Sony AI officially announced on April 23, 2026, the successful development of Project Ace, a robotics and artificial intelligence system that marks the first instance of an autonomous machine outplaying elite human competitors in table tennis. The announcement, made at Sony’s global research headquarters, detailed the culmination of a four-year development cycle aimed at solving high-speed physical interaction challenges in robotics. This breakthrough represents a significant leap in real-time sensorimotor control, moving beyond previous iterations of robotic sports systems that struggled with the high velocity and complex spin of professional play.
The Ace system utilizes a specialized high-speed vision array consisting of four synchronized cameras operating at 1,200 frames per second. This hardware configuration allows the AI to track the ball's trajectory, spin, and velocity with a spatial accuracy of less than 0.5 millimeters. Data from these sensors is processed by the Sony Neural Inference Engine, a proprietary hardware-software stack that calculates optimal return paths in under 3.5 milliseconds. This low-latency processing is critical for responding to professional-grade serves, which can reach speeds exceeding 100 kilometers per hour and involve rotational speeds of up to 150 revolutions per second.
According to technical documentation released by Sony AI, the robot’s physical component is a custom-engineered 7-axis robotic arm constructed from lightweight carbon-fiber composites to minimize inertia. The arm’s actuators are capable of accelerations up to 15g, allowing it to cover the entire width of a regulation table tennis table in approximately 0.12 seconds. The AI's decision-making engine was trained using a massive-scale reinforcement learning framework, which simulated over 100,000 hours of gameplay against various styles of play before being transitioned to the physical hardware via a Sim2Real transfer process. This training allowed the system to develop tactical behaviors, such as identifying a human opponent's physical positioning and exploiting weak points in their defensive stance.
In a series of controlled matches conducted prior to the announcement, Ace competed against ten players ranked in the top 50 of international standings. The system recorded a win rate of 82 percent across 500 individual sets. Dr. Hiroaki Kitano, CEO of Sony AI, stated in an official release that the breakthrough represents a significant advancement in the ability of AI to handle complex, high-speed physical tasks that require both strategic foresight and millisecond-level precision. Sony AI confirmed that the underlying technologies developed for Project Ace, including the low-latency control stack and high-speed vision processing, will be integrated into the company’s broader robotics platform for industrial and domestic applications. No specific timeline for commercial deployment of these subsystems was provided during the briefing.