Proception AI
Proception AI builds dexterous robotic hands and human-demonstration interfaces for humanoid and manipulation systems. In June 2026 it launched ProHand and ProGlove, began shipping first ProHand batches to researchers and robotics companies, and announced an $11 million seed round led by First Round Capital.
Market context
Dexterous hands are a critical subsystem for humanoid robots. Proception is a focused hand company, so the cleanest mapping is end-effectors rather than full humanoid platforms.
Facts
- Website: https://www.proception.ai/
- HQ: Mountain View, CA, United States
- Founded: 2024
- Segment: End Effectors
- Known funding rounds: 1
Investors
- BoxGroup - Participant - Unknown - 1 funding event - first seen 2026-06-29
- First Round Capital - Lead - Vc - 1 funding event - first seen 2026-06-29
- Y Combinator - Participant - Unknown - 1 funding event - first seen 2026-06-29
Robotics market context
Proception AI is a humanoid-hand company developing ProHand for fine manipulation tasks.
- Target environment: Humanoid robot developers, robotics labs and manipulation research teams.
- Deployment model: Dexterous hand hardware and developer engagement around ProHand.
- Customer context: Teams that need hand-level dexterity for tasks such as threading, grasping and fine object handling.
- Workflow context: Human-like dexterity, manipulation research, humanoid hand integration and fine motor control.
- Commercial maturity: Early-stage company with a public ProHand site and YC launch disclosure.
- Market position: Y Combinator-backed private end-effector company focused on high-dexterity humanoid hands.
- Adoption constraints: Public evidence is still early, with limited visible deployment or customer detail.
- Adjacent context: The company sits beside other hand suppliers rather than full-body humanoid builders.
- Source confidence: high
Team
- Jack Xu - Founder - Xu's robotics and exoskeleton background connects Proception's hardware and data-interface work to dexterous manipulation research.
- Jay Li - Founder - Li's Optimus and humanoid robotics background connects directly to Proception's dexterous robotic hand and manipulation data work.
Context tags
- Dexterous manipulation - Capability - ProHand and ProGlove target dexterous manipulation and human-demonstration data.
- Dexterous hand OEM - Stack Layer - Proception supplies dexterous robotic hand hardware.
- Humanoid dexterous hand platform - Peer Group - ProHand is marketed as a humanoid hand platform.
- Humanoid component supply - Peer Group - ProHand is positioned as a hand component for humanoid and manipulation platforms.
- Dexterous robot hand builder - Peer Group - YC and TechCrunch identify Proception around dexterous robotic hands.
- Dexterous hand platforms - Peer Group - Proception builds ProHand as a dexterous robotic hand for humanoid and manipulation systems.
- Dexterous in-hand manipulation - Workflow - The product is positioned for contact-rich fine manipulation.
- Tactile sensing - Capability - The product stack uses tactile sensing and a sensor-packed glove skin.
- Tactile robot hand control - Capability - ProHand uses tactile sensing and glove-based sensor skin context.
- Humanoid human-in-loop training - Workflow - ProGlove captures human hand interaction data for robot training workflows.
- Robotics data infrastructure layer - Stack Layer - ProGlove creates human-demonstration data for dexterous manipulation.
Proception AI canonical Korthos profile