Problem

Most humanoid robots can navigate, perceive, and reason, but their simple grippers remain a bottleneck for interacting with the physical world. Grippers work when objects are rigid, predictable, and precisely positioned; they fail at contact-rich tasks requiring grip adjustment, multiple fingers, tactile feedback, handling soft or delicate objects, and recovery when something slips. The economic pain is the lost usefulness and automation potential of expensive humanoid and robotics systems that cannot reliably perform real-world manipulation. The central use case is dexterous manipulation—tasks such as tool use, picking delicate objects, and ultimately human-like work such as threading a needle.

The problem also has a data dimension. Teleoperating robots can produce training data, but collection scales poorly because it depends on the number of available robots and operator hours, while remote operators do not directly feel the object contact and pressure that guide human manipulation. Proception is therefore addressing both the hardware bottleneck and the scarcity of high-quality tactile manipulation data needed to train capable physical-AI systems.

Product / Service

Proception sells advanced robotic-hand hardware to researchers and robotics companies. ProHand 1.0 is a human-scale, tendon-driven hand with five fingers, a two-degree-of-freedom wrist, 22 total degrees of freedom, integrated tactile sensing, onboard control, and software interfaces for robotics integration. It is designed for dexterous grasping, contact-rich manipulation, teleoperation, and robot-learning workflows, with the benefit of bringing human-like motion and touch to robots that otherwise rely on rigid, limited grippers.

ProGlove extends the system into a wearable data-collection platform. It uses the same skin-like tactile sensing approach, allowing a person to manipulate real objects while capturing hand motion and contact data without a robot in the loop; those movements and skills can then be transferred to ProHand. This combination gives robotics teams both a manipulation end-effector and a way to collect richer human demonstrations more efficiently than robot-only teleoperation.

Market

Proception competes in the emerging physical-AI and humanoid-robotics hardware market, specifically the segment for dexterous robotic hands, tactile sensing, teleoperation, and robot-learning infrastructure. The clearest direct comparator is Shadow Robot, which markets dexterous humanoid hands for research and industrial use; broader humanoid and robotics companies may also compete for the same research and automation budgets. Proception’s differentiation is the integration of a 22-degree-of-freedom hand, tactile sensing, teleoperation support, and a matching human-worn data glove.

The company is an early-stage but financed and launching business rather than a purely conceptual project. Founded in 2024, it was listed as a 10-person Y Combinator company and reports $11 million in seed funding led by First Round Capital, with participation from Y Combinator and BoxGroup. As of August 2026, the first ProHand batch was shipping to researchers and robotics companies and broader ordering had opened. No verified revenue figure or customer count is disclosed in the reviewed primary sources, so its commercial traction is best characterized as initial product shipments and market launch rather than demonstrated scaled revenue.