Eka robotics' breakthrough in dexterous manipulation sparks industry interest

Eka Robotics, founded by MIT professor Pulkit Agrawal and former Google DeepMind researcher Tuomas Haarnoja, has closed an $150 million Series A round, valuing the company at $1.45 billion post-money. The startup’s success is attributed to a novel manipulation architecture that combines a force-sensitive gripper with a vision-force-action model, which focuses on the relationship between language, force inputs, and actions. This approach differs from traditional vision-language-action models and addresses the limitations of current generative AI robotics models by incorporating force and touch. Eka’s robot can pick up about 1,500 objects weighing up to 1 kg in an hour with a 95% success rate, surpassing human capabilities. The robot’s ability to adapt to various objects without retraining has impressed experts like Kris Hauser, who compared the experience to the early days of legged locomotion. Agrawal emphasized that vision is not sufficient for precise manipulation, and the robot’s success relies on its ability to react to forces during interaction. Despite the progress, Eka faces challenges in scaling its technology to real-world industrial applications. Agrawal aims to increase the gripper’s durability to 20 million duty cycles, aligning with standard industrial arms. The company is currently testing its technology in secretive pilots with industrial customers, but the transition from lab to production remains a significant hurdle.