Vega

Dexmate's Vega is a versatile 'Transformer-like' humanoid from the US-based NVIDIA Inception startup founded by PhDs from MIT, UCSD, and Carnegie Mellon. Its foldable torso and arms can compact to just 66 cm for storage/transport or extend up to 2.2 meters for high-reach tasks, unique adaptability in the humanoid market.

Standing 171 cm and weighing 135 kg, Vega moves on an omnidirectional mobile base at 4 km/h. Each 7-DOF arm lifts 15 kg and ends in a 12-DOF five-fingered hand (10 fingers total), providing 36 degrees of freedom overall. Built-in RGB-D cameras, LiDAR, IMUs, ultrasonic sensors, and force/torque sensing feed real-time perception to Intel x86 CPU paired with NVIDIA Jetson AGX Orin GPU (32GB or 64GB configurations).

Vega's high-capacity battery delivers over 10 hours of continuous operation, extending to 30 hours under lighter loads, among the longest runtimes available. The Linux-based system runs Dexmate's ROS-compatible SDK with Python API, URDF/USD simulation files, and optional LLM integration for natural-language tasking. VR and exoskeleton teleoperation support enables intuitive skill programming.

Priced at approximately $89,999 with lead times under four months and a $999 deposit, Vega lowers barriers to full-scale humanoid automation. Rich I/O (Ethernet, WiFi, USB, Bluetooth, DisplayPort, 5V/12V rails) supports custom payloads. Dexmate positions Vega as a Swiss Army Knife platform for researchers and developers needing capable, production-ready mobile manipulation.