Eka Robotics raises $150M for physical AI object manipulation

A robotics startup has developed a physical AI model focused on object manipulation, combining a force-sensitive gripper with a deep-learning system trained through physics simulation rather than video datasets. The approach differs from other robotic AI systems by prioritizing force and touch sensing alongside vision, allowing the robot to perform delicate tasks like grasping without crushing fragile objects. Experts in the field have noted the system's ability to adapt to different objects without retraining and its speed compared to human workers, though the company still needs to demonstrate reliable performance in real industrial settings over extended periods. The startup's $150 million funding round at a $1.45 billion valuation reflects investor confidence that this method could bridge the gap between robots that are general-purpose but slow versus robots that are fast but limited to specific tasks.