement Learning
• Physical AI

Best for: AI Scientist, Robotics Engineer, Research Scientist

Related on AIssential

• This robot can beat you at table tennis — The "Ace" robot demonstrates AI's ability to surpass elite human performance in complex physical sports.
• AI-powered robot beats elite table tennis players — Sony AI's Ace robot demonstrates advanced real-time perception and control in complex human-robot interaction.
• Watch Sony’s elite ping-pong robot beat top-ranked players — Sony's Ace robot demonstrates AI's capability to excel in complex physical sports like table tennis against top human players.
• Sony AI table tennis robot outplays elite human players — Ace is the first robot to defeat elite human players in a competitive physical sport.
• Table tennis robot defeats some of world’s best players – why this has major implications for robotics — Ace robot demonstrates AI's ability to achieve superhuman performance in complex, real-world physical environments.
• Meet Ace, the table-tennis robot that can beat elite players — High-speed perception systems enable robots to achieve elite performance in complex, dynamic tasks like table tennis.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Big Data & AI News - EE Times.