Building AI-powered mobile manipulators that maximize dexterity, reach, and industrial productivity.
Dexmate is a robotics company based in Santa Clara, California, founded in 2024 by researchers from MIT, UC San Diego, and Carnegie Mellon University. The company develops AI-powered mobile manipulation systems for industrial automation, combining advanced robotics with embodied AI to automate tasks requiring dexterity, reach, and continuous operation. Its approach prioritizes practical deployment through wheeled mobility and flexible manipulation rather than bipedal locomotion. Its flagship platform, Vega, combines dual robotic arms, dexterous hands, an omnidirectional mobile base, and a foldable torso that extends its working range from ground level to overhead tasks. Designed for manufacturing, logistics, and research, Vega offers long operating times, open developer tools, and high-performance manipulation for real-world industrial environments.
Platform maturity, autonomy stack, and flagship-system specifications in one view.
A concise view of platform maturity and deployment footprint.
Capabilities and model layers highlighted by the company.
Dexmate was founded in 2024 by CEO Tao Chen (MIT PhD, CoRL Best Paper), CTO Yuzhe Qin (UCSD PhD), and COO Chongyang Wang (MIT). They set out to bridge the gap in robotic dexterous manipulation.
Vega is a 171cm, 135kg dual-armed mobile humanoid with omni-directional base and foldable torso reaching 2.2m overhead. It runs 10+ hours continuously at $89,999.
As an NVIDIA Inception startup, Vega won the Red Dot Design Award in 2025. Dexmate sponsored the $200,000 WBCD Challenge at ICRA 2025 and launched an open-source teleoperation device.
Vega units are in testing with publicly traded companies across manufacturing, logistics, and retail.
