Ziegler highlights Dexmate Vega as a robot research platform
Dexmate’s Vega 1 Pro costs $75,000 at this year’s pricing and ships roughly four weeks after ordering, robotics commentator Lukas M. Ziegler said in a LinkedIn post. He highlighted the bimanual mobile robot as a platform used across several research efforts, arguing that labs have fewer options for buying capable hardware than for sourcing robot intelligence. His post presents Vega as an answer to a practical research bottleneck: obtaining the machine on which a policy can be tested.
According to the specifications Ziegler cited, Vega 1 Pro has 24 degrees of freedom, a torso that spans floor level to 2.2 meters, a 5 kg payload per arm, an omnidirectional base and approximately 20 hours of operation per charge. The listed hardware includes Jetson AGX Thor compute, a stereo head, two LiDAR sensors and four base cameras. Ziegler also lists Python, C++ and Rust APIs, alongside SimReady URDF and USD models matching the shipped robot.
Ziegler points to Dexmate Vega appearances in DeepMind’s Gemini Robotics 2 launch video, Amazon FAR’s real world reinforcement learning work and DUET, a dual robot manipulation framework from USC, Stanford and TRI. He also cites a 100 hour tactile dataset from Berkeley, NVIDIA and Stanford. His central argument is about procurement rather than model performance: teams validating robot policies can otherwise face a wait for a hardware partnership or spend a year building a platform that was never the focus of their research.
Source: Lukas M. Ziegler on LinkedIn
