Sony has introduced a new milestone in robotics with “Project Ace,” an advanced AI-driven robot capable of competing with high-level table tennis players. The development highlights how artificial intelligence is rapidly expanding beyond digital environments into complex physical tasks.

Project Ace is designed to operate in one of the fastest and most dynamic sports—table tennis—where success depends on split-second decision-making, precision, and adaptability. According to Sony, the robot has demonstrated the ability to perform at near expert human level, marking a significant step forward for what is often referred to as “physical AI.”

At the core of Ace’s capabilities is a sophisticated system of nine high-speed pixel-sensor cameras that track the ball in real time, mapping its position in 3D space. Additional sensors help measure speed, spin, and trajectory, allowing the robot to react with remarkable accuracy. This sensory input is combined with a reinforcement learning-based control system, enabling the robot to make decisions dynamically rather than relying on pre-programmed responses.

Unlike traditional robotic systems that follow fixed instructions, Ace continuously learns and adapts through gameplay. This allows it to adjust strategies, refine movements, and respond to different playing styles—an essential requirement in a sport known for its unpredictability.

In testing scenarios, Project Ace faced five elite and two professional-level table tennis players. The robot reportedly won three out of five matches against elite players and demonstrated strong serving performance, scoring twice as many direct points as its human opponents in certain instances. These results suggest that AI systems are beginning to match, and in some cases exceed, human performance in controlled physical environments.

The announcement comes at a time when the broader AI industry is heavily focused on “agentic AI”—systems designed to perform tasks autonomously in digital settings. However, Sony’s work underscores that physical AI, which operates in real-world environments, is advancing at a comparable pace.

Experts suggest that mastering a sport like table tennis could have far-reaching implications. The same technologies—real-time perception, adaptive learning, and precision control—can be applied to fields such as manufacturing, healthcare robotics, and service automation.

While Project Ace is still a demonstration of capability rather than a commercial product, it offers a glimpse into a future where AI-powered machines are not only assisting humans but competing with them in complex physical tasks. As physical AI continues to evolve, innovations like this may redefine the boundaries of human-machine interaction.