The world of table tennis could change dramatically after Sony’s AI division unveiled Ace, an autonomous robot that can compete with professional table tennis players.
Sony AI officials said in a statement that this is the first time a robot has achieved “expert-level play in a competitive sport commonly practiced in the physical world.”
Article continues below
you may like
“This breakthrough is much bigger than table tennis,” Peter Stone, chief scientist at Sony AI, said in a statement. “This is a landmark moment in AI research, demonstrating for the first time that an AI system can effectively perceive, reason, and act effectively in complex, rapidly changing real-world environments that demand precision and speed.”
A study detailing how the robot works was published April 22 in the journal Nature.
Where hardware and software meet
AI systems are already showing great capabilities in strategy games such as Go, chess, and role-playing games.
However, incorporating AI into the body of a robot that combines quick reflexes and physical movements can be more difficult. Here, software and hardware components must work seamlessly. And in table tennis, where speed and hand-eye coordination are important, this combination needs to work to win.
“Table tennis is an extremely complex game that requires not only speed and power, but also split-second decisions,” Peter Duerr, director of Sony AI in Zurich and project leader of Ace, said in a statement. “This research breakthrough highlights the potential of physical AI agents to perform real-time interactive tasks and represents an important step toward creating robots with broader applications in fast, accurate, real-time human interaction.”
Ace’s strategy builds on Sony AI’s previous research on the AI agent Gran Turismo Sophie, which uses advanced sensors and high-speed software to perceive its environment. These sensors include nine active pixel sensor cameras that help Ace determine the exact location of the ball in 3D space, and three line-of-sight systems that use mirrors and event-based vision cameras to measure the spin and angular velocity of the ball as it moves through the air.
Powering these cameras is Sony AI’s proprietary AI control system. It is based on model-free reinforcement learning, where the AI agent learns directly from interactions in the environment without first creating a predictive model. This technology allows Ace to adapt and make decisions more quickly without relying on pre-programmed models.
What to read next
Finally, Ace’s robot body, which features a rotating arm with a paddle-like appendage at the end, was created with the company’s robotic hardware.
beat the pros
In April 2025, scientists pitted Ace against five elite athletes (each with over 10 years of experience and approximately 20 hours of training each week) and two professional table tennis players (Minami Ando and Sho Sone, both from the Japanese Pro League). Both classes of players are skilled at table tennis, but while professional athletes make a living from playing table tennis, elite athletes may not be talented enough to make a living from playing table tennis.
Ace won three out of five matches against elite players and boasted a serve return rate of 75%. Its autonomous system also allowed the robot to return unusual shots, such as a ball bouncing off the net. However, he lost both matches against professionals.
Then, in December 2025, Sony AI had Ace play a series of separate matches against two professional players and two elite players. This time, Ace defeated both elite and professional players. According to company representatives, the robot moved closer to the edge of the table, increasing its shot speed and firing fast-paced volleys at opponents.
Considering that less than two years ago, Google DeepMind’s table tennis robot lost to an elite player, Ace’s victory shows how rapidly this field of robotics has progressed in a short period of time.
“If AI can operate at the level of human experts in these situations, it opens the door to a whole new class of real-world applications that were previously out of reach,” Stone said in a statement.
Source link
