Their birthplace is just off the Charles River in Cambridge, Massachusetts, a two-mile walk to historic Fenway Park, so naturally they wear the colors of the local team, the Boston Red Sox. They were born to play baseball.
They are two robots, with surprising dexterity and agility that allows them to play catch and hit a baseball. They were created at RAI, the Robotics and AI Institute, which was founded by Marc Raibert, a former Boston Dynamics engineer.
“We need to make robots smarter, more agile and dexterous, and generally easier to use, more like people,” Mr. Raibert said. “Once we do that, robots and other types of intelligent systems will increase productivity, free people from dangerous work, care for the disabled, and help people live better lives.”
One of the robots can throw a fastball with the velocity of a high school junior varsity pitcher. The robots don’t quite look the part. They are fixed to a platform and have a cylinder-like body with several thick cords attached from the arms to the back. Their torsos are packed with wiring, sensors, and other gizmos that give them function, contained within carbon fiber. Positioned about 23 feet apart in a lab, the two bots pass the time by playing catch and taking swings. One wears a blue Red Sox hat, the other red. A specially designed mitt locks onto one robot’s hand, while the other grips a bat. Their arms, made of lightweight components such as carbon fiber rods, move on flexible joints. Electric actuators give them short bursts of power. The heaviest component — the battery — sits low in the body to provide stability and keep their motions smooth rather than wobbly.
As robotic as they look, when they pitch, swing or field a ball, their movements are strikingly human-like, from the windup before a throw to the arc of a follow-through. The very point of their design is to eliminate the stiff, mechanical quirks that often make robots seem out of sync with the physical world. If fluidity is the goal, these machines are a home run. The robots can move and react quickly to unexpected changes in the environment, a quality that engineers say could be far more important than excelling at batting practice.
“Although not every human being is an athlete, we all possess high levels of athletic intelligence. In humans, athletic intelligence is what enables balance while walking, running, and climbing, utilizing real-time perception to maneuver through terrain and around obstacles and opponents, grasp and manipulate objects, and manage motion energetics to achieve remarkable endurance and strength. The Institute’s work on athletic intelligence will address capabilities of robots that physically move in the world, including their mobility, navigation, and dexterity.”
Seeing robots manage a baseball might look like a niche experiment, but researchers argue the underlying capabilities matter across a wide range of fields, from logistics and sports training to rescue missions, advanced prosthetics and planetary exploration. Baseball provides a tidy proving ground. The sport demands timing, spatial awareness and fine motor coordination — ideal conditions for testing whether a robot can interpret movement and respond appropriately. Behind each throw or swing is a complex combination of motion prediction, fast actuation and split-second choices, the same competencies required in many industrial and scientific environments.
The institute posted a video on YouTube on Nov. 19, 2025. The robots played catch with a human, hit a ball thrown by a human and mimicked a person’s movements — including a short dance sequence reminiscent of the Macarena. It had the unmistakable look of roboticists having fun with their creations, showing off what they could do.
“In this demo, robots play a game of catch and participate in batting practice, both with each other and with skilled humans. The robots are capable of throwing 70mph, approaching the speed of a strong high school pitcher. The robots can catch and bat at short distances (23 feet) requiring quick reaction times to catch balls thrown at up to 41 mph, and hit balls pitched at up to 30 mph.”
The system itself runs on standard computer hardware, paired with software that fuses live camera footage with predictive motion models. As the ball moves, the software tracks its position in three dimensions and determines how the robot should adjust. Sensors feed constant updates into the algorithms, allowing the machine to reposition almost instantly. If the ball suddenly dips or drifts, the robot recalculates in milliseconds — a quick reflex that helps it handle unpredictable trajectories.
The home field is a Cambridge lab marked not by chalk lines, but stacked with hardware. Baseball may be the starting point, but the technology behind these robots is less about the sport and more about teaching machines how to move through the world with the same ease as humans.
